Family Practice

Bridgwood, Bernadeta;Park, John;Hawcroft, Claire;Kay, Natasha;Tang, Eugene

2017 Family Practice

doi: 10.1093/fampra/cmx101pmid: 29045622

Abstract This systematic review describes how international exchange programmes in primary care have been received and evaluated. Electronic databases (MEDLINE, Embase, PsycINFO, EBM reviews, CAB abstracts and PubMED) were searched to identify articles where the main focus of the study was exchanges undertaken in primary care/family medicine until March 2016. Articles were included if they (i) discussed participant exchanges in primary care; (ii) presented associated outcome data—this included (a) individual/group experience of exchange; (b) mechanism of exchange and (c) observations during the exchange. A narrative synthesis was performed of the heterogeneous data identified. Twenty-nine studies were included. Exchange locations varied across the world with the largest number in Europe. Participants came from a range of backgrounds including medical students, nurses, General Practitioners (GP), GP trainees (GPTs) and visiting scholars/professors. Exchange duration ranged from 3 days to 2 years. Key themes were identified from analysis of the studies with illustrative quotes from the included studies provided. Four key areas were discussed in relation to exchange experience: learning opportunities and new knowledge; comparative observation; knowledge gained and translational learning. Primary care international exchanges provide a rich source of cross-country learning. This review identified that exchange participants benefit both personally and professionally, equipping them with translatable skills to improve the care provided to their patients. Academic medicine, access to healthcare, continuing medical education, family health, immigrant health, International health, primary care Introduction At the forefront of many healthcare systems are primary care/family physicians who provide the foundations of healthcare provision, offering continuity of care for their patients and families. In some settings, these physicians have the role of specialist generalists; in others they act as ‘gatekeepers’, allocating community resources, as well as referring patients as necessary for specialist management. Increasingly, there is a limitation of resources, aging populations and a larger emphasis on delivering care in the community (1). Furthermore, the face of our population is changing—with our world today reflecting an interconnected multicultural society. It is important for healthcare professionals to have both knowledge and understanding of the varied cultural backgrounds and beliefs of the populations that they serve, to provide the best and most appropriate health care (1–3). There is growing evidence that exposure to other healthcare systems enhances cultural understanding, empathy and communication (4,5). The challenges provided by the changing face of healthcare demands innovation and change to provide improved efficiency and outcomes. Primary care/family physicians are positioned at the front line in this period of rapid change, and may have unique experience and perspectives (2). One mechanism to enhance knowledge, which may help to inform and develop primary care, is that of exchange programs (1). The concept of gaining a working insight into different healthcare systems has grown in popularity (6,7). The notion of undertaking an exchange has shifted from an initial mindset of medical students enjoying an elective, to professionals of all levels seeking a learning opportunity for personal and professional development. What is not clear is how this type of experience has been specifically received and evaluated by participants undertaking exchanges within primary care/family medicine and we seek to address this within this systematic review. Objective The objective of this study was to conduct a systematic review of the literature relating to exchanges within primary care/family medicine. It aimed to identify who undertakes exchanges in primary care, what exchanges are undertaken and why, and the value of such exchanges in terms of learning outcomes and experience gained. Methods Article identification PRISMA guidelines were followed to identify articles where the main focus of the study was exchanges undertaken in primary care/family medicine (Fig. 1). A search of electronic databases (MEDLINE, Embase, PsycINFO, EBM reviews, CAB abstracts and PubMED) was conducted using the following terms: exchange, general practitioners/physicians, family physician, primary care, general practice, GP, family medicine, family practi*, primary health care, primary care physician, family doctor, primary medical care. The search was performed by a reviewer (ET) of all articles available until the search date of the 4 March 2016. Figure 1. View largeDownload slide PRISMA flow diagram of information through the different phases of the systematic review Figure 1. View largeDownload slide PRISMA flow diagram of information through the different phases of the systematic review Inclusion criteria Articles were included if they (i) discussed participant exchanges in primary care; (ii) presented associated outcome data—this included (a) individual/group experience of exchange; (b) mechanism of exchange and (c) observations during the exchange. All articles types were considered including non-English articles. Data extraction Two independent reviewers (BB, JP) assessed the article titles and identified relevant articles. Any discrepancies were resolved by discussion with a third reviewer (ET). A structured data extraction form was used whereby categories were devised during the initial phase of the data extraction. This included exchange participant details; exchange details; participant experience; host details and participant exchange evaluation. Three reviewers (CH, NK, BB) independently evaluated full text articles, extracted and combined data. Due to the heterogeneity of the studies, results are presented as a narrative synthesis and no additional analysis was performed. Results are reported as—publication number, PN (level of training) and either ‘direct quote from a participant’ or an extract from the publication. Data analysis Here narrative analysis was employed to describe the exchange demographics, exchange objectives, mechanisms of exchange and exchange outcomes (8). The program NVivo 11 was used to aid the data assessment. Results Exchange demographics Twenty-nine papers were identified which fulfilled the search criteria (Fig. 1). Table 1 summarizes the characteristics of the participants, location of exchange, mechanism of exchange and duration. Exchange locations varied across the world with the largest number in Europe (PN 2,4,6,8,11,12,13,14,16,18,19,20,24,26,27,29). Participants came from a range of backgrounds including medical students (PN 5,12,19,25,28) through to GP scholars/professors (PN 7). Exchange duration ranged from 3 days (PN 14) to 2 years (PN 15). One report discussed a 1-day exchange in a practice before attending a primary care conference (conference exchange, PN 16). The review identified a number of exchange programs including Hippokrates (PN 2,4,11,13,16,18,20,26,27) Erasmus exchanges (PN 12,19), an Irish/Australian registrar exchange program (PN 1,8), International Healthcare Fellowship, FM360 (Family Medicine 360, PN 10), SanteSud (PN 3) and international health fellowship (PN 5,9). The remainder were either organized through universities/work-place (PN 14,17,25,28) or were self organized (PN 6,7,15,21-24). Table 1. A summary of the information extracted from the publications segregated into publication number; Stage of Training—General practitioner/family doctor (GP), General Practitioner Trainers (T), GPs in training (GTP), medical students (MS), Nurses (N) Professor (P); Country/continent visited; Duration of exchange and type of exchange Publication no Stage of training Location Duration Type of exchange (self organized/ organized and program) Reference 1 GPT Donegal, Ireland 3 months Organized—Irish Registrar Exchange Program Wearne E. Aust fam physician. 2008;37(3):158. 2 GPT/GP Portugal 2 weeks Organized—Hippokrates VdGM Hiam L. InnovAiT. 2014;8(8):507. 3 GP Mali 2 weeks Organized—Sante-Sud Van Dormael. Inter Fam Med Ed. 2008;40(3):211. 4 GP England 1 week RCGP/APM-CG Jelley D.EJGP. 2002;8(2):75. 5 MS Ghana and Nigeria 8 months Organized—International Health Fellowship Smilkstein G. Acad med. 1990;65(12):781. 6 GP France, Czech Republic 1 week Organized—GP network Cembrowicz S. BJGP. 2002;52(474)78. 7 P Venezuela 18 months Self-organized Ventres W. Fam pract. 1995;12(3):324. 8 GPT Ireland 2–3 months Organized—Irish Registrar Exchange Program Kinsella P. Aust Fam Physician. 2008;37(9):739. 9 GPT Malawi and Australia 4 months Organized—International Health Fellowship Dowling S. Educ Prim Care 2015;26(6):388. 10 GPT/GP Worldwide 4 weeks Organized—FM 360 Barata A. J Fam Med Prim Care. 2015;4(3):305. 11 GPT/GP Germany N/A Organized—Hippokrates VdGM Carmienke S. Z Allg Med. 2014; 90(1):43. 12 MS Europe not given Organized—EU Socrates Programme van Weel C EJGP. 2005;11(3–4):122. 13 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S. Educ Prim Care. 2015;26(4):282. 14 T Netherlands 3 days Organized—University partnership Zwart S. Educ Prim Care. 2013;24(25):380. 15 GPT Worldwide 1–2 years Self-organized Munir S. Educ Prim Care.2013;24(5)303 16 GPT UK 1-day pre-conference Organized—Hippokrates VdGM Villanueva T. Innovait. 2010;3(1):697. 17 N New Zealand 2 weeks Organized—NHS equal opportunities unit Gould E. Nurs New Zeal. 1998;4(7):23. 18 GPT/GP Italy 2 weeks Organized—Hippokrates VdGM Loveday L. InnovAiT. 2012;5(8):484. 19 MS Slovenia 7 weeks Erasmus through Ljubljana medical school Rotar-Pavlic D. Acta Med Acad. 2012;41(1):47. 20 GPT/GP France 2 weeks Organized—Hippokrates VdGM Metcalfe N. Educ Prim Care. 2015;25(5):356. 21 GP Australia 4 months Self-organized Rhodes C. J RCGP. 1979;29(202):302. 22 GP Australia N/A Self-organized Pearce C. Aus J Rural health. 2000;8(4):218. 23 GP Canada 5 month Self-organized Marsh G. BMJ. 1971;1(5744):33. 24 GP UK 5months Self-organized Sweeny G. BMJ. 1971;1(5744):337. 25 MS USA and ‘others’ 1 year Organized—ministry of health, labor and welfare Kitamura K. Fam Med. 2002;34(10):761. 26 GPT/GP Poland 2 weeks Organized—Hippokrates VdGM Willoughby H. BJGP. 2012;62(602):486. 27 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S.EJGP. 2014;20(1):58 28 MS Argentina 6 weeks Organized—Mcgill medical student incentive Trop I. Can Fam Physician 1993;39:2600 29 GP Turkey Not given Organized—VdGM Van Hoorick, J Fam Med Prim Care. 2016;5(2):220. Publication no Stage of training Location Duration Type of exchange (self organized/ organized and program) Reference 1 GPT Donegal, Ireland 3 months Organized—Irish Registrar Exchange Program Wearne E. Aust fam physician. 2008;37(3):158. 2 GPT/GP Portugal 2 weeks Organized—Hippokrates VdGM Hiam L. InnovAiT. 2014;8(8):507. 3 GP Mali 2 weeks Organized—Sante-Sud Van Dormael. Inter Fam Med Ed. 2008;40(3):211. 4 GP England 1 week RCGP/APM-CG Jelley D.EJGP. 2002;8(2):75. 5 MS Ghana and Nigeria 8 months Organized—International Health Fellowship Smilkstein G. Acad med. 1990;65(12):781. 6 GP France, Czech Republic 1 week Organized—GP network Cembrowicz S. BJGP. 2002;52(474)78. 7 P Venezuela 18 months Self-organized Ventres W. Fam pract. 1995;12(3):324. 8 GPT Ireland 2–3 months Organized—Irish Registrar Exchange Program Kinsella P. Aust Fam Physician. 2008;37(9):739. 9 GPT Malawi and Australia 4 months Organized—International Health Fellowship Dowling S. Educ Prim Care 2015;26(6):388. 10 GPT/GP Worldwide 4 weeks Organized—FM 360 Barata A. J Fam Med Prim Care. 2015;4(3):305. 11 GPT/GP Germany N/A Organized—Hippokrates VdGM Carmienke S. Z Allg Med. 2014; 90(1):43. 12 MS Europe not given Organized—EU Socrates Programme van Weel C EJGP. 2005;11(3–4):122. 13 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S. Educ Prim Care. 2015;26(4):282. 14 T Netherlands 3 days Organized—University partnership Zwart S. Educ Prim Care. 2013;24(25):380. 15 GPT Worldwide 1–2 years Self-organized Munir S. Educ Prim Care.2013;24(5)303 16 GPT UK 1-day pre-conference Organized—Hippokrates VdGM Villanueva T. Innovait. 2010;3(1):697. 17 N New Zealand 2 weeks Organized—NHS equal opportunities unit Gould E. Nurs New Zeal. 1998;4(7):23. 18 GPT/GP Italy 2 weeks Organized—Hippokrates VdGM Loveday L. InnovAiT. 2012;5(8):484. 19 MS Slovenia 7 weeks Erasmus through Ljubljana medical school Rotar-Pavlic D. Acta Med Acad. 2012;41(1):47. 20 GPT/GP France 2 weeks Organized—Hippokrates VdGM Metcalfe N. Educ Prim Care. 2015;25(5):356. 21 GP Australia 4 months Self-organized Rhodes C. J RCGP. 1979;29(202):302. 22 GP Australia N/A Self-organized Pearce C. Aus J Rural health. 2000;8(4):218. 23 GP Canada 5 month Self-organized Marsh G. BMJ. 1971;1(5744):33. 24 GP UK 5months Self-organized Sweeny G. BMJ. 1971;1(5744):337. 25 MS USA and ‘others’ 1 year Organized—ministry of health, labor and welfare Kitamura K. Fam Med. 2002;34(10):761. 26 GPT/GP Poland 2 weeks Organized—Hippokrates VdGM Willoughby H. BJGP. 2012;62(602):486. 27 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S.EJGP. 2014;20(1):58 28 MS Argentina 6 weeks Organized—Mcgill medical student incentive Trop I. Can Fam Physician 1993;39:2600 29 GP Turkey Not given Organized—VdGM Van Hoorick, J Fam Med Prim Care. 2016;5(2):220. Summary—Africa (3) Australasia (4) Canada (1) Ireland (2) Europe (12) S America (2) UK (3) USA (1) Worldwide (2). View Large Table 1. A summary of the information extracted from the publications segregated into publication number; Stage of Training—General practitioner/family doctor (GP), General Practitioner Trainers (T), GPs in training (GTP), medical students (MS), Nurses (N) Professor (P); Country/continent visited; Duration of exchange and type of exchange Publication no Stage of training Location Duration Type of exchange (self organized/ organized and program) Reference 1 GPT Donegal, Ireland 3 months Organized—Irish Registrar Exchange Program Wearne E. Aust fam physician. 2008;37(3):158. 2 GPT/GP Portugal 2 weeks Organized—Hippokrates VdGM Hiam L. InnovAiT. 2014;8(8):507. 3 GP Mali 2 weeks Organized—Sante-Sud Van Dormael. Inter Fam Med Ed. 2008;40(3):211. 4 GP England 1 week RCGP/APM-CG Jelley D.EJGP. 2002;8(2):75. 5 MS Ghana and Nigeria 8 months Organized—International Health Fellowship Smilkstein G. Acad med. 1990;65(12):781. 6 GP France, Czech Republic 1 week Organized—GP network Cembrowicz S. BJGP. 2002;52(474)78. 7 P Venezuela 18 months Self-organized Ventres W. Fam pract. 1995;12(3):324. 8 GPT Ireland 2–3 months Organized—Irish Registrar Exchange Program Kinsella P. Aust Fam Physician. 2008;37(9):739. 9 GPT Malawi and Australia 4 months Organized—International Health Fellowship Dowling S. Educ Prim Care 2015;26(6):388. 10 GPT/GP Worldwide 4 weeks Organized—FM 360 Barata A. J Fam Med Prim Care. 2015;4(3):305. 11 GPT/GP Germany N/A Organized—Hippokrates VdGM Carmienke S. Z Allg Med. 2014; 90(1):43. 12 MS Europe not given Organized—EU Socrates Programme van Weel C EJGP. 2005;11(3–4):122. 13 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S. Educ Prim Care. 2015;26(4):282. 14 T Netherlands 3 days Organized—University partnership Zwart S. Educ Prim Care. 2013;24(25):380. 15 GPT Worldwide 1–2 years Self-organized Munir S. Educ Prim Care.2013;24(5)303 16 GPT UK 1-day pre-conference Organized—Hippokrates VdGM Villanueva T. Innovait. 2010;3(1):697. 17 N New Zealand 2 weeks Organized—NHS equal opportunities unit Gould E. Nurs New Zeal. 1998;4(7):23. 18 GPT/GP Italy 2 weeks Organized—Hippokrates VdGM Loveday L. InnovAiT. 2012;5(8):484. 19 MS Slovenia 7 weeks Erasmus through Ljubljana medical school Rotar-Pavlic D. Acta Med Acad. 2012;41(1):47. 20 GPT/GP France 2 weeks Organized—Hippokrates VdGM Metcalfe N. Educ Prim Care. 2015;25(5):356. 21 GP Australia 4 months Self-organized Rhodes C. J RCGP. 1979;29(202):302. 22 GP Australia N/A Self-organized Pearce C. Aus J Rural health. 2000;8(4):218. 23 GP Canada 5 month Self-organized Marsh G. BMJ. 1971;1(5744):33. 24 GP UK 5months Self-organized Sweeny G. BMJ. 1971;1(5744):337. 25 MS USA and ‘others’ 1 year Organized—ministry of health, labor and welfare Kitamura K. Fam Med. 2002;34(10):761. 26 GPT/GP Poland 2 weeks Organized—Hippokrates VdGM Willoughby H. BJGP. 2012;62(602):486. 27 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S.EJGP. 2014;20(1):58 28 MS Argentina 6 weeks Organized—Mcgill medical student incentive Trop I. Can Fam Physician 1993;39:2600 29 GP Turkey Not given Organized—VdGM Van Hoorick, J Fam Med Prim Care. 2016;5(2):220. Publication no Stage of training Location Duration Type of exchange (self organized/ organized and program) Reference 1 GPT Donegal, Ireland 3 months Organized—Irish Registrar Exchange Program Wearne E. Aust fam physician. 2008;37(3):158. 2 GPT/GP Portugal 2 weeks Organized—Hippokrates VdGM Hiam L. InnovAiT. 2014;8(8):507. 3 GP Mali 2 weeks Organized—Sante-Sud Van Dormael. Inter Fam Med Ed. 2008;40(3):211. 4 GP England 1 week RCGP/APM-CG Jelley D.EJGP. 2002;8(2):75. 5 MS Ghana and Nigeria 8 months Organized—International Health Fellowship Smilkstein G. Acad med. 1990;65(12):781. 6 GP France, Czech Republic 1 week Organized—GP network Cembrowicz S. BJGP. 2002;52(474)78. 7 P Venezuela 18 months Self-organized Ventres W. Fam pract. 1995;12(3):324. 8 GPT Ireland 2–3 months Organized—Irish Registrar Exchange Program Kinsella P. Aust Fam Physician. 2008;37(9):739. 9 GPT Malawi and Australia 4 months Organized—International Health Fellowship Dowling S. Educ Prim Care 2015;26(6):388. 10 GPT/GP Worldwide 4 weeks Organized—FM 360 Barata A. J Fam Med Prim Care. 2015;4(3):305. 11 GPT/GP Germany N/A Organized—Hippokrates VdGM Carmienke S. Z Allg Med. 2014; 90(1):43. 12 MS Europe not given Organized—EU Socrates Programme van Weel C EJGP. 2005;11(3–4):122. 13 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S. Educ Prim Care. 2015;26(4):282. 14 T Netherlands 3 days Organized—University partnership Zwart S. Educ Prim Care. 2013;24(25):380. 15 GPT Worldwide 1–2 years Self-organized Munir S. Educ Prim Care.2013;24(5)303 16 GPT UK 1-day pre-conference Organized—Hippokrates VdGM Villanueva T. Innovait. 2010;3(1):697. 17 N New Zealand 2 weeks Organized—NHS equal opportunities unit Gould E. Nurs New Zeal. 1998;4(7):23. 18 GPT/GP Italy 2 weeks Organized—Hippokrates VdGM Loveday L. InnovAiT. 2012;5(8):484. 19 MS Slovenia 7 weeks Erasmus through Ljubljana medical school Rotar-Pavlic D. Acta Med Acad. 2012;41(1):47. 20 GPT/GP France 2 weeks Organized—Hippokrates VdGM Metcalfe N. Educ Prim Care. 2015;25(5):356. 21 GP Australia 4 months Self-organized Rhodes C. J RCGP. 1979;29(202):302. 22 GP Australia N/A Self-organized Pearce C. Aus J Rural health. 2000;8(4):218. 23 GP Canada 5 month Self-organized Marsh G. BMJ. 1971;1(5744):33. 24 GP UK 5months Self-organized Sweeny G. BMJ. 1971;1(5744):337. 25 MS USA and ‘others’ 1 year Organized—ministry of health, labor and welfare Kitamura K. Fam Med. 2002;34(10):761. 26 GPT/GP Poland 2 weeks Organized—Hippokrates VdGM Willoughby H. BJGP. 2012;62(602):486. 27 GPT/GP Europe 2 weeks Organized—Hippokrates VdGM Rigon S.EJGP. 2014;20(1):58 28 MS Argentina 6 weeks Organized—Mcgill medical student incentive Trop I. Can Fam Physician 1993;39:2600 29 GP Turkey Not given Organized—VdGM Van Hoorick, J Fam Med Prim Care. 2016;5(2):220. Summary—Africa (3) Australasia (4) Canada (1) Ireland (2) Europe (12) S America (2) UK (3) USA (1) Worldwide (2). View Large Exchange objectives Exchange objectives appeared to focus on providing a broader intercultural and clinical learning experience. Participants discussed the opportunity to observe different healthcare systems and to potentially identify examples of best practice to translate into their own practice. Additionally, exchanges were seen to facilitate improving language and interpersonal skills and establishing personal and professional bonds. PN 18 (GPT/GP) ‘…to encourage mobility and sharing of knowledge among young doctors across Europe’. PN 21 (GP) …gaining varied experiences of the delivery of primary care, seeking knowledge to improve care, exchange of ideas. Organising an exchange Exchanges were organized either through direct contact with potential hosts via adverts or through discussion. Alternatively exchanges were arranged through dedicated exchange programs or organisations. The success of establishing an exchange appeared to depend on the dedication and accessibility of a host and appeared easier when organisations offered ‘readily-available’ and willing hosts. PN 6 (GP) ‘Successful exchanges seem to depend on a single dependable contact abroad, a language in common, and a coherent group’. PN 10 (GPT/GP) Established structure - used ERASMUS guidance. For exchanges based beyond Europe, participants discussed some difficulty regarding bureaucratic requirements of working and travelling abroad, such as obtaining visas. PN 22 (GP) ‘It’s multiple bureaucracies… time consuming’. Furthermore, UK trainees and those undertaking longer exchanges identified different levels of support in different training deaneries or parts of the NHS. PN 15 (GPT) ‘Trainees face challenges at all stages … from the planning and application process…. Across the UK, there is disparity in the level of deanery support for trainees wishing to take OOPE(Out of Program Exchange)…’. Exchange evaluation and experience From the views of participants, it was clear that for many, exchanges were an opportunity to gain new knowledge in a new environment. PN 3 (GP) ‘(exchange can)…stimulate reflection and inspire practice innovations (…) poorer countries can get useful insights on improving individual clinical care (…) richer countries can learn about their population and public health responsibilities’. PN 25 (MS) All trainees reported an increase in skills developed and medical knowledge. Furthermore, participants identified the exchange as an opportunity to extend their personal and professional networks. Participants overwhelmingly compared their own home practice to their host practice. Comparisons were typically made of the structure of practice, duties of a GP, patient expectations and other multidisciplinary team members (MDT). PN 2 (GPT/GP) ‘Overall, what struck me is the universal nature of GP. Despite notable differences, we all face similar challenges’. PN 17 (N) ‘…structural differences with nurses in New Zealand - more GPs see patients than nurses. UK has more scope to develop nurse role.’ PN 1 (GPT) I was witness to new knowledge, a new way of doing things; then at times I felt very strongly how similar our issues and experiences of training are. On completion of an exchange, there was reflection on learning outcomes. Participants discussed personal gains—including an improvement in networking skills, increased understanding of another culture, positive attitudes towards migrant health needs, language skills and augmented organisation and flexibility. PN 7 (P) (…participants) reported improved knowledge and skills, and increased awareness of health issues in another country. Moreover, they learned about cultural differences and had an opportunity to reflect and grow personally as doctors. PN 26 (GPT/GP) ‘It has broadened my perspective of general practice, opened doors of opportunity and, most importantly, enabled me to make some wonderful friends for life’. Evidence of improved professional skills included enhanced communication and consulting skills, a range of practical clinical skills, understanding and utilisation of resources—particularly of resource limitations—and enhanced intercultural medical professionalism. PN 15 (GPT) It was felt that trainees returned with a broader exposure to clinical presentations and greater confidence in their UK practice. PN 28 (MS) ‘The project (…) allowed us to perceive health care in a different setting ripe with challenges for growth and development; to highlight the importance of primary care in assuring the health of a population’. Benefits of exchange As participants identified the opportunity to discuss and share experience and practice with both their host and home organisations, this conferred the potential to share areas of best practice, including teaching practice. PN 4 (GPT/GP) ‘there is much to be learned from observing another system and comparing this to our NHS. I feel the NHS could be improved by focusing on health promotion and health education… A growing international community provides a forum to discuss and share experiences, and consider how we can improve’. PN 12 (MS) A further benefit of the exchange programme lies in the transfer of teaching innovations between universities. PN 29 (GP) We learned about each other’s habits and cultures but also about each other’s incertitude and fears. Moreover, research and education networks were formed, upon which strong foundations for future exchanges were established. PN 28 (MS) Strong educational links have resulted from this sojourn: a resident from the CEMIC has (already) completed 3 months of family medicine at McGill University. Importantly, participants described that they intended to translate their new knowledge and skills into their own practice within their home countries. PN 24 (GP) ‘I returned to Canada feeling quite humbled and took back with me ideas and concepts that should effectively help our group to reorganize its efforts to deliver better community first-contact health care’. PN 9 (GPT) All trainees reported an increase in skills development and medical knowledge as a result of the (…exchange). Trainees stated that they became less reliant on diagnostic tests and placed stronger emphasis on history taking and physical examination. Discussion This review aimed to evaluate the objectives and experiences of undertaking an exchange in primary care in addition to identifying the individual and organisational benefits. Our review supports the theory that, for any activity with educational value, the importance of setting learning objectives and evaluating attainment is important (8–10). There is evidence that exchange participants aimed to understand global healthcare and share knowledge (10). However, to facilitate positive learning outcomes, adequate preparation and supervision are key elements in addition to pastoral support such as adjusting to the local culture (11). Evaluating the exchange experience in primary care highlighted four key areas—learning opportunities; comparative observation; knowledge gained and translational learning. International exchanges are well established in the medical student population (12,13), however have filtrated through all MDT roles and levels of experience (14–16). International exchanges in primary care have evolved from curiosity between individual practitioners (13,15,16), to established networks such as the Vasco da Gama Movement (VdGM) (6,17,18). There is obvious scope for sharing good practice in a coordinated manner to reduce duplication of organisational effort and resource (19). Indeed, there are forums for sharing exchange experiences on an international level during the World Organisation for Family Medicine (WONCA) Conference (6,20). Engaging in a global health partnership would actively seek to make better connections and promote learning and improve global healthcare standards. Primary care operates within multicultural populations; however, problems with basic concepts such as communication are widespread (21–23). The importance of improving skills and knowledge in global health is being increasingly discussed within medical schools, and even in family medicine training, to ensure that graduates are appropriately prepared to work in an international world (23,24). International student exchanges across disciplines have evidenced benefits such as developing professionalism; broadening subject knowledge, cultural awareness and personal skills (17,25,26). Within this review, exchange participants evidenced educational gains offering potential benefits to both the primary care system of the exchange participant and the host. Participants acknowledged an increased confidence and gain in clinical skills, while simultaneously developing communication skills and an appreciation of varied cultural etiquette (1,5,27,28). Importantly, exchange participants identified an increased cultural understanding and discussed a deeper empathy towards their patients (1). This is driven by the individual’s requirement to personally engage with a different culture and engage in the learning process to adequately acquire an intercultural identity and develop competencies such as communication (28,29). Implications for practice An increased demand on resources within primary care has placed an emphasis on the transformation of existing models of care. Moreover, social and economic migration has resulted in a culturally diverse population that requires practitioners to have a broader perspective on the communication and care provided (20,21,26,30). General practitioners need to be clinically competent and have a cultural understanding of both their patients and work teams. This enables practitioners to adapt to different social settings and provide high standard, holistic care in multiethnic teams and environments (5,18,27,28,31,32). Primary care is a relatively new speciality in some countries and well established in others. There are an enormous number of valuable insights, procedures, structures and mechanisms practised within primary care globally, which are available to observe and evaluate (8,20). International exchanges provide a route to allow exposure to global healthcare practice, finding similarities, common challenges and possible solutions (1,12). Strengths and limitations This is the first review of publications concerned with international exchange in primary care. We were able to identify the range of professional positions/training levels who undertake exchanges, identify available exchange opportunities and discuss the potential value of the exchanges in terms of personal and professional benefits. Participants were able to compare practice in home and host organisations and identify areas of practice which may conceivably be implemented into their own practice. Some of the practicalities of organising an exchange—including potential problems—are discussed. This gives potential participants realistic expectations of undertaking such an activity. To mitigate any researcher bias during data analysis, the study employed a pre-defined extraction template and narrative synthesis to describe the data and ensure reproducibility. Conclusion Primary care physicians are often the initial contact patients have with the healthcare system and are under pressure to provide patient-centred, community focused care with reducing resources. How primary care is delivered worldwide varies in terms of information systems, team structures, payment incentives and guidelines. Exchanges provide a rich source of cross-country learning by which healthcare physicians can experience alternative practice and consider methods to improve best practice and develop improved models of care. By providing a global perspective, this review has identified that exchange participants benefit both personally and professionally equipping them with translatable skills to improve the care provided to their patients. It is urged that international exchange opportunities are promoted—and organisational challenges reduced—at every stage of physician training to enable more primary care physicians to participate in this great learning experience for the benefit of our primary care community and our patients. To learn more about how you can contact these exchange organisations and participant in an exchange, please see Table 2. Table 2. List of current opportunities for international exchanges Exchange organisation Website address Hippokrates exchange http://vdgm.woncaeurope.org/content/exchanges Family medicine 360 http://www.vdgm.woncaeurope.org Sante Sud http://www.santesud.org/ Global health fellowship https://gprecruitment.hee.nhs.uk/Recruitment/GHF McGill university International exchange https://www.mcgill.ca/familymed/global-health/projects/studentexchange Erasmus plus exchanges http://www.rcgp.org.uk/jic General International opportunities http://www.rcgp.org.uk/rcgp-near-you/rcgp-international/international-opportunities.aspx https://www.cfhi.org Exchange organisation Website address Hippokrates exchange http://vdgm.woncaeurope.org/content/exchanges Family medicine 360 http://www.vdgm.woncaeurope.org Sante Sud http://www.santesud.org/ Global health fellowship https://gprecruitment.hee.nhs.uk/Recruitment/GHF McGill university International exchange https://www.mcgill.ca/familymed/global-health/projects/studentexchange Erasmus plus exchanges http://www.rcgp.org.uk/jic General International opportunities http://www.rcgp.org.uk/rcgp-near-you/rcgp-international/international-opportunities.aspx https://www.cfhi.org View Large Table 2. List of current opportunities for international exchanges Exchange organisation Website address Hippokrates exchange http://vdgm.woncaeurope.org/content/exchanges Family medicine 360 http://www.vdgm.woncaeurope.org Sante Sud http://www.santesud.org/ Global health fellowship https://gprecruitment.hee.nhs.uk/Recruitment/GHF McGill university International exchange https://www.mcgill.ca/familymed/global-health/projects/studentexchange Erasmus plus exchanges http://www.rcgp.org.uk/jic General International opportunities http://www.rcgp.org.uk/rcgp-near-you/rcgp-international/international-opportunities.aspx https://www.cfhi.org Exchange organisation Website address Hippokrates exchange http://vdgm.woncaeurope.org/content/exchanges Family medicine 360 http://www.vdgm.woncaeurope.org Sante Sud http://www.santesud.org/ Global health fellowship https://gprecruitment.hee.nhs.uk/Recruitment/GHF McGill university International exchange https://www.mcgill.ca/familymed/global-health/projects/studentexchange Erasmus plus exchanges http://www.rcgp.org.uk/jic General International opportunities http://www.rcgp.org.uk/rcgp-near-you/rcgp-international/international-opportunities.aspx https://www.cfhi.org View Large Declaration Funding: BB and CH are supported by the NIHR as Academic Clinical Fellows. Eugene Tang is supported by a NIHR Doctoral Research Fellowship (DRF-2015-08-006). Ethical approval: none. Conflict of interest: none. Acknowledgements To the Junior International Committee (JIC) a subcommittee of the RCGP who provided pastoral support. References 1. RCGP, NHS Health Education England. General practice forward view . https://www.england.nhs.uk/wp-content/uploads/2016/04/gpfv.pdf (Accessed on 7 April 2017 ). 2. Such E , Walton E , Delaney B , Harris J , Salway S . New Migrants in Primary Healthcare – How are services adapting? Summary and Mini Case Book . http://scharr.dept.shef.ac.uk/healthequity/wp-content/uploads/sites/4/2016/07/New-Migrants-in-Primary-Healthcare-Summary-FINAL5.pdf (accessed on 7 April 2017 ). 3. Lindenmeyer A , Redwood S , Griffith L , Teladia Z , Phillimore J . Experiences of primary care professionals providing healthcare to recently arrived migrants: a qualitative study . BMJ Open 2016 ; 6 : e012561 . Google Scholar CrossRef Search ADS PubMed 4. Gupta AR , Wells CK , Horwitz RI , Bia FJ , Barry M . The International Health Program: the fifteen-year experience with Yale University’s Internal Medicine Residency Program . Am J Trop Med Hyg 1999 ; 61 : 1019 – 23 . Google Scholar CrossRef Search ADS PubMed 5. Crisp N . Global Health Partnerships: The UK Contribution to Health in Developing Countries . 2007 . www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_065374 (accessed on 7 April 2017 ). 6. Barata AN , Rigon S . Family medicine 360°: Global exchanges in family medicine . J Family Med Prim Care 2015 ; 4 : 305 – 9 . Google Scholar CrossRef Search ADS PubMed 7. van Weel C , Mattsson B , Freeman GK , de Meyere M , von Fragstein M ; EU Socrates Programme ‘Primary Health Care’ . General practice based teaching exchanges in Europe . Eur J Gen Pract 2005 ; 11 : 122 – 6 . Google Scholar CrossRef Search ADS PubMed 8. Popay J , Roberts H , Sowden A et al. Guidance on the conduct of narrative synthesis in systematic reviews . ESRC Methods Programme 2006 ; 15 : 47 – 71 . 9. Pettigrew L , Seo HN . Tips for GP trainees wishing to undertake an international experience . Br J Gen Pract 2012 ; 62 : 221 – 2 . Google Scholar CrossRef Search ADS PubMed 10. Dowling S , Rouse M , Farrell J et al. Experiences of Irish general practice trainees who participant in 4-month rotations in Malawi and Australia . Ed Prim Car 2015 ; 26 : 388 – 94 . Google Scholar CrossRef Search ADS 11. Niemantsvertsverdriet S , van der Vletuen C et al. An explorative study into learning on international traineeships: experiential. Experiential learning processes dominate . Med Educ 2005 ; 39 : 1236 – 42 . Google Scholar CrossRef Search ADS PubMed 12. Milne A , Cowie J . Promoting culturally competent care: the Erasmus exchange programme . Nurs Stand 2013 ; 27 : 42 – 6 . Google Scholar CrossRef Search ADS PubMed 13. Pavlic D . Erasmus exchange in the field of family medicine in Slovenia . Acta Med Acad 2012 ; 41 : 47 – 51 . Google Scholar CrossRef Search ADS PubMed 14. Jelley D . A UK-Portugal general practice exchange programme . EJGP 2002 ; 8 : 75 – 6 . 15. Sweeny GP . Canadian view of British general practice . Br Med J 1971 ; 1 : 337 – 8 . Google Scholar CrossRef Search ADS PubMed 16. Rhodes CD , Marsh SG . Anglo-Australian exchange in general practice . J R Coll Gen Pract 1979 ; 29 : 302 – 3 . Google Scholar PubMed 17. Gilin B , Young T . Educational benefits of international experiential learning in an MSW program . Am J Med 1986 ; 80 : 1019 – 21 . Google Scholar CrossRef Search ADS PubMed 18. Munir S , Holliste C . International out of programme experience (OOPE): is it worthwhile ? Ed Prim Car 2013 ; 24 : 303 – 7 . Google Scholar CrossRef Search ADS 19. Tooke J . Aspiring to Excellence. Final Report of the Independent Inquiry into Modernising Medical Careers 2008 . www.mmcinquiry.org.uk/Final_8_Jan_08_MMC_all.pdf (accessed on 7 April 2017 ). 20. Watson J , Attridge M , Begg S . Primary care training: a global perspective . Innovait 2012 ; 5 : 472 – 82 . 21. Ullrich S , Briel D , Nesterko Y , Hiemisch A , Brähler E , Glaesmer H . Communication with migrant patients and their parents in inpatient general pediatric care . Gesundheitswesen 2016 ; 78 : 209 – 14 . Google Scholar CrossRef Search ADS PubMed 22. Cheng IH , Drillich A , Schattner P . Refugee experiences of general practice in countries of resettlement: a literature review . Br J Gen Pract 2015 ; 65 : e171 – 6 . Google Scholar CrossRef Search ADS PubMed 23. Brewer TF , Saba N , Clair V . From boutique to basic: a call for standardised medical education in global health . Med Educ 2009 ; 43 : 930 – 3 . Google Scholar CrossRef Search ADS PubMed 24. Evert J , Bazemore A , Hixon A , Withy K . Going global: considerations for introducing global health into family medicine training programs . Fam Med 2007 ; 39 : 659 – 65 . Google Scholar PubMed 25. Hansel B , Grove N . International student exchange programs are the educational benefits real ? NASSP Bull 1984; 70 : 84 – 90 . 26. Gupta AR , Wells CK , Horwitz RI , Bia FJ , Barry M . The International Health Program: the fifteen-year experience with Yale University’s Internal Medicine Residency Program . Am J Trop Med Hyg 1999 ; 61 : 1019 – 23 . Google Scholar CrossRef Search ADS PubMed 27. Frenk J , Chen L , Bhutta ZA et al. Health professionals for a new century: transforming education to strengthen health systems in an interdependent world . Lancet 2010 ; 376 : 1923 – 58 . Google Scholar CrossRef Search ADS PubMed 28. Metcalfe N , Frances P . Learning from a GP exchange to France . Ed Prim Car 2015 ; 26 : 356 – 7 . Google Scholar CrossRef Search ADS 29. Heyward M . From international to intercultural redefining the international school for a globalized world . J Res Int Educ 2002 ; 1 : 9 – 32 . 30. United Nations, Department of Economic and Social Affairs, Population Division (2015). Trends in International Migrant Stock: The 2015 Revision. (United Nations database, POP/DB/MIG/Stock/Rev.2015). 31. van Olmen J , Criel B , Devadasan N et al. Primary Health Care in the 21st century: primary care providers and people’s empowerment . Trop Med Int Health 2010 ; 15 : 386 – 90 . Google Scholar PubMed 32. van der Horst F , Lemmens P . Medical education and professionalism across different cultures . Med Educ 2012 ; 46 : 238 – 9 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Lai, Shih-Han;Tsai, Yi-Wen;Chen, Yi-Chuan;Chang, Shy-Shin

2017 Family Practice

doi: 10.1093/fampra/cmx110pmid: 29092063

Abstract Background Obesity is associated with increased risk of cardiovascular disease and chronic kidney disease (CKD). Hyperhomocysteinaemia refers to increased oxidative stress and has been associated with the risk of CKD. Objectives We investigated the association among body mass index (BMI), homocysteine level and impaired renal function in a Taiwanese adult population. Methods This was a retrospective cross-sectional study involving 24826 subjects who underwent a health check-up from January 2013 to December 2015. A multivariate linear regression model was developed to analyse the relationship among BMI, serum homocysteine and estimated glomerular filtration rate (eGFR). A multivariate logistic regression model was used to assess the relationship among weight categories, hyperhomocysteinaemia and CKD. Results The prevalence of CKD in the quartile groups of homocysteine were 2.5%, 2.7%, 3.4% and 5.2% (P < 0.01). For every one-unit increase in BMI (kg/m2), the eGFR decreased by 0.50 ml/min/1.73 m2. Overweight/obese subjects with high homocysteine levels had a higher odds ratio (OR) for CKD, as compared with normal weight subjects (1.84 versus 1.38, respectively; P < 0.01 versus P = 0.02, respectively). Overweight/obese female subjects with hyperhomocysteinaemia had an OR of 3.40 [P < 0.01; 95% confidence interval (CI): 2.06–5.61] for CKD; in males, the OR was 1.66 (P < 0.01; 95% CI: 1.38–1.99). Conclusions Patients who are overweight/obese with higher homocysteine levels have an increased risk of CKD, especially females. Additional studies exploring whether the effect of weight loss or homocysteine-lowering therapies such as folic acid, vitamin B12 supplements that may prevent or slow the progression of declining renal function, is warranted. Body mass index, chronic, glomerular filtration rate, homocysteine, inflammation, obesity, renal insufficiency Introduction Chronic kidney disease (CKD) poses a significant challenge in 21st century global health policy because of its emerging health and economic burden (1). Primary health care is important for CKD because of high global prevalence (11–13%) of stage 3 underdiagnosed and undertreated patients with CKD (2). In Taiwan, the prevalence of CKD is as high as 11.93%, reaching about 2.74 million patients, potentially resulting in a significant mortality rate because of cardiovascular diseases (CVD), and accounting for tremendous health care expenditures (3). Therefore, early identification of the risk factors for CKD is critical for preventing the development of kidney damage and adverse outcomes. At present, >2.1 billion people are overweight or obese worldwide. As overweight and obesity are the fifth leading cause of death worldwide, accounting for nearly 3.4 million deaths annually (4), the increasing prevalence should be considered in primary care setting. Obesity is strongly associated with diabetes and hypertension and was demonstrated as a risk factor for the development of CKD (5). The crucial pathogenic role of obesity-induced chronic renal disease may be related to excess nutrients in metabolic cells, leading to the activation of several bioactive mediators (6) and an increase in the endogenous production of proinflammatory cytokines (7). Homocysteine is an amino acid formed by the conversion of methionine to cysteine, and an elevated plasma homocysteine level has been recognized as an independent factor for CVD (8). Homocysteine is believed to impair implantation by interfering with oxidative injury to vascular endothelial cells, and their vascular integrity may contribute to intrarenal arteriosclerosis, along with a subsequent reduction in renal perfusion pressure (9) and eventually reduced estimated glomerular filtration rate (eGFR). Preliminary investigations suggest that elevated homocysteine levels may be a risk factor or risk marker of future CKD (10,11). However, the association between weight status and different plasma homocysteine levels in CKD is not yet well established. The purpose of this study was to determine the relationship between body mass index (BMI) and serum homocysteine levels in CKD and to evaluate other associated risk factors. Methods Subjects This cross-sectional study involved subjects aged ≥18 years who underwent annual heath check-ups at the Linkou (northern Taiwan) and Chiayi (southern Taiwan) branches of Chang Gung Memorial Hospital from January 2013 to December 2015. All subjects enrolled were factory workers from northern and southern Taiwan who participated in the annual health check-ups at Chang Gung Memorial Hospital Linkou or Chiayi branch. The general characteristics of the population were age ranging from 30 to 55 years and being predominantly male. Each subject was invited to answer the questionnaires regarding his or her personal and past medical history. Trained nurses provided assistance while the participants were answering the questionnaire during the health examination. The original number of participants was 29728, and 3828 of them were excluded because of incomplete answers to the questionnaire. Among the remaining participants, 152 pregnant women and 922 participants who reported with underlying chronic diseases that might alter the metabolic state or kidney function tests, such as thyroid or hypothalamic diseases, adrenal disease, renal cancer, glomerulonephritis, renal failure on haemodialysis or peritoneal dialysis, liver cirrhosis, or use of diuretic renal replacement therapy were also excluded. The remaining 24826 participants were enrolled in this study. Informed consent was obtained from all participants. The Institutional Review Board of Chang Gung Memorial Hospital approved this study. Data collection Trained nurses took the anthropometric measurements for all participants in accordance with standard operating procedures. The questionnaires consisted of two main parts: the survey of personal and past medical history. Questions on personal history included smoking habits, alcohol drinking history, betel nut chewing history and pregnancy status (if female). Questions on past medical history included chronic diseases, medication and operation history. Blood pressure (BP) was measured with an automatic sphygmomanometer and repeated two to three times after at least 10 min of rest when subjects had BP measurements higher than 120/80 mmHg (Welch Allyn, Skaneateles Falls, NY, USA; based on yearly calibrations). Height and weight were measured using an automatic scale with a sensitivity of 0.1 kg and a resolution of 0.1 cm. BMI was calculated as a ratio between weight and height in metre squared (kg/m2). Waist circumference was measured by two trained examiners using a measuring tape placed horizontally around the subjects’ abdomen at the midpoint between the lower border of the rib cage and the upper iliac crest. Biochemical measurements Venous blood samples were collected in vacuum tubes by venipuncture in the morning after a 12-h fast; the samples were stored at 4°C in a refrigerator before analysis by the hospital laboratory department. All blood analyses were done in the Clinical Laboratory Department of Linkuo or Chiayi Chang Gung Memorial Hospital; both laboratories are certified by the College of American Pathologists. Urine specimens were obtained in the morning and scheduled to avoid menstrual periods. Laboratory measurements included high-sensitivity C reactive protein (hsCRP), fasting plasma glucose (FPG), total cholesterol (TChol), triglyceride (TG) and high-density lipoprotein cholesterol (HDL-C) levels. Creatinine (Cr) and hsCRP levels were measured with a Hitachi 7600 Modular Chemistry Analyzer (Hitachi, Tokyo, Japan). FPG was measured using the hexokinase method. TChol and TG levels were measured using an enzymatic colorimetric test. HDL-C was measured using a selective-inhibition method. Definition of measurement cut-offs and calculations BMI categories were defined as follows: obesity, ≥27 kg/m2; overweight, 24–26.9 kg/m2; and normal weight, <23.9 kg/m2, according to the ranges established for Asian populations by the Ministry of Health and Welfare of Taiwan (12). The cut-off for waist circumference for abdominal obesity was ≥90 cm for men and ≥80 cm for women, using the Asian-specific cut-off points established by the International Diabetes Federation (13). The eGFR was calculated using equations for the Modification of Diet in Renal Disease for Chinese patients, with CKD (14) measured in the following manner: 175 × (Scr)–1.234 × (Age)–0.179 × 0.79 (if female). CKD was defined as an eGFR of <60 ml/min per 1.73 m2 of body surface (ml/min/1.73 m2), according to the definition from the Kidney Disease Outcomes Quality Initiative (K/DOQI) (15) for CKD ≥ stage 3. A high homocysteine level was defined using the upper quartile for the serum homocysteine level. The upper homocysteine quartile in our study population was 11.81 µmol/l. Diagnostic criteria for metabolic syndrome (MetS) were established according to the 2004 Taiwan Ministry of Health criteria that were adapted from the Asian modification of the US National Cholesterol Education Program criteria (16). A diagnosis of MetS required three or more of the following criteria: (a) high BP (systolic BP ≥130 mmHg, diastolic BP ≥85 mmHg); (b) high serum TG (TG ≥ 150 mg/dl); (c) increased HDL-C (<40 mg/dl for males and <50 mg/dl for females); (d) hyperglycaemia (FPG ≥ 100 mg/dl); and (e) abdominal obesity (using modified waist circumference cut-offs for Asian populations). Statistical analysis Continuous variables are presented as median (interquartile range). Categorical variables are shown as count and percentage. Differences for categorical variables between normal weight and overweight/obese groups were examined using the chi-square test. Participants were classified into quartiles according to their serum homocysteine levels. The four independent homocysteine groups were statistically analysed using a Kruskal–Wallis test with a post hoc Bonferroni correction for repeated comparisons. A multivariate linear regression model was established to study the association of renal function (eGFR), BMI and serum homocysteine levels with age, gender, hypertension and diabetes. Finally, we established two multivariate logistic regression models: one was developed to evaluate associations between different combined weight groups and serum homocysteine levels with the risk of CKD after adjusting for age, gender, smoking status, hypertension, diabetes mellitus and hyperlipidaemia. The other was developed to investigate the presence of hyperhomocysteinaemia and the risk of CKD among different weight groups based on the total population and sex with completely separate analyses. SPSS software package, version 20.0 (IBM Corporation, Chicago, USA), was used for statistical analysis. All statistic assessments were evaluated using a two-sided α level of 0.05. Results Baseline characteristics among different weight groups A total of 24826 subjects were enrolled in this study. All participants were divided into two groups based on their BMI: one with normal weight subjects (BMI <24 kg/m2; n = 12093, 48.71%) and the other with overweight/obese subjects (BMI ≥24 kg/m2; n = 12733, 51.29%). Among all participants, the medium age was 37 (33, 41) years and 44 (38, 51) years in the normal weight and overweight/obese groups, respectively. There were 19076 (76.8%) males and 5750 (23.2%) females in this study. Among the males, 77% were overweight or obese, and 23% of the females were overweight or obese. There were significant differences in demographic and cardiometabolic risk factors between the two groups (Table 1). Compared with the normal weight group, the overweight/obese group was older and had higher body fat percentages (%BF), waist circumferences, systolic/diastolic BP, TChol, LDL, TG, Chol-T/HDL, FPG, uric acid, Cr, hsCRP and homocysteine levels. In addition, the prevalence of smoking, proteinuria, hypertension, diabetes mellitus, hyperlipidaemia and CKD (eGFR ≥ 60 ml/min/1.73 m2 and/or proteinuria of ≥1+) was higher in the overweight/obese group than in the normal weight group. The percentages of CKD were 2.1% and 4.8% [odds ratio (OR): 2.38; 95% CI: 2.05–2.76; P < 0.01] in the normal weight and overweight/obese groups, respectively. Table 1. Baseline characteristics of study subjects aged ≥18 years who underwent annual health check-ups during 2013–2015 based on BMI groups (N = 24826) Characteristics Normal weight; BMI <24 (n = 12093) Overweight/obesity; BMI ≥24 (n = 12733) P value Age 37 (33, 41) 44 (38, 51) <0.01* Gender (n, %) 0.61  Male 9275 (76.7) 9801 (77.0)  Female 2818 (23.3) 2932 (23.0) Smoking (n, %) <0.01*  Non-smokers 9280 (76.7) 9093 (71.4)  Past smokers 730 (6.0) 1192 (9.4)  Current smokers 2083 (17.2) 2448 (19.2) BMI (kg/m2) 22.46 (20.78, 23.95) 26.77 (25.0, 28.82) <0.01* Body fat percentage (%) 30.18 (25.10, 32.31) 36.41 (34.34, 39.09) <0.01* Waist circumference (cm) (n, %) 77.5 (71.0, 82.0) 88.0 (82.0, 94.0) <0.01* Normal 11700 (96.8) 6454 (50.7) Abnormal (male ≥90, female ≥80) 393 (3.2) 6279 (49.3) SBP (mmHg) 117 (108, 126) 127 (118, 136) <0.01* DBP (mmHg) 73 (67, 79) 80 (73, 87) <0.01* Total cholesterol (mg/dl) 180 (161, 202) 193 (172, 214) <0.01* LDL cholesterol(mg/dl) 113 (95, 133) 125 (106, 145) <0.01* Triglycerides (mg/dl) 84 (62, 119) 120 (86, 173) <0.01* HDL cholesterol (mg/dl) 51 (44, 59) 46 (40, 53) <0.01* Chol-T/HDL 3.46 (2.90, 4.17) 4.19 (3.50, 4.97) <0.01* Fasting glucose (mg/dl) 85 (81, 90) 90 (85, 97) <0.01* Creatinine (mg/dl) 0.83 (0.71, 0.93) 0.86 (0.73, 0.98) <0.01* eGFR (ml/min/1.73 m2) 112.8 (98.93, 130.7) 104.8 (90.02, 121.93) <0.01* CKD (n, %) <0.01*  eGFR ≥60 11844 (97.9) 12127 (95.2)  eGFR ＜60 or proteinuria ≥1+ 249 (2.1) 606 (4.8) Proteinuria (n, %) <0.01*  Absent 11862 (98.1) 12219 (96)  Present 231 (1.9) 514 (4.0) Uric acid (mg/dl) 6.0 (5.1, 6.9) 6.6 (5.5, 7.5) <0.01* Homocysteine (µmol/l) 9.82 (8.30, 9.82) 10.1 (8.60, 11.98) <0.01* hsCRP (μg/ml) 0.76 (0.43, 1.47) 1.4 (0.77, 2.62) <0.01* Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 11984 (99.1) 11648 (91.5)  Present 109 (0.9) 1085 (8.5)  Diabetes mellitus (n, %) <0.01*  Absent 12062 (99.7) 12456 (97.8)  Present 31 (0.3) 277 (2.2)  Hyperlipidaemia (n, %) <0.01*  Absent 12058 (99.7) 12513 (98.3)  Present 35 (0.3) 220 (1.7) Characteristics Normal weight; BMI <24 (n = 12093) Overweight/obesity; BMI ≥24 (n = 12733) P value Age 37 (33, 41) 44 (38, 51) <0.01* Gender (n, %) 0.61  Male 9275 (76.7) 9801 (77.0)  Female 2818 (23.3) 2932 (23.0) Smoking (n, %) <0.01*  Non-smokers 9280 (76.7) 9093 (71.4)  Past smokers 730 (6.0) 1192 (9.4)  Current smokers 2083 (17.2) 2448 (19.2) BMI (kg/m2) 22.46 (20.78, 23.95) 26.77 (25.0, 28.82) <0.01* Body fat percentage (%) 30.18 (25.10, 32.31) 36.41 (34.34, 39.09) <0.01* Waist circumference (cm) (n, %) 77.5 (71.0, 82.0) 88.0 (82.0, 94.0) <0.01* Normal 11700 (96.8) 6454 (50.7) Abnormal (male ≥90, female ≥80) 393 (3.2) 6279 (49.3) SBP (mmHg) 117 (108, 126) 127 (118, 136) <0.01* DBP (mmHg) 73 (67, 79) 80 (73, 87) <0.01* Total cholesterol (mg/dl) 180 (161, 202) 193 (172, 214) <0.01* LDL cholesterol(mg/dl) 113 (95, 133) 125 (106, 145) <0.01* Triglycerides (mg/dl) 84 (62, 119) 120 (86, 173) <0.01* HDL cholesterol (mg/dl) 51 (44, 59) 46 (40, 53) <0.01* Chol-T/HDL 3.46 (2.90, 4.17) 4.19 (3.50, 4.97) <0.01* Fasting glucose (mg/dl) 85 (81, 90) 90 (85, 97) <0.01* Creatinine (mg/dl) 0.83 (0.71, 0.93) 0.86 (0.73, 0.98) <0.01* eGFR (ml/min/1.73 m2) 112.8 (98.93, 130.7) 104.8 (90.02, 121.93) <0.01* CKD (n, %) <0.01*  eGFR ≥60 11844 (97.9) 12127 (95.2)  eGFR ＜60 or proteinuria ≥1+ 249 (2.1) 606 (4.8) Proteinuria (n, %) <0.01*  Absent 11862 (98.1) 12219 (96)  Present 231 (1.9) 514 (4.0) Uric acid (mg/dl) 6.0 (5.1, 6.9) 6.6 (5.5, 7.5) <0.01* Homocysteine (µmol/l) 9.82 (8.30, 9.82) 10.1 (8.60, 11.98) <0.01* hsCRP (μg/ml) 0.76 (0.43, 1.47) 1.4 (0.77, 2.62) <0.01* Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 11984 (99.1) 11648 (91.5)  Present 109 (0.9) 1085 (8.5)  Diabetes mellitus (n, %) <0.01*  Absent 12062 (99.7) 12456 (97.8)  Present 31 (0.3) 277 (2.2)  Hyperlipidaemia (n, %) <0.01*  Absent 12058 (99.7) 12513 (98.3)  Present 35 (0.3) 220 (1.7) Continuous data are reported as median (interquartile range) for non-normal distribution data and compared using the Mann–Whitney U test of non-parametric analysis; categorical data are shown as number (percentage) and compared using the chi-square test. Asterisk indicates a significant difference between BMI < 24 kg/m2 and BMI ≥ 24 kg/m2. BMI, body mass index; SBP, systolic blood pressure; DPB, diastolic blood pressure; LDL, low-density lipoprotein lipase cholesterol; HDL, high-density lipoprotein lipase cholesterol; Chol-T, total cholesterol; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; hsCRP, high sensitivity C reactive protein. View Large Table 1. Baseline characteristics of study subjects aged ≥18 years who underwent annual health check-ups during 2013–2015 based on BMI groups (N = 24826) Characteristics Normal weight; BMI <24 (n = 12093) Overweight/obesity; BMI ≥24 (n = 12733) P value Age 37 (33, 41) 44 (38, 51) <0.01* Gender (n, %) 0.61  Male 9275 (76.7) 9801 (77.0)  Female 2818 (23.3) 2932 (23.0) Smoking (n, %) <0.01*  Non-smokers 9280 (76.7) 9093 (71.4)  Past smokers 730 (6.0) 1192 (9.4)  Current smokers 2083 (17.2) 2448 (19.2) BMI (kg/m2) 22.46 (20.78, 23.95) 26.77 (25.0, 28.82) <0.01* Body fat percentage (%) 30.18 (25.10, 32.31) 36.41 (34.34, 39.09) <0.01* Waist circumference (cm) (n, %) 77.5 (71.0, 82.0) 88.0 (82.0, 94.0) <0.01* Normal 11700 (96.8) 6454 (50.7) Abnormal (male ≥90, female ≥80) 393 (3.2) 6279 (49.3) SBP (mmHg) 117 (108, 126) 127 (118, 136) <0.01* DBP (mmHg) 73 (67, 79) 80 (73, 87) <0.01* Total cholesterol (mg/dl) 180 (161, 202) 193 (172, 214) <0.01* LDL cholesterol(mg/dl) 113 (95, 133) 125 (106, 145) <0.01* Triglycerides (mg/dl) 84 (62, 119) 120 (86, 173) <0.01* HDL cholesterol (mg/dl) 51 (44, 59) 46 (40, 53) <0.01* Chol-T/HDL 3.46 (2.90, 4.17) 4.19 (3.50, 4.97) <0.01* Fasting glucose (mg/dl) 85 (81, 90) 90 (85, 97) <0.01* Creatinine (mg/dl) 0.83 (0.71, 0.93) 0.86 (0.73, 0.98) <0.01* eGFR (ml/min/1.73 m2) 112.8 (98.93, 130.7) 104.8 (90.02, 121.93) <0.01* CKD (n, %) <0.01*  eGFR ≥60 11844 (97.9) 12127 (95.2)  eGFR ＜60 or proteinuria ≥1+ 249 (2.1) 606 (4.8) Proteinuria (n, %) <0.01*  Absent 11862 (98.1) 12219 (96)  Present 231 (1.9) 514 (4.0) Uric acid (mg/dl) 6.0 (5.1, 6.9) 6.6 (5.5, 7.5) <0.01* Homocysteine (µmol/l) 9.82 (8.30, 9.82) 10.1 (8.60, 11.98) <0.01* hsCRP (μg/ml) 0.76 (0.43, 1.47) 1.4 (0.77, 2.62) <0.01* Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 11984 (99.1) 11648 (91.5)  Present 109 (0.9) 1085 (8.5)  Diabetes mellitus (n, %) <0.01*  Absent 12062 (99.7) 12456 (97.8)  Present 31 (0.3) 277 (2.2)  Hyperlipidaemia (n, %) <0.01*  Absent 12058 (99.7) 12513 (98.3)  Present 35 (0.3) 220 (1.7) Characteristics Normal weight; BMI <24 (n = 12093) Overweight/obesity; BMI ≥24 (n = 12733) P value Age 37 (33, 41) 44 (38, 51) <0.01* Gender (n, %) 0.61  Male 9275 (76.7) 9801 (77.0)  Female 2818 (23.3) 2932 (23.0) Smoking (n, %) <0.01*  Non-smokers 9280 (76.7) 9093 (71.4)  Past smokers 730 (6.0) 1192 (9.4)  Current smokers 2083 (17.2) 2448 (19.2) BMI (kg/m2) 22.46 (20.78, 23.95) 26.77 (25.0, 28.82) <0.01* Body fat percentage (%) 30.18 (25.10, 32.31) 36.41 (34.34, 39.09) <0.01* Waist circumference (cm) (n, %) 77.5 (71.0, 82.0) 88.0 (82.0, 94.0) <0.01* Normal 11700 (96.8) 6454 (50.7) Abnormal (male ≥90, female ≥80) 393 (3.2) 6279 (49.3) SBP (mmHg) 117 (108, 126) 127 (118, 136) <0.01* DBP (mmHg) 73 (67, 79) 80 (73, 87) <0.01* Total cholesterol (mg/dl) 180 (161, 202) 193 (172, 214) <0.01* LDL cholesterol(mg/dl) 113 (95, 133) 125 (106, 145) <0.01* Triglycerides (mg/dl) 84 (62, 119) 120 (86, 173) <0.01* HDL cholesterol (mg/dl) 51 (44, 59) 46 (40, 53) <0.01* Chol-T/HDL 3.46 (2.90, 4.17) 4.19 (3.50, 4.97) <0.01* Fasting glucose (mg/dl) 85 (81, 90) 90 (85, 97) <0.01* Creatinine (mg/dl) 0.83 (0.71, 0.93) 0.86 (0.73, 0.98) <0.01* eGFR (ml/min/1.73 m2) 112.8 (98.93, 130.7) 104.8 (90.02, 121.93) <0.01* CKD (n, %) <0.01*  eGFR ≥60 11844 (97.9) 12127 (95.2)  eGFR ＜60 or proteinuria ≥1+ 249 (2.1) 606 (4.8) Proteinuria (n, %) <0.01*  Absent 11862 (98.1) 12219 (96)  Present 231 (1.9) 514 (4.0) Uric acid (mg/dl) 6.0 (5.1, 6.9) 6.6 (5.5, 7.5) <0.01* Homocysteine (µmol/l) 9.82 (8.30, 9.82) 10.1 (8.60, 11.98) <0.01* hsCRP (μg/ml) 0.76 (0.43, 1.47) 1.4 (0.77, 2.62) <0.01* Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 11984 (99.1) 11648 (91.5)  Present 109 (0.9) 1085 (8.5)  Diabetes mellitus (n, %) <0.01*  Absent 12062 (99.7) 12456 (97.8)  Present 31 (0.3) 277 (2.2)  Hyperlipidaemia (n, %) <0.01*  Absent 12058 (99.7) 12513 (98.3)  Present 35 (0.3) 220 (1.7) Continuous data are reported as median (interquartile range) for non-normal distribution data and compared using the Mann–Whitney U test of non-parametric analysis; categorical data are shown as number (percentage) and compared using the chi-square test. Asterisk indicates a significant difference between BMI < 24 kg/m2 and BMI ≥ 24 kg/m2. BMI, body mass index; SBP, systolic blood pressure; DPB, diastolic blood pressure; LDL, low-density lipoprotein lipase cholesterol; HDL, high-density lipoprotein lipase cholesterol; Chol-T, total cholesterol; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; hsCRP, high sensitivity C reactive protein. View Large Different characteristics among homocysteine subgroup quartiles Table 2 shows the different characteristics according to the quartiles of serum homocysteine levels. Significant differences were observed in most characteristics among low-to-high homocysteine subgroups, except for diabetes mellitus (0.9%, 1.1%, 1.4% and 1.6%, the first to fourth quartiles, respectively). Subjects with higher plasma homocysteine levels were more likely to have a lower eGFR. The eGFRs in the first, second, third and fourth quartiles were 124.4, 109.9, 105.1 and 99.1 ml/min/1.73 m2, respectively. The prevalence of smoking, hypertension, hyperlipidaemia, proteinuria and CKD (2.5%, 2.7%, 3.4% and 5.2% for the respective quartiles) increased with increasing plasma homocysteine levels. The medians for BMI, %BF, waist circumference, systolic BP, diastolic BP, TG, Chol/HDL, Cr and uric acid levels all significantly increased with increasing homocysteine levels (all P values <0.01); however, the HDL levels decreased with increasing homocysteine levels, and all post hoc analyses with a Bonferroni correction reached significance between the two groups (all P values <0.01). However, no significant difference was identified in TChol, LDL, FPG and hsCRP between the third and fourth quartiles. Table 2. Characteristics represented across quartiles of homocysteine (N = 24826) by participants aged ≥18 years who underwent annual health check-ups during 2013–2015 Homocysteine quartiles Group 1; 0–8.50 µmol/l (n = 6485) Group 2; 8.51–10.00 µmol/l (n = 6140) Group 3; 10.01–11.80 µmol/l (n = 6189) Group 4; >11.81 µmol/l (n = 6012) P value Age 40 (35, 46） 40 (35, 46) 40 (35, 47) 40 (35, 47) <0.01*bc Gender (n, %) <0.01*  Male 3065 (47.3) 4830 (78.7) 5504 (88.9) 5677 (94.4)  Female 3420 (52.7) 1310 (21.3) 685 (11.1) 335 (5.6) BMI (kg/m2) 23.37 (21.0, 25.8) 24.4 (22.2, 26.7) 24.8 (22.8, 27.1) 25.1 (23.1, 27.5) <0.01*abcdef Normal weight (BMI <24) 3412 (52.6) 3062 (49.9) 2940 (47.5) 2679 (48.7) <0.01* Overweight/obese (BMI ≥24) 3073 (47.4) 3078 (50.1) 3249 (52.5) 3333 (55.4) Body fat percentage (%) 27.3 (18.9, 33.8) 32.6 (28.0, 36.3) 33.6 (30.2, 37.0) 34.4 (31.1, 37.7) <0.01*abcedf Waist circumference (cm) (n, %) 78.0 (70.0, 85.0) 82.0 (76, 89) 84 (78, 90) 85.0 (79.0, 91.5) <0.01*abcdef  Normal 5075 (78.3) 4587 (74.7) 4443 (71.8) 4049 (67.3)  Abnormal (male ≥90, female ≥80) 1410 (21.7) 1553 (25.3) 1746 (28.2) 1963 (32.7) <0.01* SBP (mmHg) 116 (106, 128) 122 (112, 131) 119 (111, 129) 126 (117, 135) <0.01*abcdef DBP (mmHg) 73 (66, 81) 76 (69, 83) 77 (71, 85) 79 (72, 86) <0.01*abcdef Total cholesterol (mg/dl) 182 (163, 204) 187 (166, 108) 189 (169, 211) 188 (168, 211) <0.01*abcde LDL cholesterol (mg/dl) 114 (95, 134) 119 (100, 139) 123 (103, 143) 122 (103, 143) <0.01*abcde Triglycerides (mg/dl) 88 (63, 128) 102 (71, 149) 106 (75, 154) 110 (78, 160) <0.01*abcdef HDL cholesterol (mg/dl) 52 (44, 60) 48 (42, 56) 47 (41, 55) 46 (40, 54) <0.01*abcdef Chol-T/HDL 3.47 (2.89, 4.23) 3.83 (3.16, 4.58) 4.0 (3.3, 4.7) 4.07 (3.37, 4.85) <0.01*abcdef Fasting glucose (mg/dl) 86 (81, 92) 87 (83, 94) 88 (83, 94) 88 (83, 95) <0.01*abcde Smoking (n, %) <0.01*  Non-smokers 5357 (82.6) 4508 (73.4) 4340 (70.1) 4168 (69.3)  Past smokers 367 (5.7) 493 (8.0) 572 (9.2) 490 (8.2)  Current smokers 761 (11.7) 1139 (18.6) 1277 (20.6) 1354 (22.5) Creatinine (mg/dl) 0.70 (0.57, 0.85) 0.84 (0.72, 0.93) 0.88 (0.78, 0.98) 0.92 (0.83, 1.03) <0.01*abcdef eGFR (ml/min/1.73 m2) 124.4 (106.3, 146.3) 109.9 (96.9, 125.9) 105.1 (92.1, 119.6) 99.1 (86.2, 113.07) <0.01*abcdef CKD (n, %) <0.01*  eGFR ≥60 6321 (97.5) 5975 (97.3) 5977 (96.6) 5698 (94.8)  eGFR＜60 or proteinuria ≥1+ 164 (2.5) 165 (2.7) 212 (3.4) 314 (5.2) Proteinuria (n, %) <0.01*  Absent 6326 (97.5) 5986 (97.5) 5992 (96.8) 5777 (96.1)  Present 159 (2.5) 154 (2.5) 197 (3.2) 235 (3.9) Uric acid (mg/dl) 5.4 (4.5, 6.4) 6.2 (5.3, 7.1) 6.6 (5.7, 7.5) 6.9 (6.0, 7.8) <0.01*abcdef Homocysteine (µmol/l) 7.5 (6.7, 8.1) 9.3 (8.9, 9.7) 10.83 (10.4, 11.3) 13.3 (12.5, 15.0) <0.01*abcdef hsCRP (μg/ml) 0.94 (0.49, 1.95) 1.03 (0.53,2.06) 1.09 (0.6, 2.08) 1.14 (0.62, 2.18) <0.01*abcde Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 6271 (96.7) 5895 (96) 5868 (94.8) 5598 (93.1)  Present 214 (3.3) 245 (4.0) 321 (5.2) 414 (6.9)  Diabetes mellitus (n, %) <0.01  Absent 6426 (99.1) 6074 (98.9) 6100 (98.6) 5918 (98.4)  Present 59 (0.9) 66 (1.1) 89 (1.4) 94 (1.6)  Hyperlipidaemia (n, %) <0.01*  Absent 6441 (99.3) 6090 (99.3) 6116 (98.8) 5918 (98.4)  Present 44 (0.7) 44 (0.7) 73 (1.2) 94 (1.6) Homocysteine quartiles Group 1; 0–8.50 µmol/l (n = 6485) Group 2; 8.51–10.00 µmol/l (n = 6140) Group 3; 10.01–11.80 µmol/l (n = 6189) Group 4; >11.81 µmol/l (n = 6012) P value Age 40 (35, 46） 40 (35, 46) 40 (35, 47) 40 (35, 47) <0.01*bc Gender (n, %) <0.01*  Male 3065 (47.3) 4830 (78.7) 5504 (88.9) 5677 (94.4)  Female 3420 (52.7) 1310 (21.3) 685 (11.1) 335 (5.6) BMI (kg/m2) 23.37 (21.0, 25.8) 24.4 (22.2, 26.7) 24.8 (22.8, 27.1) 25.1 (23.1, 27.5) <0.01*abcdef Normal weight (BMI <24) 3412 (52.6) 3062 (49.9) 2940 (47.5) 2679 (48.7) <0.01* Overweight/obese (BMI ≥24) 3073 (47.4) 3078 (50.1) 3249 (52.5) 3333 (55.4) Body fat percentage (%) 27.3 (18.9, 33.8) 32.6 (28.0, 36.3) 33.6 (30.2, 37.0) 34.4 (31.1, 37.7) <0.01*abcedf Waist circumference (cm) (n, %) 78.0 (70.0, 85.0) 82.0 (76, 89) 84 (78, 90) 85.0 (79.0, 91.5) <0.01*abcdef  Normal 5075 (78.3) 4587 (74.7) 4443 (71.8) 4049 (67.3)  Abnormal (male ≥90, female ≥80) 1410 (21.7) 1553 (25.3) 1746 (28.2) 1963 (32.7) <0.01* SBP (mmHg) 116 (106, 128) 122 (112, 131) 119 (111, 129) 126 (117, 135) <0.01*abcdef DBP (mmHg) 73 (66, 81) 76 (69, 83) 77 (71, 85) 79 (72, 86) <0.01*abcdef Total cholesterol (mg/dl) 182 (163, 204) 187 (166, 108) 189 (169, 211) 188 (168, 211) <0.01*abcde LDL cholesterol (mg/dl) 114 (95, 134) 119 (100, 139) 123 (103, 143) 122 (103, 143) <0.01*abcde Triglycerides (mg/dl) 88 (63, 128) 102 (71, 149) 106 (75, 154) 110 (78, 160) <0.01*abcdef HDL cholesterol (mg/dl) 52 (44, 60) 48 (42, 56) 47 (41, 55) 46 (40, 54) <0.01*abcdef Chol-T/HDL 3.47 (2.89, 4.23) 3.83 (3.16, 4.58) 4.0 (3.3, 4.7) 4.07 (3.37, 4.85) <0.01*abcdef Fasting glucose (mg/dl) 86 (81, 92) 87 (83, 94) 88 (83, 94) 88 (83, 95) <0.01*abcde Smoking (n, %) <0.01*  Non-smokers 5357 (82.6) 4508 (73.4) 4340 (70.1) 4168 (69.3)  Past smokers 367 (5.7) 493 (8.0) 572 (9.2) 490 (8.2)  Current smokers 761 (11.7) 1139 (18.6) 1277 (20.6) 1354 (22.5) Creatinine (mg/dl) 0.70 (0.57, 0.85) 0.84 (0.72, 0.93) 0.88 (0.78, 0.98) 0.92 (0.83, 1.03) <0.01*abcdef eGFR (ml/min/1.73 m2) 124.4 (106.3, 146.3) 109.9 (96.9, 125.9) 105.1 (92.1, 119.6) 99.1 (86.2, 113.07) <0.01*abcdef CKD (n, %) <0.01*  eGFR ≥60 6321 (97.5) 5975 (97.3) 5977 (96.6) 5698 (94.8)  eGFR＜60 or proteinuria ≥1+ 164 (2.5) 165 (2.7) 212 (3.4) 314 (5.2) Proteinuria (n, %) <0.01*  Absent 6326 (97.5) 5986 (97.5) 5992 (96.8) 5777 (96.1)  Present 159 (2.5) 154 (2.5) 197 (3.2) 235 (3.9) Uric acid (mg/dl) 5.4 (4.5, 6.4) 6.2 (5.3, 7.1) 6.6 (5.7, 7.5) 6.9 (6.0, 7.8) <0.01*abcdef Homocysteine (µmol/l) 7.5 (6.7, 8.1) 9.3 (8.9, 9.7) 10.83 (10.4, 11.3) 13.3 (12.5, 15.0) <0.01*abcdef hsCRP (μg/ml) 0.94 (0.49, 1.95) 1.03 (0.53,2.06) 1.09 (0.6, 2.08) 1.14 (0.62, 2.18) <0.01*abcde Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 6271 (96.7) 5895 (96) 5868 (94.8) 5598 (93.1)  Present 214 (3.3) 245 (4.0) 321 (5.2) 414 (6.9)  Diabetes mellitus (n, %) <0.01  Absent 6426 (99.1) 6074 (98.9) 6100 (98.6) 5918 (98.4)  Present 59 (0.9) 66 (1.1) 89 (1.4) 94 (1.6)  Hyperlipidaemia (n, %) <0.01*  Absent 6441 (99.3) 6090 (99.3) 6116 (98.8) 5918 (98.4)  Present 44 (0.7) 44 (0.7) 73 (1.2) 94 (1.6) Continuous data are reported as median (interquartile range) for non-normal distribution data and compared using the Kruskal–Wallis test for post hoc analysis; categorical data are shown as number (percentage) and compared using the chi-square test. Asterisk indicates a statistical significance among the quartiles of homocysteine. Group 1 indicated as homocysteine level lower than 8.50 µmol/l; Group 2 indicated as homocysteine level between 8.50 and 10.00 µmol/l; Group 3 indicated as homocysteine level between 10.01 and 11.80 µmol/l; Group 4 indicated as homocysteine level higher than 11.81 µmol/l. BMI, body mass index; SBP, systolic blood pressure; DPB, diastolic blood pressure; LDL, low-density lipoprotein lipase cholesterol; HDL, high-density lipoprotein lipase cholesterol; Chol-T, total cholesterol; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; hsCRP, high sensitivity C reactive protein. aSignificant differences between G1 and G2. bSignificant differences between G1 and G3. cSignificant differences between G1 and G4. dSignificant differences between G2 and G3. eSignificant differences between G2 and G4. fSignificant differences between G3 and G4. View Large Table 2. Characteristics represented across quartiles of homocysteine (N = 24826) by participants aged ≥18 years who underwent annual health check-ups during 2013–2015 Homocysteine quartiles Group 1; 0–8.50 µmol/l (n = 6485) Group 2; 8.51–10.00 µmol/l (n = 6140) Group 3; 10.01–11.80 µmol/l (n = 6189) Group 4; >11.81 µmol/l (n = 6012) P value Age 40 (35, 46） 40 (35, 46) 40 (35, 47) 40 (35, 47) <0.01*bc Gender (n, %) <0.01*  Male 3065 (47.3) 4830 (78.7) 5504 (88.9) 5677 (94.4)  Female 3420 (52.7) 1310 (21.3) 685 (11.1) 335 (5.6) BMI (kg/m2) 23.37 (21.0, 25.8) 24.4 (22.2, 26.7) 24.8 (22.8, 27.1) 25.1 (23.1, 27.5) <0.01*abcdef Normal weight (BMI <24) 3412 (52.6) 3062 (49.9) 2940 (47.5) 2679 (48.7) <0.01* Overweight/obese (BMI ≥24) 3073 (47.4) 3078 (50.1) 3249 (52.5) 3333 (55.4) Body fat percentage (%) 27.3 (18.9, 33.8) 32.6 (28.0, 36.3) 33.6 (30.2, 37.0) 34.4 (31.1, 37.7) <0.01*abcedf Waist circumference (cm) (n, %) 78.0 (70.0, 85.0) 82.0 (76, 89) 84 (78, 90) 85.0 (79.0, 91.5) <0.01*abcdef  Normal 5075 (78.3) 4587 (74.7) 4443 (71.8) 4049 (67.3)  Abnormal (male ≥90, female ≥80) 1410 (21.7) 1553 (25.3) 1746 (28.2) 1963 (32.7) <0.01* SBP (mmHg) 116 (106, 128) 122 (112, 131) 119 (111, 129) 126 (117, 135) <0.01*abcdef DBP (mmHg) 73 (66, 81) 76 (69, 83) 77 (71, 85) 79 (72, 86) <0.01*abcdef Total cholesterol (mg/dl) 182 (163, 204) 187 (166, 108) 189 (169, 211) 188 (168, 211) <0.01*abcde LDL cholesterol (mg/dl) 114 (95, 134) 119 (100, 139) 123 (103, 143) 122 (103, 143) <0.01*abcde Triglycerides (mg/dl) 88 (63, 128) 102 (71, 149) 106 (75, 154) 110 (78, 160) <0.01*abcdef HDL cholesterol (mg/dl) 52 (44, 60) 48 (42, 56) 47 (41, 55) 46 (40, 54) <0.01*abcdef Chol-T/HDL 3.47 (2.89, 4.23) 3.83 (3.16, 4.58) 4.0 (3.3, 4.7) 4.07 (3.37, 4.85) <0.01*abcdef Fasting glucose (mg/dl) 86 (81, 92) 87 (83, 94) 88 (83, 94) 88 (83, 95) <0.01*abcde Smoking (n, %) <0.01*  Non-smokers 5357 (82.6) 4508 (73.4) 4340 (70.1) 4168 (69.3)  Past smokers 367 (5.7) 493 (8.0) 572 (9.2) 490 (8.2)  Current smokers 761 (11.7) 1139 (18.6) 1277 (20.6) 1354 (22.5) Creatinine (mg/dl) 0.70 (0.57, 0.85) 0.84 (0.72, 0.93) 0.88 (0.78, 0.98) 0.92 (0.83, 1.03) <0.01*abcdef eGFR (ml/min/1.73 m2) 124.4 (106.3, 146.3) 109.9 (96.9, 125.9) 105.1 (92.1, 119.6) 99.1 (86.2, 113.07) <0.01*abcdef CKD (n, %) <0.01*  eGFR ≥60 6321 (97.5) 5975 (97.3) 5977 (96.6) 5698 (94.8)  eGFR＜60 or proteinuria ≥1+ 164 (2.5) 165 (2.7) 212 (3.4) 314 (5.2) Proteinuria (n, %) <0.01*  Absent 6326 (97.5) 5986 (97.5) 5992 (96.8) 5777 (96.1)  Present 159 (2.5) 154 (2.5) 197 (3.2) 235 (3.9) Uric acid (mg/dl) 5.4 (4.5, 6.4) 6.2 (5.3, 7.1) 6.6 (5.7, 7.5) 6.9 (6.0, 7.8) <0.01*abcdef Homocysteine (µmol/l) 7.5 (6.7, 8.1) 9.3 (8.9, 9.7) 10.83 (10.4, 11.3) 13.3 (12.5, 15.0) <0.01*abcdef hsCRP (μg/ml) 0.94 (0.49, 1.95) 1.03 (0.53,2.06) 1.09 (0.6, 2.08) 1.14 (0.62, 2.18) <0.01*abcde Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 6271 (96.7) 5895 (96) 5868 (94.8) 5598 (93.1)  Present 214 (3.3) 245 (4.0) 321 (5.2) 414 (6.9)  Diabetes mellitus (n, %) <0.01  Absent 6426 (99.1) 6074 (98.9) 6100 (98.6) 5918 (98.4)  Present 59 (0.9) 66 (1.1) 89 (1.4) 94 (1.6)  Hyperlipidaemia (n, %) <0.01*  Absent 6441 (99.3) 6090 (99.3) 6116 (98.8) 5918 (98.4)  Present 44 (0.7) 44 (0.7) 73 (1.2) 94 (1.6) Homocysteine quartiles Group 1; 0–8.50 µmol/l (n = 6485) Group 2; 8.51–10.00 µmol/l (n = 6140) Group 3; 10.01–11.80 µmol/l (n = 6189) Group 4; >11.81 µmol/l (n = 6012) P value Age 40 (35, 46） 40 (35, 46) 40 (35, 47) 40 (35, 47) <0.01*bc Gender (n, %) <0.01*  Male 3065 (47.3) 4830 (78.7) 5504 (88.9) 5677 (94.4)  Female 3420 (52.7) 1310 (21.3) 685 (11.1) 335 (5.6) BMI (kg/m2) 23.37 (21.0, 25.8) 24.4 (22.2, 26.7) 24.8 (22.8, 27.1) 25.1 (23.1, 27.5) <0.01*abcdef Normal weight (BMI <24) 3412 (52.6) 3062 (49.9) 2940 (47.5) 2679 (48.7) <0.01* Overweight/obese (BMI ≥24) 3073 (47.4) 3078 (50.1) 3249 (52.5) 3333 (55.4) Body fat percentage (%) 27.3 (18.9, 33.8) 32.6 (28.0, 36.3) 33.6 (30.2, 37.0) 34.4 (31.1, 37.7) <0.01*abcedf Waist circumference (cm) (n, %) 78.0 (70.0, 85.0) 82.0 (76, 89) 84 (78, 90) 85.0 (79.0, 91.5) <0.01*abcdef  Normal 5075 (78.3) 4587 (74.7) 4443 (71.8) 4049 (67.3)  Abnormal (male ≥90, female ≥80) 1410 (21.7) 1553 (25.3) 1746 (28.2) 1963 (32.7) <0.01* SBP (mmHg) 116 (106, 128) 122 (112, 131) 119 (111, 129) 126 (117, 135) <0.01*abcdef DBP (mmHg) 73 (66, 81) 76 (69, 83) 77 (71, 85) 79 (72, 86) <0.01*abcdef Total cholesterol (mg/dl) 182 (163, 204) 187 (166, 108) 189 (169, 211) 188 (168, 211) <0.01*abcde LDL cholesterol (mg/dl) 114 (95, 134) 119 (100, 139) 123 (103, 143) 122 (103, 143) <0.01*abcde Triglycerides (mg/dl) 88 (63, 128) 102 (71, 149) 106 (75, 154) 110 (78, 160) <0.01*abcdef HDL cholesterol (mg/dl) 52 (44, 60) 48 (42, 56) 47 (41, 55) 46 (40, 54) <0.01*abcdef Chol-T/HDL 3.47 (2.89, 4.23) 3.83 (3.16, 4.58) 4.0 (3.3, 4.7) 4.07 (3.37, 4.85) <0.01*abcdef Fasting glucose (mg/dl) 86 (81, 92) 87 (83, 94) 88 (83, 94) 88 (83, 95) <0.01*abcde Smoking (n, %) <0.01*  Non-smokers 5357 (82.6) 4508 (73.4) 4340 (70.1) 4168 (69.3)  Past smokers 367 (5.7) 493 (8.0) 572 (9.2) 490 (8.2)  Current smokers 761 (11.7) 1139 (18.6) 1277 (20.6) 1354 (22.5) Creatinine (mg/dl) 0.70 (0.57, 0.85) 0.84 (0.72, 0.93) 0.88 (0.78, 0.98) 0.92 (0.83, 1.03) <0.01*abcdef eGFR (ml/min/1.73 m2) 124.4 (106.3, 146.3) 109.9 (96.9, 125.9) 105.1 (92.1, 119.6) 99.1 (86.2, 113.07) <0.01*abcdef CKD (n, %) <0.01*  eGFR ≥60 6321 (97.5) 5975 (97.3) 5977 (96.6) 5698 (94.8)  eGFR＜60 or proteinuria ≥1+ 164 (2.5) 165 (2.7) 212 (3.4) 314 (5.2) Proteinuria (n, %) <0.01*  Absent 6326 (97.5) 5986 (97.5) 5992 (96.8) 5777 (96.1)  Present 159 (2.5) 154 (2.5) 197 (3.2) 235 (3.9) Uric acid (mg/dl) 5.4 (4.5, 6.4) 6.2 (5.3, 7.1) 6.6 (5.7, 7.5) 6.9 (6.0, 7.8) <0.01*abcdef Homocysteine (µmol/l) 7.5 (6.7, 8.1) 9.3 (8.9, 9.7) 10.83 (10.4, 11.3) 13.3 (12.5, 15.0) <0.01*abcdef hsCRP (μg/ml) 0.94 (0.49, 1.95) 1.03 (0.53,2.06) 1.09 (0.6, 2.08) 1.14 (0.62, 2.18) <0.01*abcde Past medical history of systemic diseases  Hypertension (n, %) <0.01*  Absent 6271 (96.7) 5895 (96) 5868 (94.8) 5598 (93.1)  Present 214 (3.3) 245 (4.0) 321 (5.2) 414 (6.9)  Diabetes mellitus (n, %) <0.01  Absent 6426 (99.1) 6074 (98.9) 6100 (98.6) 5918 (98.4)  Present 59 (0.9) 66 (1.1) 89 (1.4) 94 (1.6)  Hyperlipidaemia (n, %) <0.01*  Absent 6441 (99.3) 6090 (99.3) 6116 (98.8) 5918 (98.4)  Present 44 (0.7) 44 (0.7) 73 (1.2) 94 (1.6) Continuous data are reported as median (interquartile range) for non-normal distribution data and compared using the Kruskal–Wallis test for post hoc analysis; categorical data are shown as number (percentage) and compared using the chi-square test. Asterisk indicates a statistical significance among the quartiles of homocysteine. Group 1 indicated as homocysteine level lower than 8.50 µmol/l; Group 2 indicated as homocysteine level between 8.50 and 10.00 µmol/l; Group 3 indicated as homocysteine level between 10.01 and 11.80 µmol/l; Group 4 indicated as homocysteine level higher than 11.81 µmol/l. BMI, body mass index; SBP, systolic blood pressure; DPB, diastolic blood pressure; LDL, low-density lipoprotein lipase cholesterol; HDL, high-density lipoprotein lipase cholesterol; Chol-T, total cholesterol; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; hsCRP, high sensitivity C reactive protein. aSignificant differences between G1 and G2. bSignificant differences between G1 and G3. cSignificant differences between G1 and G4. dSignificant differences between G2 and G3. eSignificant differences between G2 and G4. fSignificant differences between G3 and G4. View Large Association analysis among BMI, homocysteine and eGFR As shown in Table 3, both the BMI and plasma homocysteine levels were negatively associated with eGFR. A significant difference was detected in the multiple linear regression model for evaluating the association among BMI, homocysteine and eGFR after adjusting for gender, age, smoking, hypertension, diabetes and hyperlipidaemia. For a one-unit (kg/m2) increase in BMI, there was a 0.50 ml/min/1.73 m2 decline in eGFR. For every unit (µmol/l) increase in homocysteine, the eGFR decreased by 1.10 ml/min/1.73 m2. Table 3. Multivariate linear regression model estimating the association among BMI, homocysteine and eGFR (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 β (95% Confidence interval) P value Gender (male) –28.79 (–29.53, –28.06) <0.01 Age –0.65 (–0.68, –0.62) <0.01 Hypertension –2.56 (–3.97, –1.16) <0.01 Diabetes 6.16 (3.53, 8.78) <0.01 BMI –0.50 (–0.58, –0.42) <0.01 Homocysteine –1.10 (–1.18, –1.11) <0.01 β (95% Confidence interval) P value Gender (male) –28.79 (–29.53, –28.06) <0.01 Age –0.65 (–0.68, –0.62) <0.01 Hypertension –2.56 (–3.97, –1.16) <0.01 Diabetes 6.16 (3.53, 8.78) <0.01 BMI –0.50 (–0.58, –0.42) <0.01 Homocysteine –1.10 (–1.18, –1.11) <0.01 The model of association analysis among BMI, homocysteine and eGFR was adjusted by gender, age, smoking, hypertension (self-reported history of hypertension or taking anti-hypertensive medication), diabetes (self-reported history of diabetes or taking anti-diabetic drugs), hyperlipidaemia (self-reported history of hyperlipidaemia or taking lipid-lowering medication), body mass index and homocysteine using stepwise multivariate linear regression analysis; constant = 185.78; F = 2000.15; r2 = 0.33; P = <0.01. BMI, body mass index; eGFR, estimated glomerular filtration rate. View Large Table 3. Multivariate linear regression model estimating the association among BMI, homocysteine and eGFR (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 β (95% Confidence interval) P value Gender (male) –28.79 (–29.53, –28.06) <0.01 Age –0.65 (–0.68, –0.62) <0.01 Hypertension –2.56 (–3.97, –1.16) <0.01 Diabetes 6.16 (3.53, 8.78) <0.01 BMI –0.50 (–0.58, –0.42) <0.01 Homocysteine –1.10 (–1.18, –1.11) <0.01 β (95% Confidence interval) P value Gender (male) –28.79 (–29.53, –28.06) <0.01 Age –0.65 (–0.68, –0.62) <0.01 Hypertension –2.56 (–3.97, –1.16) <0.01 Diabetes 6.16 (3.53, 8.78) <0.01 BMI –0.50 (–0.58, –0.42) <0.01 Homocysteine –1.10 (–1.18, –1.11) <0.01 The model of association analysis among BMI, homocysteine and eGFR was adjusted by gender, age, smoking, hypertension (self-reported history of hypertension or taking anti-hypertensive medication), diabetes (self-reported history of diabetes or taking anti-diabetic drugs), hyperlipidaemia (self-reported history of hyperlipidaemia or taking lipid-lowering medication), body mass index and homocysteine using stepwise multivariate linear regression analysis; constant = 185.78; F = 2000.15; r2 = 0.33; P = <0.01. BMI, body mass index; eGFR, estimated glomerular filtration rate. View Large Table 4 shows that both the BMI and plasma homocysteine levels were significantly related to an increased risk of CKD after adjusting for gender, age, smoking, hypertension, diabetes and hyperlipidaemia. As for overweight/obese (BMI ≥24 kg/m2), the OR for risk of CKD was 1.94 (95% CI: 1.66–2.27, P < 0.01). The OR for hyperhomocysteinaemia (>11.81 µmol/l) for CKD was 1.70 (95% CI: 1.47–1.97, P < 0.01). Table 4. Multivariate linear regression model estimating the association among BMI, homocysteine and CKD (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 OR (95% confidence interval) P value Age 1.02 (1.01, 1.02) <0.01 Hypertension 2.42 (1.96, 3.00) <0.01 Diabetes 3.12 (2.25, 4.34) <0.01 BMI 1.11 (1.10, 1.11) <0.01 Homocysteine 1.05 (1.03, 1.06) <0.01 OR (95% confidence interval) P value Age 1.02 (1.01, 1.02) <0.01 Hypertension 2.42 (1.96, 3.00) <0.01 Diabetes 3.12 (2.25, 4.34) <0.01 BMI 1.11 (1.10, 1.11) <0.01 Homocysteine 1.05 (1.03, 1.06) <0.01 The model of association analysis among BMI, homocysteine and eGFR was adjusted by gender, age, smoking, hypertension (self-reported history of hypertension or taking anti-hypertensive medication), diabetes (self-reported history of diabetes or taking anti-diabetic drugs), hyperlipidaemia (self-reported history of hyperlipidaemia or taking lipid-lowering medication), body mass index and homocysteine using stepwise multivariate logistic regression analysis; constant = 185.78; F = 2000.154; r2 = 0.33; P = <0.01. BMI, body mass index; eGFR, estimated glomerular filtration rate. View Large Table 4. Multivariate linear regression model estimating the association among BMI, homocysteine and CKD (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 OR (95% confidence interval) P value Age 1.02 (1.01, 1.02) <0.01 Hypertension 2.42 (1.96, 3.00) <0.01 Diabetes 3.12 (2.25, 4.34) <0.01 BMI 1.11 (1.10, 1.11) <0.01 Homocysteine 1.05 (1.03, 1.06) <0.01 OR (95% confidence interval) P value Age 1.02 (1.01, 1.02) <0.01 Hypertension 2.42 (1.96, 3.00) <0.01 Diabetes 3.12 (2.25, 4.34) <0.01 BMI 1.11 (1.10, 1.11) <0.01 Homocysteine 1.05 (1.03, 1.06) <0.01 The model of association analysis among BMI, homocysteine and eGFR was adjusted by gender, age, smoking, hypertension (self-reported history of hypertension or taking anti-hypertensive medication), diabetes (self-reported history of diabetes or taking anti-diabetic drugs), hyperlipidaemia (self-reported history of hyperlipidaemia or taking lipid-lowering medication), body mass index and homocysteine using stepwise multivariate logistic regression analysis; constant = 185.78; F = 2000.154; r2 = 0.33; P = <0.01. BMI, body mass index; eGFR, estimated glomerular filtration rate. View Large Hyperhomocysteinaemia and risk of CKD among different weight groups Table 5 shows the association of hyperhomocysteinaemia and risk of CKD based on different weights and gender. For the total population, the OR for CKD in hyperhomocysteinaemia for the normal weight group was 1.38 (95% CI: 1.05–1.83, P = 0.02) and 1.84 for the overweight/obesity group (95% CI: 1.56–2.19, P < 0.01). Similar patterns and trends were observed in both genders. For males, the OR for CKD with hyperhomocysteinaemia for the normal weight and overweight/obese groups was 1.44 (95% CI: 1.06–1.95, P = 0.02) and 1.66 (95% CI: 1.38–1.99, P < 0.01), respectively; for females, the OR for CKD with hyperhomocysteinaemia for the normal weight and overweight/obese groups was 1.95 (95% CI: 0.77–4.94, P = 0.16) and 3.40 (95% CI: 2.06–5.61, P < 0.01), respectively. Table 5. Hyperhomocysteinaemia and risk of chronic kidney disease among different weight groups by multivariate logistic regression analysis in total population and in stratification of gender (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 Odds ratio P value 95% Confidence interval Total (N = 24826) Normal weight (BMI <24 kg/m2) 1.38* 0.02 1.05–1.83 Overweight/obesity (BMI ≥24 kg/m2) 1.84* <0.01 1.56–2.19 Male (n = 19,076) Normal weight (BMI <24 kg/m2) 1.44* 0.02 1.06–1.95 Overweight/obesity (BMI ≥24 kg/m2) 1.66* <0.01 1.38–1.99 Female (n = 5750) Normal weight (BMI <24 kg/m2) 1.95 0.16 0.77–4.94 Overweight/obesity (BMI ≥24 kg/m2) 3.40* <0.01 2.06–5.61 Odds ratio P value 95% Confidence interval Total (N = 24826) Normal weight (BMI <24 kg/m2) 1.38* 0.02 1.05–1.83 Overweight/obesity (BMI ≥24 kg/m2) 1.84* <0.01 1.56–2.19 Male (n = 19,076) Normal weight (BMI <24 kg/m2) 1.44* 0.02 1.06–1.95 Overweight/obesity (BMI ≥24 kg/m2) 1.66* <0.01 1.38–1.99 Female (n = 5750) Normal weight (BMI <24 kg/m2) 1.95 0.16 0.77–4.94 Overweight/obesity (BMI ≥24 kg/m2) 3.40* <0.01 2.06–5.61 Asterisk indicates significant statistical difference. In normal weight groups, omnibus test for total, male and female model = 0.01, 0.01, 0.00; in overweight/obesity groups, omnibus test for total, male and female model = 0.00, 0.00, 0.00; chi-square test = 43.37; P = 0.00. All models are adjusted for age, gender, smoking status, hypertension, diabetes mellitus and hyperlipidaemia. BMI, body mass index. View Large Table 5. Hyperhomocysteinaemia and risk of chronic kidney disease among different weight groups by multivariate logistic regression analysis in total population and in stratification of gender (N = 24826) of participants aged ≥18 years who underwent annual health check-ups during 2013–2015 Odds ratio P value 95% Confidence interval Total (N = 24826) Normal weight (BMI <24 kg/m2) 1.38* 0.02 1.05–1.83 Overweight/obesity (BMI ≥24 kg/m2) 1.84* <0.01 1.56–2.19 Male (n = 19,076) Normal weight (BMI <24 kg/m2) 1.44* 0.02 1.06–1.95 Overweight/obesity (BMI ≥24 kg/m2) 1.66* <0.01 1.38–1.99 Female (n = 5750) Normal weight (BMI <24 kg/m2) 1.95 0.16 0.77–4.94 Overweight/obesity (BMI ≥24 kg/m2) 3.40* <0.01 2.06–5.61 Odds ratio P value 95% Confidence interval Total (N = 24826) Normal weight (BMI <24 kg/m2) 1.38* 0.02 1.05–1.83 Overweight/obesity (BMI ≥24 kg/m2) 1.84* <0.01 1.56–2.19 Male (n = 19,076) Normal weight (BMI <24 kg/m2) 1.44* 0.02 1.06–1.95 Overweight/obesity (BMI ≥24 kg/m2) 1.66* <0.01 1.38–1.99 Female (n = 5750) Normal weight (BMI <24 kg/m2) 1.95 0.16 0.77–4.94 Overweight/obesity (BMI ≥24 kg/m2) 3.40* <0.01 2.06–5.61 Asterisk indicates significant statistical difference. In normal weight groups, omnibus test for total, male and female model = 0.01, 0.01, 0.00; in overweight/obesity groups, omnibus test for total, male and female model = 0.00, 0.00, 0.00; chi-square test = 43.37; P = 0.00. All models are adjusted for age, gender, smoking status, hypertension, diabetes mellitus and hyperlipidaemia. BMI, body mass index. View Large Discussion To our knowledge, this is the first study to identify the association of weight and serum homocysteine levels with risk of CKD in a Taiwanese adult population. Hyperhomocysteinaemia has been recognized as an independent factor for CVD, deep vein thrombosis and CKD development (8). Elevated homocysteine levels were associated with oxidative injury to vascular endothelial cells and the inhibition of endothelial mediators such as nitric oxide; generation of superoxide radicals that inhibit the relaxation of vessels (17); increased proliferation of vascular smooth muscle cells (9); and decreased production of adenosine, which is thought to be associated with vasodilation and vessel remodelling (18). Meanwhile, in hyperhomocysteinaemic conditions, homocysteine burden leads to the expression of an endoplasmic reticulum stress gene that causes cellular injury in cultured podocytes suggesting a link to kidney damage (19) and eventually leading to focal or global glomerulosclerosis, tubular atrophy, interstitial fibrosis and reduced GFRs (20). Recent evidence has suggested that hormones and cytokines secreted from adipose tissue also contribute to CKD (21). Visceral fats, along with other risk factors in the pathogenesis of MetS, are strongly correlated with insulin resistance. As a result, proatherogenic and inflammatory cytokine production increases; interferes with insulin signalling; and contributes to the development of insulin resistance (22), vascular wall inflammation and CKD (23). Hyperinsulinemia and hyperhomocysteinaemia are now well-accepted risk factors for atherosclerosis. The mechanism of homocysteine angiotoxicity–related kidney injury seems to involve the nitric oxide system by inducing oxidative stress (24). Oxidative stress has been suggested to cause insulin resistance (IR) and may be linked to atherosclerosis. The association between IR and elevated homocysteine levels in healthy, non-obese patients has been proposed, suggesting that IR may contribute to the development of hyperhomocysteinaemia and therefore have implications to premature vascular disease. However, previous study by De Pergola et al. (25) indicated that plasma homocysteine levels are independently associated with IR in apparently healthy normal weight, overweight and obese premenopausal women, thus suggesting a possible role in IR, hyperinsulinemia, or both in increasing plasma homocysteine levels. Consistent estimated high prevalence of CKD and obesity was observed globally. Therefore, identifying patients with CKD at an earlier stage in the primary care setting is vital, so that treatment can be initiated to delay progression and prevent renal failure complications. Our findings indicated that inflammation-related glomerulopathy may be one of the potential causal pathways for IR caused by obesity and hyperhomocysteinaemia (26), and surrogate markers such as BMI and homocysteine may be useful for predicting CKD. Gender differences were also identified by combining BMI and hyperhomocysteinaemia to estimate the risk of CKD among different weight groups based on the multivariate logistic regression analysis of the total population and in different sexes with completely separate analyses (Table 5). For females, the OR for risk of CKD in overweight/obese groups with hyperhomocysteinaemia was greater than that of males (3.40 versus 1.66, respectively), implying that BMI and homocysteine levels may be more influential in females than males. This may be because of different body compositions and fat mass deposition in each gender. With the same BMI level, females usually have greater body fat composition, whereas males have more lean muscle mass (27). Visceral fat contains greater amounts of inflammatory mediators than subcutaneous fat, and these mediators are thought to contribute to the development of IR (28). However, BMI measurements, particularly %BF, are the major determinants of IR in non-diabetic patients with stages 3 and 4 CKD (29). Inflammation-related IR and elevated homocysteine levels secondary to excess body fat at the same BMI levels may have a greater influence in females than males. Prospective studies are needed to define more clearly how different body compositions or fat mass deposition interferes with renal function changes and to determine whether interventions targeting IR or how homocysteine-lowering therapy in this patient population can decrease cardiovascular morbidity and mortality and progression to end-stage renal disease. Nevertheless, certain limitations should be mentioned. First, because of the cross-sectional design of this study, we cannot draw any causal inferences from the data. Continued longitudinal analyses will be aimed at exploring the pathophysiological relationship among overweight/obese, insulin resistance, chronic inflammation, accumulated homocysteine level and subsequent renal function deterioration. Second, it was not a national survey so the results are only representative of healthy adults aged 30–55 years in a Taiwanese population. More prospective data are needed to determine the predictive value of BMI and hyperhomocysteinaemia in CKD development in children as cardiometabolic risk factors begin early in life and continue to adulthood. Moreover, we used the same cut-off value for homocysteine levels for men and women, an assumption that may confound results if a gender difference exists. Further longitudinal research focusing on the relationship between inflammation-related IR and elevated homocysteine levels in CKD patients may help in elucidating the causal relationship and determining whether any interventions, such as vitamins supplementation, modified dietary habits and intensive exercise by changing the body composition or weight loss could prevent CKD progression by lowering the IR or homocysteine level. Conclusion Both BMI and homocysteine levels are independent risk factors for CKD. Our findings may provide the clinical physician a method for early detection of renal function impairment and thus help prevent CKD by identifying overweight or obese patients in the primary care. Moreover, elevated serum homocysteine levels were positively associated with the risk for CKD, and the association was more profound in overweight/obese females than overweight/obese males. Moreover, future clinical research on the effects of weight loss, intensive exercise, lifestyle modification, nutritional status or homocysteine-lowering therapy such as folic acid, vitamin B6, and B12 supplementation, with the aim of unravelling the role of adipose tissue and serum homocysteine levels in associated comorbidities, is highly warranted. Author contribution The authors’ contributions were as follows: SHL and YCC were responsible for data collection. SHL planned the research, analysed the data and prepared the draft of the manuscript. YWT and SSC designed the research, planned the data analysis, interpretation and reviewed the final draft. YWT and SSC were in equal contribution to this work. All authors read and approved the final version of the manuscript. Declaration Funding: None. Ethical approval: Chang Gung Memorial Hospital Institution Review Board (Ethical Review Committee). Ethical approval number: 201600399B0. Conflict of interest: none declared. Acknowledgements Y-WT and S-SC are in equal contribution to this work. The authors would like to thank the staff of the Health Examination Centers in the Keelung, Linkou, and Chiayi branches of Cheng Gung Memorial Hospital for assistance with data collection. References 1. Nugent RA , Fathima SF , Feigl AB et al. The burden of chronic kidney disease on developing nations: a 21st century challenge in global health . Nephron Clin Pract 2011 ; 118 : c269 – 77 . Google Scholar CrossRef Search ADS PubMed 2. Hill NR , Fatoba ST , Oke JL et al. Global prevalence of chronic kidney disease—a systematic review and meta-analysis . PLoS One 2016 ; 11 : e0158765 . Google Scholar CrossRef Search ADS PubMed 3. Wen CP , Cheng TY , Tsai MK et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan . Lancet 2008 ; 371 : 2173 – 82 . Google Scholar CrossRef Search ADS PubMed 4. Smith KB , Smith MS . Obesity statistics . Prim Care 2016 ; 43 : 121 – 35, ix . Google Scholar CrossRef Search ADS PubMed 5. Sabanayagam C , Wong TY , Liao J et al. Body mass index and preclinical kidney disease in Indian adults aged 40 years and above without chronic kidney disease . Clin Exp Nephrol 2014 ; 18 : 919 – 24 . Google Scholar CrossRef Search ADS PubMed 6. Briffa JF , McAinch AJ , Poronnik P et al. Adipokines as a link between obesity and chronic kidney disease . Am J Physiol Renal Physiol 2013 ; 305 : F1629 – 36 . Google Scholar CrossRef Search ADS PubMed 7. Fogo AB . Mechanisms of progression of chronic kidney disease . Pediatr Nephrol 2007 ; 22 : 2011 – 22 . Google Scholar CrossRef Search ADS PubMed 8. Homocysteine Studies C . Homocysteine and risk of ischemic heart disease and stroke: a meta-analysis . JAMA . 2002 ; 288 : 2015 – 22 . CrossRef Search ADS PubMed 9. Haynes WG . Hyperhomocysteinemia, vascular function and atherosclerosis: effects of vitamins . Cardiovasc Drugs Ther 2002 ; 16 : 391 – 9 . Google Scholar CrossRef Search ADS PubMed 10. Chao MC , Hu SL , Hsu HS et al. Serum homocysteine level is positively associated with chronic kidney disease in a Taiwan Chinese population . J Nephrol 2014 ; 27 : 299 – 305 . Google Scholar CrossRef Search ADS PubMed 11. Chen JY , Tsai YW , Chen SY et al. The association of leptin and homocysteine with renal function impairment in a population of Taiwanese adults . Clin Nutr 2015 ; 34 : 943 – 50 . Google Scholar CrossRef Search ADS PubMed 12. Choo V . WHO reassesses appropriate body-mass index for Asian populations . Lancet 2002 ; 360 : 235 . Google Scholar CrossRef Search ADS PubMed 13. Alberti KG , Zimmet P , Shaw J . Metabolic syndrome—a new world-wide definition. A Consensus Statement from the International Diabetes Federation . Diabet Med 2006 ; 23 : 469 – 80 . Google Scholar CrossRef Search ADS PubMed 14. Ma YC , Zuo L , Chen JH et al. Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease . J Am Soc Nephrol 2006 ; 17 : 2937 – 44 . Google Scholar CrossRef Search ADS PubMed 15. National Kidney Foundation . K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification . Am J Kidney Dis 2002 ; 39 ( 2 suppl 1 ): S1 –S 266 . PubMed 16. National Cholesterol Education Program Expert Panel on Detection, Evaluation, Treatment of High Blood Cholesterol in Adults . Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report . Circulation 2002 ; 106 : 3143 – 421 . PubMed 17. Vyssoulis G , Karpanou E , Kyvelou SM et al. Associations between plasma homocysteine levels, aortic stiffness and wave reflection in patients with arterial hypertension, isolated office hypertension and normotensive controls . J Hum Hypertens 2010 ; 24 : 183 – 9 . Google Scholar CrossRef Search ADS PubMed 18. Ho MF , Rose’Meyer RB . Vascular adenosine receptors; potential clinical applications . Curr Vasc Pharmacol 2013 ; 11 : 327 – 37 . Google Scholar CrossRef Search ADS PubMed 19. Inagi R , Nangaku M , Onogi H et al. Involvement of endoplasmic reticulum (ER) stress in podocyte injury induced by excessive protein accumulation . Kidney Int 2005 ; 68 : 2639 – 50 . Google Scholar CrossRef Search ADS PubMed 20. Kubo M , Kiyohara Y , Kato I et al. Risk factors for renal glomerular and vascular changes in an autopsy-based population survey: the Hisayama study . Kidney Int 2003 ; 63 : 1508 – 15 . Google Scholar CrossRef Search ADS PubMed 21. Garland JS . Elevated body mass index as a risk factor for chronic kidney disease: current perspectives . Diabetes Metab Syndr Obes 2014 ; 7 : 347 – 55 . Google Scholar CrossRef Search ADS PubMed 22. Kern PA , Ranganathan S , Li C et al. Adipose tissue tumor necrosis factor and interleukin-6 expression in human obesity and insulin resistance . Am J Physiol Endocrinol Metab 2001 ; 280 : E745 – 51 . Google Scholar CrossRef Search ADS PubMed 23. Thomas G , Sehgal AR , Kashyap SR et al. Metabolic syndrome and kidney disease: a systematic review and meta-analysis . Clin J Am Soc Nephrol 2011 ; 6 : 2364 – 73 . Google Scholar CrossRef Search ADS PubMed 24. Polidori MC , Cherubini A , Senin U et al. Hyperhomocysteinemia and oxidative stress in ischemic stroke . Stroke 2001 ; 32 : 275 – 8 . Google Scholar CrossRef Search ADS PubMed 25. De Pergola G , Pannacciulli N , Zamboni M et al. Homocysteine plasma levels are independently associated with insulin resistance in normal weight, overweight and obese pre-menopausal women . Diabetes Nutr Metab 2001 ; 14 : 253 – 8 . Google Scholar PubMed 26. Atamer A , Demir B , Bayhan G et al. Serum levels of leptin and homocysteine in women with polycystic ovary syndrome and its relationship to endocrine, clinical and metabolic parameters . J Int Med Res 2008 ; 36 : 96 – 105 . Google Scholar CrossRef Search ADS PubMed 27. Tsai YW , Lu MC , Lin YH et al. Combined body mass index with high-sensitivity C-reactive protein as independent predictors for chronic kidney disease in a relatively healthy population in Taiwan . Eur J Clin Nutr 2016 ; 70 : 766 – 71 . Google Scholar CrossRef Search ADS PubMed 28. Fain JN , Madan AK , Hiler ML et al. Comparison of the release of adipokines by adipose tissue, adipose tissue matrix, and adipocytes from visceral and subcutaneous abdominal adipose tissues of obese humans . Endocrinology 2004 ; 145 : 2273 – 82 . Google Scholar CrossRef Search ADS PubMed 29. Trirogoff ML , Shintani A , Himmelfarb J et al. Body mass index and fat mass are the primary correlates of insulin resistance in nondiabetic stage 3-4 chronic kidney disease patients . Am J Clin Nutr 2007 ; 86 : 1642 – 8 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Miquel, Laia;López-Pelayo, Hugo;Nuño, Laura;Arbesú, José Ángel;Zarco, José;Manthey, Jakob;Rehm, Jürgen;Gual, Antoni

2017 Family Practice

doi: 10.1093/fampra/cmx107pmid: 29106526

Abstract Background Alcohol intake and hypertension (HT) are interrelated public health problems with cost-effective interventions at the primary care level that, to date, are poorly implemented. Objective This study aims to explore the barriers to implementing alcohol interventions for people with HT in primary care. Methods As part of the project BASIS (Baseline Alcohol Screening and Intervention Survey), an internet survey from five European countries was developed to determine the role of alcohol in the management of HT in primary care practice. The survey contained 28 core items and 7 country-specific items. We present answers from Spanish general practitioners (GPs), who were reached through the main professional and scientific societies via e-mail and asked to take the online survey. Results In total, 867 GPs answered the survey (65.1% women, 70.4% > 30 years old). As indicated by the Alcohol Use Disorders Identification Test-C scores, 12.4% of GPs who responded were risky drinkers (21.3% of men versus 7.1% of women). GPs reported considering alcohol relatively unimportant in HT treatment, as well as a difficult condition to deal with. The three main barriers to implement screening for alcohol consumption in HT patients were the lack of time (50.0%), considering alcohol unimportant for HT (28.4%) and stigma (16.5%). Conclusions GPs did not consider alcohol consumption a relevant factor for HT and, additionally, found it difficult to deal with alcohol problems. Some of the barriers for alcohol screening could be overcome through structural changes in the health system, such as empowering GPs to treat alcohol use disorders (rather than a single focus on implementing preventive strategies) by enhancing training in alcohol diagnosis and treatment. Alcohol, barriers, hypertension, primary care, screening, treatment Introduction Alcohol is a risk factor for more than 200 health conditions (1). One way in which alcohol affects health is by its impact on blood pressure and hypertension (HT) (2), which are themselves major risk factors for cardiovascular diseases, and therefore related to increased mortality (3). Globally, cardiovascular disease is the largest cause of death (4). In Europe, 30–40% of the European population suffers from HT (5) (defined as blood pressure ≥140/90 (6)), while 15% of European citizens are risky drinkers (7) (defined as a quantity or pattern of alcohol use that places individuals at risk for adverse health and social outcomes). For research and clinical purposes, risky drinking is defined considering quantity and frequency of drinking by using alcohol standard units per day (SDU) or the Alcohol Use Disorders Identification Test (AUDIT)-C which includes information on SDU. Alcohol use and HT are among eight factors (e.g. tobacco use, high body mass index, high cholesterol, hyperglycaemia, poor diet and physical inactivity) which account for 61% of healthy life years lost due to cardiovascular disease. As determinants, alcohol combined with HT represent 1369 years of life lost and 1478 disability-adjusted life years (8). Several risk factors are linked to HT, including older age, gender (young men and women >65 years old), family history of HT, race, overweight, poor diet, lack of physical activity, stress, tobacco and alcohol use (9). The relationship between alcohol consumption and blood pressure is well established. Regular alcohol use increases blood pressure in HT patients under treatment, and alcohol-related HT decreases with abstinence or reductions in alcohol consumption (10). A recent comprehensive meta-analysis has shown that the effect of reducing alcohol intake on improving HT is more significant at higher levels of drinking (11). Moreover, risky drinkers have a 1.5 times greater risk of presenting uncontrolled HT (12)—defined as persistent high blood pressure, despite lifestyle changes and taking diuretic and antihypertensive medications (13)). In Spain, despite the improvement of HT management in the last decade, <25% of hypertensive patients have controlled blood pressure levels (6). In rural areas, the prevalence of HT is high because the level of awareness of high blood pressure is low. Regarding gender, men have lower level of awareness about the risk of having high blood pressure than women, as in other countries (14). In addition, the risk perception of risky alcohol use is low in the Spanish general population, especially among risky drinkers, men in general and young people (15,16). Part of the reason for insufficient levels of blood pressure control is the difficulties in adhering to lifestyle advice (17). It is worth noting that the reduction of raised blood pressure is one of the nine strategic goals identified in the World Health Organization’s ‘Global Action Plan for Prevention and Control of Non-Communicable Diseases (NCDs) 2013–2020’. In Spanish primary care, 17.9% of the population older than 14 years has a diagnosis of HT (18.7% women). For the population older than 40 years, this figure increases to 32%, and for those older than 70 years, it is between 60% and 70% (18). Alcohol, consumed by 77.6% of the Spanish population (19), is a lifestyle factor that should be taken into account in the management of HT, since screening and brief intervention (BI) tools have demonstrated their effectiveness in tackling this problem (20). However, several barriers hamper the implementation of screening, brief intervention and referral to treatment (SBIRT) strategies in primary care—lack of time and resources, insufficient knowledge and skills and negative attitudes towards SBIRT (21). In Europe, research shows that training, support and financial incentives for SBIRT should lead to improved access to this cost-effective strategy (22). In other countries, such as the USA, SBIRT programmes with committed leadership and the use of specialists to deliver the service have been demonstrated to be useful (23). Reducing alcohol consumption in hypertensive patients would improve HT management and reduce pharmacological treatment. Unfortunately, screening and BI for risky drinking are underused in general, and also among hypertensive patients (24). There is a paucity of studies analysing the barriers to identification and treatment of risky drinkers among this specific patient population in routine practice. These patients are a target population for SBIRT because they attend primary care and suffer from an alcohol-related condition, clearly warranting BI. Studies to date which focus on SBIRT barriers have several limitations. They do not analyse in depth why these barriers appear in the first place. They generally do not assess the impact of GPs’ attitude, socio-demographic characteristics, GPs’ drinking patterns, or their training in alcohol management of those barriers. Furthermore, they do not analyse differences between those who regularly implement SBIRT and those who do not, and do not study a specific target population as hypertensive patients. We designed our study to correct these limitations. This study aims to (i) identify GPs’ attitudes and possible barriers in the identification and clinical management of alcohol use among HT patients; and (ii) explore whether any of the following GP’s characteristics were associated with these barriers: sex, age, GP training (graduate and postgraduate), number of patient visits per day and GP’s own pattern of alcohol consumption. These factors represent either commonly cited barriers of alcohol screening among GPs (e.g. lack of training, workload) (25) or personal characteristics (i.e. sex, age and pattern of alcohol consumption), which were hypothesized to impact on the GP’s perception of HT patients’ alcohol consumption. Here, we present data from the Spanish sample of the Baseline Alcohol Screening and Intervention Survey (BASIS) study on alcohol management in hypertensive patients in primary care practices of five European countries (for the European results of BASIS see (24)). Method Data collection A 35-item survey was developed to explore GPs’ attitudes, opinions and perceived barriers towards alcohol management in HT patients attending their practices. A pilot study using an English version of the survey was conducted among 41 health professionals of five different countries. The final versions (translated into French, German, Italian and Spanish) were administered online using SurveyMonkey© (http://www.surveymonkey.com/). Further details about the procedure and assessment of the survey have previously been published (24). The Spanish version of the questionnaire is available in the supplementary material. The following data were obtained from all participating GPs: demographics, alcohol use (AUDIT-C questionnaire (26)) (contained in the survey as the last three questions), number of patient visits per day and number of patients with HT diagnosis per day (for more details, see Supplementary material). Outcome measures GPs were also asked about the three most relevant risk factors for HT and how easy they found it to deal with them (respondents could choose from overweight, salt intake, sleep apnoea, alcohol use, physical activity, smoking and stress). The survey also asked GPs about the perceived capacity of their HT patients to reduce their blood pressure and/or avoid medication through lifestyle changes; and about their own ability and training to deal with risky drinking and alcohol dependence. Those GPs who reported screening for alcohol consumption in <30% of their HT patients were asked about the reasons that deter them from screening. For more details, see the questionnaire in the Supplementary material. Participants Different organizations—Semfyc (Sociedad Española de Medicina de Familia y Comunitaria, https://www.semfyc.es/medicos/) and Semergen (Sociedad Española de Médicos de Atención Primaria, http://www.semergen.es/)—Spanish professional and scientific GP societies, disseminated the survey link to their members via e-mail (n = 20620). Just over 4% (867 GPs) completed the survey, 65.1% being women and 70.4% being older than 30 years. As the study was an anonymous survey, it was exempt from the research ethics committee approval. All respondents were given a brief description of the aims of the study before the actual survey started. Consent to participate was therefore a precondition for taking the survey. Statistical analyses A descriptive analysis of the sample was carried out. Continuous variables were described using means (SD), whereas categorical variables were described using counts and percentages (95% CI). GPs with low alcohol screening rates (<30%) were compared with those who had high alcohol screening rates using the Student’s t-test for continuous variables (number of patients per day, number of patients with HT per day), and the Pearson’s chi-square test or Fisher’s exact test for categorical variables (sex, age >30 years old, alcohol graduate training, alcohol postgraduate training). Those variables with a P value of <0.1 in the univariate analyses were introduced in the logistic regression analysis with GPs with low alcohol screening rate being the dependent variable. To analyse the variables related to each barrier (age, gender, risky drinking, graduate and postgraduate training, number of visits per day), comparisons between screening rate groups were performed using the Student’s t-test for continuous variables and the Pearson’s chi-square test or Fisher’s exact test for categorical variables for each barrier identified. Categorical variables were re-coded into two categories where necessary. Risky drinking was defined as an AUDIT-C score of >4, which has been shown to provide high specificity in both men and women (0.96/0.98)). A P value of <0.05 was required for significance. Bonferroni correction was done, with P = 0.05/6 = 0.0083. All analyses were carried out using the SPSS statistical package (SPSS Inc., version 23.0, Chicago, IL). Results In total, 20620 Spanish GPs were contacted. The response rate was 5.6%, 1146 started the survey and 867 completed it (75.7%). Of the total sample, 65.1% (95% CI: 64.8–66.1%) were women and 70.4% (95% CI: 69.8–71.0%) were >30 years of age. Risky drinking among GPs was estimated at 12.4% (95% CI: 10.2–14.6%): 21.3% (95% CI: 18.6–24.0%) for men and 7.1% (95% CI: 5.4–8.8%) for women, with differences between gender being statistically significant (χ21df = 32.564, P < 0.001). On average, GPs attended 36.8 patients (SD = 18.3), including 13 HT patients per day (SD = 7.9). GPs’ opinions on the relationship between alcohol and HT Of all presented lifestyle factors related to HT, GPs judged alcohol as the least relevant and the second least easy to deal with (stress was the only risk factor considered more difficult to handle than alcohol, see Table 1). Furthermore, GPs estimated that giving advice on lifestyle changes could lead to patients avoiding HT medication in about every fourth case (95% CI: 23.5–28.5%) of a patient with HT diagnosis. More specifically, GPs reported 16% (95% CI: 13.9–18.1%) of HT patients would follow advice to change alcohol intake to avoid HT medication. Table 1. General practitioners’ perception of risk factors’ relevance for HT and easiness to deal with them (n = 867). Data collected between 29 September and 1 December 2015 Relevance (%, 95% CI) Easiness to deal with (%, 95% CI) Overweight/obesity 91.3 (89.4–93.2) 30.1 (27.1–33.2) High salt intake 53.9 (50.6–57.2) 74.5 (71.6–77.4) Smoking 51.4 (48.1–54.7) 20.8 (18.1–23.5) Lack of physical activity 49.8 (46.5–53.1) 46.7 (43.4–50.0) Stress 20.8 (18.1–23.5) 6.1 (4.51–7.69) Sleep apnoea 20.2 (17.5–22.9) 13.1 (10.9–15.4) Alcohol use 12.6 (10.4–14.8) 8.7 (6.8–10.6) Relevance (%, 95% CI) Easiness to deal with (%, 95% CI) Overweight/obesity 91.3 (89.4–93.2) 30.1 (27.1–33.2) High salt intake 53.9 (50.6–57.2) 74.5 (71.6–77.4) Smoking 51.4 (48.1–54.7) 20.8 (18.1–23.5) Lack of physical activity 49.8 (46.5–53.1) 46.7 (43.4–50.0) Stress 20.8 (18.1–23.5) 6.1 (4.51–7.69) Sleep apnoea 20.2 (17.5–22.9) 13.1 (10.9–15.4) Alcohol use 12.6 (10.4–14.8) 8.7 (6.8–10.6) CI, confidence interval. View Large Table 1. General practitioners’ perception of risk factors’ relevance for HT and easiness to deal with them (n = 867). Data collected between 29 September and 1 December 2015 Relevance (%, 95% CI) Easiness to deal with (%, 95% CI) Overweight/obesity 91.3 (89.4–93.2) 30.1 (27.1–33.2) High salt intake 53.9 (50.6–57.2) 74.5 (71.6–77.4) Smoking 51.4 (48.1–54.7) 20.8 (18.1–23.5) Lack of physical activity 49.8 (46.5–53.1) 46.7 (43.4–50.0) Stress 20.8 (18.1–23.5) 6.1 (4.51–7.69) Sleep apnoea 20.2 (17.5–22.9) 13.1 (10.9–15.4) Alcohol use 12.6 (10.4–14.8) 8.7 (6.8–10.6) Relevance (%, 95% CI) Easiness to deal with (%, 95% CI) Overweight/obesity 91.3 (89.4–93.2) 30.1 (27.1–33.2) High salt intake 53.9 (50.6–57.2) 74.5 (71.6–77.4) Smoking 51.4 (48.1–54.7) 20.8 (18.1–23.5) Lack of physical activity 49.8 (46.5–53.1) 46.7 (43.4–50.0) Stress 20.8 (18.1–23.5) 6.1 (4.51–7.69) Sleep apnoea 20.2 (17.5–22.9) 13.1 (10.9–15.4) Alcohol use 12.6 (10.4–14.8) 8.7 (6.8–10.6) CI, confidence interval. View Large Over half of the GPs regarded their graduate training on alcohol management as insufficient (62.5%, 95% CI: 59.7–65.3%) and only 53% (95% CI: 50.1–55.9%) had received some specific postgraduate training on alcohol management. One in four of GPs judged their graduate training on HT as insufficient (76.7%, 95% CI 73.9–79.5%) and 54.9% (95% CI 51.6–58.2%) had received specific postgraduate training on HT. Only 21.9% (95% CI 19.5–24.3%) of Spanish GPs who responded felt capable of dealing with both alcohol dependence and risky drinking, and 61.4% (95% CI: 58.6–64.2%) perceived themselves as capable of dealing only with risky drinking and 14.2% (95% CI: 12.2–16.2%) did not feel competent dealing with either risky drinking or alcohol dependence. Very few of the respondents felt capable of dealing with alcohol dependence but not with risky drinking (2.5%, 95% CI 1.7–3.8%). In total, 83.3% (95% CI 80.6–85.7%) of GPs perceived themselves as capable of dealing with risky drinkers (with or without being able to manage alcohol dependence in patients); meanwhile, 24.4% (95% CI 21.6–27.5%) perceived themselves as capable of dealing with alcohol dependence (irrespective of their ability to manage risky drinking). The description of GPs’ characteristics and the differences between low-screening respondents and high-screening respondents are reported in Table 2. Those GPs seeing the highest number of patients with HT per day (OR = 0.97 95% CI 0.95–0.99), with higher levels of graduate training (OR = 0.65 95% CI 0.47–0.91) and postgraduate training in alcohol (OR = 0.54 95% CI 0.39–0.74), were less likely to be low screening respondents (Table 3). Table 2. GPs’ sample characteristics and differences between low-screening and high-screening general practitionersGPs. Data collected between 29 September and 1 December 2015 All (n = 867), n (%) GPs’ low screeninga ( n = 218), n (%) GPs’ high screening (n = 649), n (%) χ2/Student’s t-test (P value) Gender (women) 564 (65.1) 140 (64.2) 424 (65.3) 0.089 (0.766) Age >30 years 610 (70.4) 149 (68.3) 461 (71.0) 0.564 (0.453) Risky drinking 94 (12.4) 22 (11.8) 72 (12.6) 0.075 (0.785) Appointments per day, mean (SD) 36.8 (18.3) 36.8 (9.1) 36.7 (20.5) –0.049 (0.901) Number of HT patients per day, mean (SD) 13.0 (7.9) 12.1 (6.6) 13.3 (8.3) 1.920 (0.055) Sufficient alcohol graduate training 323 (37.5) 66 (30.4) 257 (39.8) 6.163 (0.013) Sufficient alcohol postgraduate training 457 (53.0) 93 (42.9) 364 (56.4) 12.016 (0.001) All (n = 867), n (%) GPs’ low screeninga ( n = 218), n (%) GPs’ high screening (n = 649), n (%) χ2/Student’s t-test (P value) Gender (women) 564 (65.1) 140 (64.2) 424 (65.3) 0.089 (0.766) Age >30 years 610 (70.4) 149 (68.3) 461 (71.0) 0.564 (0.453) Risky drinking 94 (12.4) 22 (11.8) 72 (12.6) 0.075 (0.785) Appointments per day, mean (SD) 36.8 (18.3) 36.8 (9.1) 36.7 (20.5) –0.049 (0.901) Number of HT patients per day, mean (SD) 13.0 (7.9) 12.1 (6.6) 13.3 (8.3) 1.920 (0.055) Sufficient alcohol graduate training 323 (37.5) 66 (30.4) 257 (39.8) 6.163 (0.013) Sufficient alcohol postgraduate training 457 (53.0) 93 (42.9) 364 (56.4) 12.016 (0.001) HT, hypertensive. aGPs’ low screening: GPs who screen for alcohol ≤3 out of 10 hypertensive patients. View Large Table 2. GPs’ sample characteristics and differences between low-screening and high-screening general practitionersGPs. Data collected between 29 September and 1 December 2015 All (n = 867), n (%) GPs’ low screeninga ( n = 218), n (%) GPs’ high screening (n = 649), n (%) χ2/Student’s t-test (P value) Gender (women) 564 (65.1) 140 (64.2) 424 (65.3) 0.089 (0.766) Age >30 years 610 (70.4) 149 (68.3) 461 (71.0) 0.564 (0.453) Risky drinking 94 (12.4) 22 (11.8) 72 (12.6) 0.075 (0.785) Appointments per day, mean (SD) 36.8 (18.3) 36.8 (9.1) 36.7 (20.5) –0.049 (0.901) Number of HT patients per day, mean (SD) 13.0 (7.9) 12.1 (6.6) 13.3 (8.3) 1.920 (0.055) Sufficient alcohol graduate training 323 (37.5) 66 (30.4) 257 (39.8) 6.163 (0.013) Sufficient alcohol postgraduate training 457 (53.0) 93 (42.9) 364 (56.4) 12.016 (0.001) All (n = 867), n (%) GPs’ low screeninga ( n = 218), n (%) GPs’ high screening (n = 649), n (%) χ2/Student’s t-test (P value) Gender (women) 564 (65.1) 140 (64.2) 424 (65.3) 0.089 (0.766) Age >30 years 610 (70.4) 149 (68.3) 461 (71.0) 0.564 (0.453) Risky drinking 94 (12.4) 22 (11.8) 72 (12.6) 0.075 (0.785) Appointments per day, mean (SD) 36.8 (18.3) 36.8 (9.1) 36.7 (20.5) –0.049 (0.901) Number of HT patients per day, mean (SD) 13.0 (7.9) 12.1 (6.6) 13.3 (8.3) 1.920 (0.055) Sufficient alcohol graduate training 323 (37.5) 66 (30.4) 257 (39.8) 6.163 (0.013) Sufficient alcohol postgraduate training 457 (53.0) 93 (42.9) 364 (56.4) 12.016 (0.001) HT, hypertensive. aGPs’ low screening: GPs who screen for alcohol ≤3 out of 10 hypertensive patients. View Large Table 3. Binary logistic regression results OR 95% CI Number of HT patients per day 0.97 0.95–0.99 Sufficient alcohol graduate traininga 0.65 0.47–0.91 Sufficient alcohol postgraduate trainingb 0.54 0.39–0.74 OR 95% CI Number of HT patients per day 0.97 0.95–0.99 Sufficient alcohol graduate traininga 0.65 0.47–0.91 Sufficient alcohol postgraduate trainingb 0.54 0.39–0.74 Factors are related to GPs having low rates of alcohol screening in patients with HT. Data collected between 29 September and 1 December 2015. aVersus insufficient alcohol graduate training; bVersus insufficient alcohol postgraduate training. View Large Table 3. Binary logistic regression results OR 95% CI Number of HT patients per day 0.97 0.95–0.99 Sufficient alcohol graduate traininga 0.65 0.47–0.91 Sufficient alcohol postgraduate trainingb 0.54 0.39–0.74 OR 95% CI Number of HT patients per day 0.97 0.95–0.99 Sufficient alcohol graduate traininga 0.65 0.47–0.91 Sufficient alcohol postgraduate trainingb 0.54 0.39–0.74 Factors are related to GPs having low rates of alcohol screening in patients with HT. Data collected between 29 September and 1 December 2015. aVersus insufficient alcohol graduate training; bVersus insufficient alcohol postgraduate training. View Large Barriers to implementing screening among low-screening respondents Just under one-fifth of the GPs (19%, 95% CI: 16.7–21.3%) screened for alcohol in, at most, 30% of their HT patients. Half of them reported lack of time as a barrier to screening for alcohol use (50%, 95% CI: 43.4–56.6%), and 28.4% (95% CI: 22.4–34.4%) considered alcohol consumption non-relevant for HT. For importance of all other barriers, see Table 4. We did not find any relevant pattern when analysing how the three most reported barriers were ranked (data not shown). Table 4. Gs’ ranking barriers to screen for alcohol use referred by low-screening GP respondents (n = 218). Data collected between 29 September and 1 December 2015. Barrier % (95% CI) Lack of time 50.0 (16.7–21.3) Alcohol considered unimportant 28.4(22.4–34.4) Stigma 16.5 (11.6–21.4) Patient’s alcohol use previously known 4.1 (1.5–6.7) Considering asking as inappropriate 3.2 (0.9–5.5) Too much effort 2.8 (0.6–5.0) Screening methods unknown 2.3 (0.3–4.3) Lack of training 0.9 (0.4–2.2) Reported other barriers 10.6 (6.5–14.7) Reported no barrier 1.8 (0–3.6) Barrier % (95% CI) Lack of time 50.0 (16.7–21.3) Alcohol considered unimportant 28.4(22.4–34.4) Stigma 16.5 (11.6–21.4) Patient’s alcohol use previously known 4.1 (1.5–6.7) Considering asking as inappropriate 3.2 (0.9–5.5) Too much effort 2.8 (0.6–5.0) Screening methods unknown 2.3 (0.3–4.3) Lack of training 0.9 (0.4–2.2) Reported other barriers 10.6 (6.5–14.7) Reported no barrier 1.8 (0–3.6) View Large Table 4. Gs’ ranking barriers to screen for alcohol use referred by low-screening GP respondents (n = 218). Data collected between 29 September and 1 December 2015. Barrier % (95% CI) Lack of time 50.0 (16.7–21.3) Alcohol considered unimportant 28.4(22.4–34.4) Stigma 16.5 (11.6–21.4) Patient’s alcohol use previously known 4.1 (1.5–6.7) Considering asking as inappropriate 3.2 (0.9–5.5) Too much effort 2.8 (0.6–5.0) Screening methods unknown 2.3 (0.3–4.3) Lack of training 0.9 (0.4–2.2) Reported other barriers 10.6 (6.5–14.7) Reported no barrier 1.8 (0–3.6) Barrier % (95% CI) Lack of time 50.0 (16.7–21.3) Alcohol considered unimportant 28.4(22.4–34.4) Stigma 16.5 (11.6–21.4) Patient’s alcohol use previously known 4.1 (1.5–6.7) Considering asking as inappropriate 3.2 (0.9–5.5) Too much effort 2.8 (0.6–5.0) Screening methods unknown 2.3 (0.3–4.3) Lack of training 0.9 (0.4–2.2) Reported other barriers 10.6 (6.5–14.7) Reported no barrier 1.8 (0–3.6) View Large Associations between GPs’ characteristics (sex, age, own risky drinking, training and workload) and each barrier were examined. Age, sex and risky drinking status were not associated with any of the reported barriers (P > 0.064). Insufficient undergraduate training was linked to higher rates of questioning the appropriateness of asking about alcohol use (7.5% versus 1.3%; χ21df = 5.7, Fisher’s exact test, P = 0.028). Furthermore, GPs with a higher mean number of appointments per day found lack of time to be a barrier (38.2 patients per day versus 35.4 per day; t = 2.3, P = 0.022). However, the observed differences did not remain significant after Bonferroni correction. Discussion In our sample, most of the responding GPs do not view alcohol screening as a priority among their HT patients because of the following reasons:(i) they largely perceive alcohol use as non-relevant for HT and (ii) they consider patients’ risky drinking as a difficult situation to handle. On top of that, they report low rates of perceived success when advising patients to decrease their alcohol intake. GPs have an important role not only in managing but also in preventing HT. This is reflected in a high percentage of GPs recommending most preventive lifestyle measures to their patients (reducing salt intake, healthy diet, physical activity, quit smoking, etc.), in accordance with clinical guidelines for the treatment of high blood pressure (27). However, reducing alcohol intake is the least recommended (28), despite alcohol and high blood pressure having a dose–response relationship (29): the higher the alcohol intake, the higher the risk of HT. Furthermore, risky drinking doubles the risk of HT (2). Giving advice to reduce alcohol consumption can lead to a reduced risk of developing HT, while non-adherence to lifestyle advice is associated with apparent treatment resistance (17). Screening for alcohol consumption and BI have demonstrated effectiveness to reduce alcohol consumption in primary care, even though implementation rates are low (30). Although Spain is one of the European countries where screening for alcohol and intervention are more often delivered (24), in our study, Spanish GPs considered alcohol as the least relevant lifestyle risk factor for HT. These results are in accordance with the fact that less than half of GPs reported a sufficient level of alcohol screening practice in HT patients (24). There is a dramatic difference between the readiness to deal with risky drinkers (83.3%) and the perceived competence to treat alcohol dependence (24.4%), which is a stigmatized mental disorder (31). GPs with the lowest rates of alcohol screening reported that time constraints and considering alcohol unimportant for HT were the main barriers to alcohol screening their HT patients. Time constraints have been identified as a barrier for screening and implementing BI in many studies (21,32,33). Spanish GPs have a high caseload, which is also a problem in many other European countries (34). In Catalonia, GPs have only 10 min per patient for a routine visit, similarly to other European countries (e.g. UK or Germany), despite feeling that they need more time to offer high-quality care. Other barriers which may be stigma related were also reported (i.e. fear of annoying the patient or feeling that having previous knowledge of the patient habits precludes repeating the questions). To foster less stigmatizing and more normalizing attitudes to alcohol screening, it has been proposed to measure AUDs as a continuum as is done with blood pressure (35). AUD is just the extreme end of the continuum of alcohol use. Understanding and talking about alcohol consumption as a continuum could help people to see their consumption as part of this continuum, rather than in discrete polar categories (healthy use/health disorder) (36). In this sense, adopting a definition and discourse of heavy use over time, which is responsible for the neurobiological changes associated with AUD and the substance-attributable burden of disease including social costs, might avoid some of the problems that current conceptualizations have and can help to reduce stigmatization. Barriers to alcohol screening in HT patients, described by GPs with low-screening rates, could be overcome through structural changes in primary care practice and more adequate training. Structural changes including proper reimbursement, sufficient time per patient, electronic health record systems that facilitate administering screening tools and enhanced professional training on alcohol management (graduate and postgraduate) can improve the implementation of screening and BI (37). The importance of professional training is highlighted in our study, where GPs who have received postgraduate education are more likely to screen and manage alcohol use in HT patients (24). However, we should avoid excessive optimism with regard to new strategies to improve BI implementation, as not all of them have been successful. In the Optimizing Delivery of Healthcare Intervention (ODHIN) study, a five-country cluster randomized factorial trial in which the researchers assessed three strategies to improve implementation of BI; internet intervention after screening failed to improve BI delivery, but training and support, and financial reimbursement were found to be valid strategies (25,38). In Catalonia, the engagement of GPs on a web-based BI was modest (data not published). According to GPs’ perceptions, financial incentives should be included in their salaries instead of being temporarily introduced and subject to a specific screening project. The ODHIN GPs felt that training and support must improve knowledge, skills and prioritization, and must be accompanied by enough time to learn techniques and to tailor them to specific barriers, according to different GPs’ points of view (39). The preventive model established in Spain for alcohol consumption seems to fail. Thus, it is important to spread the message among primary care professionals that managing AUD is, in fact, treating an illness, falling within their remit because they are trained to do so, in the same way they do with high blood pressure or hyperhcolesteromia (40). Alcohol consumption in Spain is slightly above the WHO European region per capita average (12.3 versus 11.9 l of pure alcohol per year), while the prevalence of AUD is lower (1.3% versus 7.5%), and the prevalence of alcohol use in the last 12 months is 73.4% (1). All these figures are typical of a viticulture society, in which there is a normative perception of high levels of consumption. We found that 12.4% of GPs who answered the survey were themselves classified as risky drinkers. Similar prevalences of risky drinking have been found among Italian resident physicians, US medical students, German hospital doctors and Belgian specialists (41–45), while the prevalence is low compared to Finnish medical students (24–49%) (46). We know that there is a misperception of alcohol use among risky drinkers, the so called ‘normative fallacy’, in which peers’ alcohol consumption is estimated to be at least as high as risky drinker’s intake (16). We expected that risky drinking GPs would be less aware of risky alcohol use among their patients, but our results did not support this hypothesis. However, we think that alcohol consumption patterns of GPs should continue to be considered in those studies that aim to explore barriers for BI, precisely because a lack of evidence is not evidence for a lack of effect or association. In Spain, a relationship exists between the patients’ awareness of their HT and living in rural areas. However, there are only a few studies that focus on alcohol consumption and living in rural areas (where, e.g., alcohol is perceived as a part of daily diet (47), and there is higher prevalence of adolescent alcohol consumption) (48), and no studies focusing on the relationship between HT awareness, alcohol consumption and rural areas. This approach is important because we already know that access to lifestyle counselling is lower in rural areas (49). Unfortunately, this point was also beyond the scope of our study. Furthermore, our study illustrates that the improved understanding of barriers can be directly translated into measures to improve the management of alcohol and HT at the primary care level. A reduction in workload could be brought about by introducing structural changes. Furthermore, promoting collaborative team-based policies between GPs, mental health care professionals, nurses and pharmacists or introducing financial incentives (37) could enhance alcohol and HT management. In addition, widespread (post-) graduate training to increase awareness of alcohol-related conditions and its management among primary care professionals, and facilitating skills to promote healthy lifestyle changes should be considered as key measures to improve knowledge and reduce the perceived stigma related to risky drinking or alcohol dependence. According to the Agency for Healthcare Research and Quality (US Department of Health and Human Services), ‘Integrated Behavioral Health Care’ is defined as ‘the care a patient experiences as a result of a team of primary care and behavioral health clinicians, working together with patients and families, using a systematic and cost-effective approach to provide patient-centered care for a defined population’(50). In our opinion, BIs for risky drinkers should be interpreted in this context. This study has some limitations. Although the original sample to whom the survey was sent was representative of Spanish GPs, the response rate was low (5.6%), which should be taken into account when generalizing the results. This response rate was similar to other countries in the same study (between 4.1 and 8.5%), but lower than in other similar studies. Web-based behaviour change intervention studies, using personalized e-mails, were found to have better response rates than studies using generic e-mails. Hence, future studies in this field should bear this approach in mind. Moreover, limitations inherent to self-reported answers (e.g. memory effects reducing reliability) also apply to this study, while the social desirability bias can be assumed to be minimal, as survey responses were given anonymously. The main strength of the study, however, is that it covers a large territory. In conclusion, as long as GPs do not consider alcohol consumption a relevant factor to deal with when treating hypertensive patients, they will not intervene. Changing this perception must be a priority, because it is well known that alcohol plays an important role in resistant HT, on top of the high burden of disease due to both alcohol and HT. Barriers described might be overcome through structural changes that could include training and support, financial incentives, conceptualizing AUD and alcohol consumption as parameters on a continuum and treating AUD as an illness in its own right, instead of regarding it as a factor in a preventive model. These strategies all show promise for reducing HT and its associated burden. Supplementary material Supplementary material is available at Family Practice online. Declaration Funding: The study was financially supported by an investigator-initiated grant to Jürgen Rehm and the GWT-TUD (Gesellschaft für Wissens - und Technologietransfer der TU Dresden mbH—company with limited liabilities for transferring knowledge and technology of the Dresden University of Technology) by Lundbeck. The study sponsor has had no role in study design, collection, analysis and interpretation of data. The study sponsor has also had no role in writing this account or the decision to submit this paper for publication. The corresponding author confirms that all authors had full access to the data in the study at all times and had final responsibility for the decision to submit for publication. This work was supported by RD12/0028/0016 project, Plan Nacional de I+D+I and financed jointly with ISCII-Subdirección General de Evaluación y Fondo Europeo de Desarrollo Regional (FEDER). The study also received the support of ‘Secretaria d’Universitats i Recerca del Departament d’Economia i Coneixement de la Generalitat de Catalunya’ (2014SGR649). Conflict of interest: LM has received honoraria from Lundbeck, outside the work for this project. HLP has received training funds from Lundbeck, Janssen, Pfizer, Lilly, Rovi and Esteve, and has also received fees from Lundbeck, Teva and Janssen. JÁA has acted as an advisor and received funding for research, publication and training from the following companies: Almirall, AstraZeneca, Glaxo-Smith-Kline, Lilly, Lundbeck, Merck, Pfizer, Servier, Esteve. JZ has acted as an advisor and has received funding for consultancy, research, publications and carrying out training activities from Lundbeck and Lilly and fees from F Glead. JR has received educational grants, travel support and honoraria from Lundbeck outside and unrelated to the work on this manuscript. AG has received financial support from Lundbeck, DyA Pharma and TEVA and has received fees by Lundbeck, DyA Pharma and Abbivie, which were unconnected to the research of this study. JM has received personal fees from Lundbeck. All other authors have no potential conflict of interest to declare. Acknowledgements The authors acknowledge all the GP associations and all GPs who participated in the survey. References 1. World Health Organization . Global Status Report on Alcohol and Health . http://apps.who.int/iris/ bitstream/10665/112736/1/9789240692763_eng. pdf?ua=1 (accessed on 12 January 2015). 2. Taylor B , Irving HM , Baliunas D et al. Alcohol and hypertension: gender differences in dose-response relationships determined through systematic review and meta-analysis . Addiction 2009 ; 104 : 1981 – 90 . Google Scholar CrossRef Search ADS PubMed 3. Rehm J , Gmel G , Serra C , Gual A . Reducing non-communicable disease via better detection of hypertension and alcohol problems in primary health care in Spain . Adicciones 2017 : 726 . 4. GBD 2015 Mortality and Causes of Death Collaborators. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015 . Lancet 2016 ; 388 : 1459 – 544 . CrossRef Search ADS PubMed 5. Banegas JR , Segura J , Sobrino J et al. Effectiveness of blood pressure control outside the medical setting . Hypertension 2007 ; 49 : 62 – 8 . Google Scholar CrossRef Search ADS PubMed 6. Banegas JR , Graciani A , de la Cruz-Troca JJ et al. Achievement of cardiometabolic goals in aware hypertensive patients in Spain: a nationwide population-based study . Hypertension 2012 ; 60 : 898 – 905 . Google Scholar CrossRef Search ADS PubMed 7. European Commission . Alcohol in Europe. A Public Health Perspective . 2006 . http://ec.europa.eu/health/archive/ph_determinants/life_style/alcohol/documents/alcohol_europe_en.pdf (accessed on September 2017). 8. Parry CD , Patra J , Rehm J . Alcohol consumption and non-communicable diseases: epidemiology and policy implications . Addiction 2011 ; 106 : 1718 – 24 . Google Scholar CrossRef Search ADS PubMed 9. American Heart Association . Know Your Risk Factors for High Blood Pressure . 2016 . http://www.heart.org/HEARTORG/Conditions/HighBloodPressure/UnderstandSymptomsRisks/Know-Your-Risk-Factors-for-High-Blood-Pressure_UCM_002052_Article.jsp#.WUpK2ezyjcs (accessed on 1 June 2017) . 10. Estruch R , Sacanella E , De la Sierra A et al. Effects of alcohol withdrawal on 24 hour ambulatory blood pressure among alcohol-dependent patients . Alcohol Clin Exp Res 2003 ; 27 : 2002 – 8 . Google Scholar CrossRef Search ADS PubMed 11. Roerecke M , Kaczorowski J , Tobe SW et al. The effect of a reduction in alcohol consumption on blood pressure: a systematic review and meta-analysis of trial data . Lancet Public Heal 2017 ; 2 : e108 – 20 . Google Scholar CrossRef Search ADS 12. Llisterri Caro JL , Rodríguez Roca GC , Alonso Moreno FJ et al. [Blood pressure control in hypertensive Spanish population attended in primary care setting. The PRESCAP 2010 study] . Med Clin (Barc) 2012 ; 139 : 653 – 61 . Google Scholar CrossRef Search ADS PubMed 13. American Heart Association . Resistant Hypertension—High Blood Pressure That’s Hard to Treat . 2016 . http://www.heart.org/HEARTORG/Conditions/HighBloodPressure/GettheFactsAboutHighBloodPressure/Resistant-Hypertension---High-Blood-Pressure-Thats-Hard-to-Treat_UCM_469327_Article.jsp#.WUpLaOzyjcs (accessed on 1 June 2017 ). 14. Banegas JR , Rodríguez-Artalejo F , de la Cruz Troca JJ et al. Blood pressure in Spain: distribution, awareness, control, and benefits of a reduction in average pressure . Hypertension 1998 ; 32 : 998 – 1002 . http://www.ncbi.nlm.nih.gov/pubmed/9856963 (accessed on 21 June 2017 ). 15. Gual A , Arbesú JÁ , Zarco J et al. Alcoholism and its treatment approach from a citizen perspective . Adicciones 28 ( 3 ): 163 – 73 . http://www.ncbi.nlm.nih.gov/pubmed/27399226 (accessed on 12 August 2016 ). 16. Gual A , Ángel Arbesú J , Zarco J et al. Risky drinkers underestimate their own alcohol consumption . Alcohol 2017 ; 52 : 516 – 7 . Google Scholar CrossRef Search ADS 17. Yaxley JP , Thambar S V . Resistant hypertension: an approach to management in primary care . J Fam Med Prim Care 2015 ; 4 : 193 – 9 . Google Scholar CrossRef Search ADS 18. Catalá-López F , Sanfélix-Gimeno G , García-Torres C et al. Control of arterial hypertension in Spain: a systematic review and meta-analysis of 76 epidemiological studies on 341 632 participants . J Hypertens 2012 ; 30 : 168 – 76 . Google Scholar CrossRef Search ADS PubMed 19. Delegación del Gobierno para Plan Nacional sobre Drogas . Encuesta Sobre Alcohol Y Drogas En Población General España Edades 2015–2016 . 2015 . 20. Barbosa C , Cowell A , Bray J , Aldridge A . The cost-effectiveness of alcohol screening, brief intervention, and referral to treatment (SBIRT) in emergency and outpatient medical settings . J Subst Abuse Treat 2015; 53 : 1 – 8 . Google Scholar CrossRef Search ADS PubMed 21. Wolstenholme A , Drummond C , Deluca P et al. Report on the Mapping of European Need and Service Provision for Early Diagnosis and Treatment of Alcohol Use Disorders . 2013 . http://amphoraproject.net/w2box/data/Deliverables/AMPHORA_WP6_D2.5.pdf. 22. Anderson P , Kaner E , Keurhorst M et al. Attitudes and learning through practice are key to delivering brief interventions for heavy drinking in primary health care: analyses from the ODHIN five country cluster randomized factorial trial . Int J Environ Res Public Health 2017 ; 14 : 121 . Google Scholar CrossRef Search ADS 23. Vendetti J , Gmyrek A , Damon D et al. Screening, brief intervention and referral to treatment (SBIRT): implementation barriers, facilitators and model migration . Addiction 2017 ; 112 ( Suppl 2 ): 23 – 33 . Google Scholar CrossRef Search ADS PubMed 24. Rehm J , Prieto JAA , Beier M et al. The role of alcohol in the management of hypertension in patients in European primary health care practices – a survey in the largest European Union countries . BMC Fam Pract 2016 ; 17 : 130 . Google Scholar CrossRef Search ADS PubMed 25. Anderson P , Bendtsen P , Spak F et al. Improving the delivery of brief interventions for heavy drinking in primary health care: outcome results of the Optimizing Delivery of Health Care Intervention (ODHIN) five-country cluster randomized factorial trial . Addiction 2016 ; 111 : 1935 – 45 . Google Scholar CrossRef Search ADS PubMed 26. Gual A , Segura L , Contel M et al. Audit-3 and Audit-4: effectiveness of two short forms of the alcohol use disorders identification test . Alcohol 2002 ; 37 : 591 – 6 . http://alcalc.oxfordjournals.org/content/37/6/591.abstract. 27. Chobanian AV , Bakris GL , Black HR et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report . JAMA 2003 ; 289 : 2560 . Google Scholar CrossRef Search ADS PubMed 28. Fang J , Ayala C , Loustalot F . Primary care providers’ recommendations for hypertension prevention, DocStyles survey, 2012 . J Prim Care Community Health 2015 ; 6 : 170 – 6 . Google Scholar CrossRef Search ADS PubMed 29. Corrao G , Bagnardi V , Zambon A et al. Exploring the dose-response relationship between alcohol consumption and the risk of several alcohol-related conditions: a meta-analysis . Addiction 1999 ; 94 : 1551 – 73 . http://www.ncbi.nlm.nih.gov/pubmed/10790907 (accessed on 15 December 2016 ). 30. Anderson P , Chisholm D , Fuhr DC . Effectiveness and cost-effectiveness of policies and programmes to reduce the harm caused by alcohol . Lancet 2009 ; 373 : 2234 – 46 . Google Scholar CrossRef Search ADS PubMed 31. Schomerus G , Matschinger H , Angermeyer MC . Attitudes towards alcohol dependence and affected individuals: persistence of negative stereotypes and illness beliefs between 1990 and 2011 . Eur Addict Res 2014 ; 20 : 293 – 9 . Google Scholar CrossRef Search ADS PubMed 32. Wangberg SC . Norwegian midwives’ use of screening for and brief interventions on alcohol use in pregnancy . Sex Reprod Healthc 2015 ; 6 : 186 – 90 . Google Scholar CrossRef Search ADS PubMed 33. Johnson M , Jackson R , Guillaume L et al. Barriers and facilitators to implementing screening and brief intervention for alcohol misuse: a systematic review of qualitative evidence . J Public Health (Oxf) 2011 ; 33 : 412 – 21 . Google Scholar CrossRef Search ADS PubMed 34. Villaescusa Sáiz R , Puente A , Pérez R et al. Cargas de trabajo de los médicos de atención primaria del Instituto Catalán de la Salud . Aten Primaria 2011 ; 43 : 214 – 5 . Google Scholar CrossRef Search ADS PubMed 35. Saha TD , Chou SP , Grant BF . Toward an alcohol use disorder continuum using item response theory: results from the National Epidemiologic Survey on Alcohol and Related Conditions . Psychol Med 2006 ; 36 : 931 – 41 . Google Scholar CrossRef Search ADS PubMed 36. Rehm J , Marmet S , Anderson P et al. Defining substance use disorders: do we really need more than heavy use ? Alcohol Alcohol 2013 ; 48 : 633 – 40 . Google Scholar CrossRef Search ADS PubMed 37. Anderson P , Bendtsen P , Spak F et al. Improving the delivery of brief interventions for heavy drinking in primary health care: outcome results of the ODHIN five country cluster randomized factorial trial . Addiction 2016 ; 111 : 1935 – 45 . Google Scholar CrossRef Search ADS PubMed 38. Bendtsen P , Mussener U , Karlsson N et al. Implementing referral to an electronic alcohol brief advice website in primary health care: results from the ODHIN implementation trial . BMJ Open 2016 : e010271, doi: 10.1136/bmjopen-2015-010271. 39. Keurhorst M , Heinen M , Colom J et al. Strategies in primary healthcare to implement early identification of risky alcohol consumption: why do they work or not? A qualitative evaluation of the ODHIN study . BMC Fam Pract 2016 ; 17 : 70 . Google Scholar CrossRef Search ADS PubMed 40. Rehm J , Anderson P , Manthey J et al. Alcohol use disorders in primary health care: what do we know and where do we go ? Alcohol Alcohol 2016 ; 51 : 422 – 7 . Google Scholar CrossRef Search ADS PubMed 41. Lamberti M , Napolitano F , Napolitano P et al. Prevalence of alcohol use disorders among under- and post-graduate healthcare students in Italy . PLoS One 2017 ; 12 : e0175719 . Google Scholar CrossRef Search ADS PubMed 42. Rosta J . Prevalence of problem-related drinking among doctors: a review on representative samples . Ger Med Sci 2005 ; 3 : Doc07 . http://www.ncbi.nlm.nih.gov/pubmed/19675724 (accessed on 29 August 2017) . 43. Rosta J . Hazardous alcohol use among hospital doctors in Germany . Alcohol Alcohol 2008 ; 43 : 198 – 203 . Google Scholar CrossRef Search ADS PubMed 44. Shah AA , Bazargan-Hejazi S , Lindstrom RW et al. Prevalence of at-risk drinking among a national sample of medical students . Subst Abus 2009 ; 30 : 141 – 9 . Google Scholar CrossRef Search ADS 45. Joos L , Glazemakers I , Dom G . Alcohol use and hazardous drinking among medical specialists . Eur Addict Res 2013 ; 19 : 89 – 97 . Google Scholar CrossRef Search ADS PubMed 46. Ketoja J , Svidkovski A-S , Heinälä P et al. Risky drinking and its detection among medical students . Addict Behav 2013 ; 38 : 2115 – 8 . Google Scholar CrossRef Search ADS PubMed 47. Mateos R , Páramo M , Carrera I et al. Alcohol consumption in a southern European region (Galicia, Spain) . Subst Use Misuse 2002 ; 37 : 1957 – 76 . http://www.ncbi.nlm.nih.gov/pubmed/12511060 (accessed on 21 June 2017 ). 48. Obradors-Rial N , Ariza C , Muntaner C . [Risky alcohol consumption and associated factors in adolescents aged 15 to 16 years in Central Catalonia (Spain): differences between rural and urban areas] . Gac Sanit 2014 ; 28 : 381 – 5 . Google Scholar CrossRef Search ADS PubMed 49. De la Cruz-Sánchez E , Aguirre-Gómez L . Health related lifestyle and preventive medical care of rural Spanish women compared to their urban counterparts . J Immigr Minor Heal 2014 ; 16 : 712 – 8 . 50. Agency for Healthcare Research and Quality . What Is Integrated Behavioral Health Care (IBHC) . 2017 . https://integrationacademy.ahrq.gov/resources/ibhc-measures-atlas/what-integrated-behavioral-health-care-ibhc#definition. © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Jadczak, Agathe Daria;Dollard, Joanne;Mahajan, Neha;Visvanathan, Renuka

2017 Family Practice

doi: 10.1093/fampra/cmx108pmid: 29145588

Abstract Background Exercise is considered to be the most effective strategy to treat, prevent and delay frailty, a prevalent geriatric syndrome observed in clinical practice. Encouraging frail older people to take up exercise is crucial in the management of this condition. The study aimed to explore pre-frail and frail older peoples’ perspectives in relation to being advised about exercise and their perceptions of the general practitioners’ (GPs) role in promoting exercise for older people. Methods Semi-structured interviews were conducted with 12 community-dwelling older (median age 83 years) participants screened pre-frail or frail using the FRAIL Screen. Their attitudes towards exercise, the advice received, their access to relevant information and their perceptions of the GP’s role in promoting exercise were explored. Thematic analysis was conducted to analyse data. Results The majority of participants had a positive attitude towards exercise, and many participants indicated a preference for being advised firstly by their GPs and then other healthcare professionals. Participants living in the community reported difficulties in accessing information on exercise and indicated that local governments and GP practices should promote exercise for older people more actively. Participants living in retirement villages, however, reported having access to relevant information and being encouraged to participate in exercise. Conclusion This research identified a gap in current practice, demonstrating that GPs, healthcare providers and local governments should promote exercise for older people more actively. Convincing health professionals to encourage regular exercise among their older patients would provide an opportunity to avoid and manage frailty in this population. Exercise, frail elderly, general practitioners, health promotion, perceptions, qualitative research Introduction Exercise is considered to be the most effective strategy for the prevention, treatment and postponement of frailty, a prevalent geriatric syndrome observed in clinical practice (1). Frailty is defined as a ‘clinically recognisable state of increased vulnerability resulting from age-associated decline in reserve and function’ (2). It includes clinical indicators, such as deficit accumulation, fatigue, sedentary behaviour, weight loss and physical function impairment (2,3), and is associated with increased hospitalization, disability, loss of independence and reduced quality of life (4,5). One of the most commonly used frameworks of frailty is the frailty phenotype by Fried et al. (6), which uses a physiological approach to categorize adults into robust (0 indicators), pre-frail (1–2 indicators) and frail (≥3 indicators) based on the accumulation of the following five physical conditions: unintentional weight loss, exhaustion, slow walking speed, weakness and low physical activity. The prevalence of frailty is reported to be 10.7% in community-dwelling older people, and it is estimated that by 2050, 4 million Australians will be either pre-frail or frail (7). Frailty is a dynamic syndrome that is treatable and reversible (8), and a critical component in its treatment is exercise (1). Exercise helps to increase and restore muscle strength and improve the overall physical function. Greater strength and mobility allow for greater independence and an enhanced quality of life (1,8,9). Promoting physical activity programmes for community-dwelling pre-frail and frail older adults and generally encouraging older people to take up exercise are critical in managing frailty (10,11). Although the benefits of exercise are well known, the uptake of exercise in older people is poor (12,13). Barriers to exercise among older people include health issues (e.g. pain), environmental factors (e.g. lack of transportation), lack of knowledge and the lack of physician advice (14). Costello et al. (15) conducted focus groups with 31 independent living, non-frail older adults and confirmed that general practitioners (GPs) play an important role in promoting exercise for older people. However, older adults perceived their exercise conversations with their GPs as inadequate, and it was suggested that healthcare providers, including GPs, should take collective responsibility for encouraging older people to take up exercise. With regard to the frail population, Broderick et al. (2015) investigated the perceptions of 29 frail older adults in order to determine their exercise motivation and behaviour. The study results indicate that family members and social support networks involving friends and peers are critical motivators of exercise activity among frail older people. However, social support networks decline as frailty increases, and family members may not only motivate but also limit frail older people in their physical activities in an effort to protect them (16). Research specifically investigating frail older people’s experiences of being advised about exercise and their opinions regarding who should advise them, and where, is sparse. By better understanding frail older people’s preferred source of exercise advice, researchers can ensure that community-based education and awareness programmes are not only appropriately composed but also that the information is made available through the most appropriate sources. The aims of this study were to (i) understand the perspectives of community-dwelling older people who are pre-frail or frail in relation to being advised about exercise and to (ii) explore their experiences with regard to any advice provided by their GPs. Methods Participants and procedures Participants were recruited from The Queen Elizabeth Hospital (TQEH), the Queen Elizabeth Specialist Centre (QE Specialist Centre) and the Geriatrics Training and Research with Aged Care Centre (G-TRAC Centre) in Adelaide, South Australia, Australia. Potential participants were asked by their TQEH geriatricians (affiliated with the researchers’ department) during a scheduled consult whether they were interested in participating in this study. Interested individuals were then referred to a researcher (ADJ) who provided more information about the study, confirmed interest and performed initial eligibility screening on-site following the consult or over the phone before enrolling them into the study and scheduling an interview. Participants aged 75 years and older, living in the community and screened pre-frail or frail using the FRAIL screen (17), were included in this study. The FRAIL screen represents the Fried frailty phenotype (6) and includes five questions about fatigue, resistance, ambulation, illness and weight loss. It categorizes individuals into pre-frail (1–2 deficits), frail (≥3 deficits) or robust (0 deficit). Participants with a medical history of dementia or who were unable to communicate in English were excluded from the study. Interviews and study questionnaires Semi-structured interviews were held at the Basil Hetzel Institute, an adjacent research centre of TQEH, at the G-TRAC Centre or at the participant’s home, depending on the participant’s preference. ADJ conducted face-to-face interviews with all participants, each lasting approximately 1 h. The interviews included 16 semi-structured questions asking about participants’ previous exercise experiences, their perspectives on receiving advice on exercise, their access to relevant information and their experiences with their GPs. The interviews were audio recorded and transcribed. The transcript of the first interview was reviewed by a member of the research team (JD) to assess ADJ’s interview technique for gathering quality data. Participants’ 5-year mortality risk was determined using the Charlson Comorbidity Index (18), and current physical activity levels were assessed using the Physical Activity Scale for the Elderly (PASE) questionnaire (19). Demographic information, recorded in a separate questionnaire, included age, gender, nationality, education, income level and participation in regular exercise. Data analysis Participants’ characteristics and demographic information were analysed using descriptive statistical methods. The interviews were analysed using thematic analysis (20). ADJ transcribed the audio-recorded interviews, reread the transcripts to ensure familiarity and coded the transcripts line by line using an inductive approach (21). Codes were then collated and categorized into potential themes, gathering all data relevant to each theme before refining the themes and generating clear definitions for each specific theme. The coding framework, the codes and the themes were then discussed with a member of the research team (JD), and a scholarly report of the analysis was prepared (20). NVivo version 10 software was used to manage data analysis. Quotations are provided in this article to support the themes. Data saturation was reached when there was enough information to replicate the study, and new information was not obtained by interviewing additional participants (22). Results Participants’ characteristics Seventeen potential participants were referred to ADJ and screened for eligibility. Three decided not to participate, another felt too unwell and another was neither pre-frail nor frail according to the FRAIL Screen (17). The selected participants provided written informed consent and were advised that participation was voluntary. Twelve participants (eight females and four males) were interviewed. By the ninth interview, data saturation was reached as it was noted that participants’ responses to the interview questions were repetitive. Recruitment was, therefore, ceased after the 12th interview. The median age of the participants was 83 years with a range of 76 to 91 years. Ten participants were pre-frail and two participants were frail. The most common deficits on the FRAIL screen included feeling fatigued (n = 6) and currently having more than five illnesses simultaneously (n = 5). Eleven participants had a low 5-year mortality risk according to the Charlson Comorbidity Index. Six participants reached the current physical activity recommendations for older people according to the PASE questionnaire, considering their age and gender. Demographic data revealed a lower income and a prevalence of men in the non-active group. However, there were no differences in age, education, frailty status and 5-year mortality risk between active and non-active participants (Table 1). Table 1. Participants’ health and demographic characteristics collected during the interviews Characteristic Total participants (n = 12) Active participants; PASE (n = 6) Non-active participants; PASE (n = 6) Age (median) 83 (range 76–91 years) 83 (range 76–91 years) 83 (range 76–90 years) Gender 8 females; 4 males 5 females; 1 male 3 females; 3 males FRAIL Screen 10 pre-frail; 2 frail 5 pre-frail; 1 frail 5 pre-frail; 1 frail Charlson Comorbidity Index 11 low 5-year mortality risk; 1 high 5-year mortality risk 6 low 5-year mortality risk; 0 high 5-year mortality risk 5 low 5-year mortality risk; 1 high 5-year mortality risk Income 8 low; 3 average; 1 high 3 low; 2 average; 1 high 5 low; 1 average; 0 high Highest education 2 primary school; 5 secondary school; 4 technical college; 1 university 0 primary school; 4 secondary school; 2 technical college; 0 university 2 primary school; 1 secondary school; 2 technical college; 1 university Nationality 8 Australian; 3 European; 1 Asian 5 Australian; 1 European; 0 Asian 3 Australian; 2 European; 1 Asian Participation in exercise programme 6 yes; 6 no 5 yes; 1 no 1 yes; 5 no Characteristic Total participants (n = 12) Active participants; PASE (n = 6) Non-active participants; PASE (n = 6) Age (median) 83 (range 76–91 years) 83 (range 76–91 years) 83 (range 76–90 years) Gender 8 females; 4 males 5 females; 1 male 3 females; 3 males FRAIL Screen 10 pre-frail; 2 frail 5 pre-frail; 1 frail 5 pre-frail; 1 frail Charlson Comorbidity Index 11 low 5-year mortality risk; 1 high 5-year mortality risk 6 low 5-year mortality risk; 0 high 5-year mortality risk 5 low 5-year mortality risk; 1 high 5-year mortality risk Income 8 low; 3 average; 1 high 3 low; 2 average; 1 high 5 low; 1 average; 0 high Highest education 2 primary school; 5 secondary school; 4 technical college; 1 university 0 primary school; 4 secondary school; 2 technical college; 0 university 2 primary school; 1 secondary school; 2 technical college; 1 university Nationality 8 Australian; 3 European; 1 Asian 5 Australian; 1 European; 0 Asian 3 Australian; 2 European; 1 Asian Participation in exercise programme 6 yes; 6 no 5 yes; 1 no 1 yes; 5 no PASE, Physical Activity Scale for the Elderly. Income per year: low (< $25 000), average ($25 000–$45 000) and high (> $45 000). View Large Table 1. Participants’ health and demographic characteristics collected during the interviews Characteristic Total participants (n = 12) Active participants; PASE (n = 6) Non-active participants; PASE (n = 6) Age (median) 83 (range 76–91 years) 83 (range 76–91 years) 83 (range 76–90 years) Gender 8 females; 4 males 5 females; 1 male 3 females; 3 males FRAIL Screen 10 pre-frail; 2 frail 5 pre-frail; 1 frail 5 pre-frail; 1 frail Charlson Comorbidity Index 11 low 5-year mortality risk; 1 high 5-year mortality risk 6 low 5-year mortality risk; 0 high 5-year mortality risk 5 low 5-year mortality risk; 1 high 5-year mortality risk Income 8 low; 3 average; 1 high 3 low; 2 average; 1 high 5 low; 1 average; 0 high Highest education 2 primary school; 5 secondary school; 4 technical college; 1 university 0 primary school; 4 secondary school; 2 technical college; 0 university 2 primary school; 1 secondary school; 2 technical college; 1 university Nationality 8 Australian; 3 European; 1 Asian 5 Australian; 1 European; 0 Asian 3 Australian; 2 European; 1 Asian Participation in exercise programme 6 yes; 6 no 5 yes; 1 no 1 yes; 5 no Characteristic Total participants (n = 12) Active participants; PASE (n = 6) Non-active participants; PASE (n = 6) Age (median) 83 (range 76–91 years) 83 (range 76–91 years) 83 (range 76–90 years) Gender 8 females; 4 males 5 females; 1 male 3 females; 3 males FRAIL Screen 10 pre-frail; 2 frail 5 pre-frail; 1 frail 5 pre-frail; 1 frail Charlson Comorbidity Index 11 low 5-year mortality risk; 1 high 5-year mortality risk 6 low 5-year mortality risk; 0 high 5-year mortality risk 5 low 5-year mortality risk; 1 high 5-year mortality risk Income 8 low; 3 average; 1 high 3 low; 2 average; 1 high 5 low; 1 average; 0 high Highest education 2 primary school; 5 secondary school; 4 technical college; 1 university 0 primary school; 4 secondary school; 2 technical college; 0 university 2 primary school; 1 secondary school; 2 technical college; 1 university Nationality 8 Australian; 3 European; 1 Asian 5 Australian; 1 European; 0 Asian 3 Australian; 2 European; 1 Asian Participation in exercise programme 6 yes; 6 no 5 yes; 1 no 1 yes; 5 no PASE, Physical Activity Scale for the Elderly. Income per year: low (< $25 000), average ($25 000–$45 000) and high (> $45 000). View Large Themes Four themes emerged from the data (Fig. 1): older peoples’ attitudes towards exercise; their difficulties in accessing information on exercise; the crucial role of GPs and healthcare professionals in promoting exercise; and the missing or limited advice on exercise provided by GPs. Figure 1. View largeDownload slide Data analysis and development of themes. Figure 1. View largeDownload slide Data analysis and development of themes. Attitudes, barriers and enablers of exercise The majority of participants (10 of 12) reported a positive attitude towards exercise and physical activity programmes, commenting on the importance of exercise and its benefits in older age, irrespective of their physical activity status (PASE). Being physically active in the past did not correlate with participants’ attitudes towards exercise reported during the interviews. Participants perceived multiple barriers to participation in exercise. These included family commitments (especially for women whose role as carer may override self-care, hobbies or other activities), physical limitations (pain and illness), transportation and seasonal climate (cold weather and darkness common during winter months in South Australia). We just couldn’t cope with the cold and both of us stopped exercising and got stuck inside. (#10) Enablers for exercise included exercising with their partner, social aspects and rehabilitation or healthcare services, such as physiotherapy after hip replacement, where participants recognized the benefits of exercise in the form of improvements in physical function and mobility. I was in hospital (hip operation) and then I went to rehabilitation. I had physiotherapy there every day. I really should have started those exercises straight away. (#4) With regard to enjoyable and preferred aspects of physical activity programmes, participants reported a preference for smaller classes with a variety of gentle, but challenging and beneficial, exercises tailored for their individual health, including strength and balance tasks. Half of the participants (6 out of 12) were engaged in regular exercise, participating in a structured exercise programme. Access to information on exercise and physical activity programmes Both active and non-active participants living in the community felt that information relating to opportunities for exercise in the community was insufficient. They suggested that the local government should be more engaged in promoting physical activity programmes for older people. The government should send out brochures to older people. They should get something in every suburb, like through the council. (#8) I don’t see any advertisement (for exercise or physical activity programs) […] and I have been to the local library and places like that […], but I’ve never seen something […]. I just don’t know where to get information. (#6) All participants living in retirement villages (3 out of 12) revealed that they had greater access to exercise information and physical activity programmes due to the services provided in their villages. Before I came here (retirement village) I had no idea where to go, I had no idea that people even went to exercise classes; you’ve never heard of it. (#12) The majority of participants (9 out of 12) preferred brochures with information, including the health benefits of exercise in older age and available physical activity programmes in the community. They also commented on their preference for personal advice from someone with a positive attitude. Role of GPs and healthcare professionals in promoting exercise Healthcare professionals appear to play a key role in the promotion of exercise for older people, especially when patients receive healthcare services like physiotherapy or rehabilitation. Five participants reported that they received information on exercise opportunities during these services, and four of these participants continued with exercise after completing their treatment. I had physiotherapy [. . . ] and he gave me some exercises [. . . ] which I still do. And they (rehabilitation centre) said there are also other exercise classes you can get involved if you like. (#8) However, half of the participants (6 out of 12) indicated a preference for being advised on exercise initially by their GPs as they see them regularly, and the GP is aware of their medical conditions and is the primary referrer to healthcare services. Initially the doctor as he knows in what condition you are and what you are able to do. (#5) At first point I would have thought [. . .], it is our GP. Older people go to see their physicians naturally. So it is first the GP. (#10) Family members were also reported to have significant influence. My daughter, she was the one that encouraged me to go to Pilates. (#4) Three participants did not nominate a particular individual as the most important adviser, instead, stating that ‘someone who knows about it’ should advise them about exercise, two participants stated that healthcare professionals should advise initially and one participant did not answer the question. Active participants did not differ from non-active participants in their opinions as to who should advise them on exercise. Advice on exercise provided by GPs Even though participants perceived that GPs play a key role in promoting exercise for older people, the majority of participants (11 out of 12; both active and non-active) reported no (n = 6) or only limited (n = 5) recollection of exercise advice being provided during consults. Advice was defined as limited if the GP suggested walking without giving detailed information on the frequency, duration or intensity, or did not refer the individual to healthcare professionals or available physical activity programmes. However, participants consistently expressed faith in their GPs and said that they would appreciate and follow their advice on exercise. I would do it (exercise) if my GP would say it. (#1) When the GP would say do this or this (exercises), yes, I would dare it. (#5) Participants also suggested that GPs should promote exercise and physical activity programmes more actively. If the doctor would have something to hand to you. A notice of who to see and what you could do (#5). One participant received information from his/her GP after taking the initiative and asking about available physical activity programmes. I want to go to an exercise class, do you know any where? And he said, yes I do, you are going to need my referral […]. They are very good, they have classes for falls and balance and all that kind of stuff, and strength. So I went down there for quite a few years. (#12) Discussion Three key findings emerged from this study. Firstly, data analysis indicated that pre-frail and frail older people had mostly positive attitudes towards exercise, irrespective of their current physical activity status. Secondly, half of the participants indicated a preference for being advised on exercise initially by their GPs, but only one participant could recollect getting advice. Thirdly, participants received information on exercise and available physical activity programmes mainly through allied healthcare services (i.e. physiotherapy and rehabilitation) and their retirement villages, while participants living outside of retirement villages or those not receiving any allied health services reported a lack of exposure to exercise and available physical activity programmes. The important role of retirement villages in successfully encouraging physical activity among older adults emerged as an unexpected finding. The mostly positive attitudes towards exercise among the pre-frail and frail older adults reported in this study are consistent with findings in other studies showing that older adults hold positive attitudes towards exercise (23–25). However, findings that report negative attitudes towards exercise among inactive older adults were not supported by this study as most of the inactive participants also reported positive attitudes (24–26). These findings might be due to participants who have a more positive attitude being more likely to consent to an interview about exercise. This bias has been reported previously, for example by Rich and Rogers (25), who used a survey to examine older adults’ attitudes towards exercise and mainly received responses from adults who were likely to exercise. However, the positive attitudes of non-exercisers provide an opportunity to coax more pre-frail and frail older people into physical activity. Attitudes towards GPs confirmed the results of other studies that have noted that older adults are more likely to accept health advice from GPs compared with any other age group (27). Older adults tend to be in regular contact with their GPs, who are not only aware of their medical conditions and limitations but also the primary referrers to healthcare services (13–15). The fact that half of the participants indicated a preference for being advised about exercise firstly by their GPs underlines the important role that GPs play. GPs should capitalize on older people’s positive attitudes towards exercise and their readiness to follow their advice by promoting exercise and available physical activity programmes more actively, especially among those who do not receive any therapy services. A shared decision-making mode where GPs collaborate with patients on treatment options and exercise strategies should be the ultimate goal rather than being paternalistic by merely telling patients what to do (28). However, exercise advice provided by physicians is reported to be rare (29) and not specific (9), which aligns with the findings of this study where only one participant recollected receiving adequate information on exercise and being referred to an exercise programme. It is suggested that specific advice, including the type, frequency, intensity and duration of exercise, can lead to a greater increase in physical activity among older people than general advice with no specifics on where to go and what to do (30). Education and awareness programmes tailored to GPs could be introduced to inform and encourage them to advise their older patients about available physical activity programmes or to refer patients to allied healthcare experts more frequently, as a lack of knowledge is reported to be the main reason why GPs do not advise their patients about exercise (31). The Enhanced Primary Care Program (currently active in Australia), for example, allows supported referrals to exercise physiologists and represents a unique opportunity to tackle frailty. Difficulty in accessing information on exercise for older people has been reported previously, as well as the need for healthcare providers and local communities to promote exercise for older people more actively (32). This study underlines this need and has identified clearly the difficulty in accessing information related to exercise and available physical activity programmes for community-dwelling pre-frail and frail older people living outside of retirement villages. Data from this study revealed that retirement villages and their associated services appear more successful in promoting exercise, as all of the participants in this study who were living in a retirement village received information on exercise and available physical activity programmes. There might be a health promotion strategy to be learnt from the retirement villages which local governments could use to integrate pre-frail and frail older people into community-based physical activity programmes. Furthermore, since brochures related to exercise and physical activity programmes were recorded as being the most popular source of information by participants in this study, local communities could strategically send out brochures to their older residents to promote available physical activity programmes and encourage participation. Healthcare services also play a crucial role in encouraging older people to take up exercise as evidenced by the comments from participants in this study. These tertiary services provided information about available physical activity programmes, and the majority (four out of five participants) continued with exercise after treatment. Healthcare services, like physiotherapy, could therefore be seen as an opportunity to introduce pre-frail and frail older people to physical activity programmes, as it has been reported before that rehabilitation or exercise programmes in a recovery setting impact positively on older people’s exercise beliefs (16). Limitations and future studies Recruiting community-dwelling frail older people to the study at the time of a geriatric consultation proved difficult due to the presence of severe medical conditions resulting in participants feeling too fatigued to be interviewed and declined to participate. This resulted in the recruitment of mostly pre-frail older adults, with the majority (9 out of 12) scoring only one deficit on the FRAIL Screen. The fact that participants were recruited from geriatric clinics only, were mostly pre-frail, were English speaking and had an average level of education limits the generalizability of the study results. Including more frail older adults with different cultural and educational backgrounds may have led to more diverse opinions in relation to exercise and being advised about exercise. Regarding sample size and data saturation, the sample size of qualitative studies depends more on the data in terms of richness and thickness rather than on the sample size (22). Twelve participants were sufficient as data saturation was reached after interviewing the ninth participant. Furthermore, half of the participants were active (PASE), despite being pre-frail or frail, suggesting that additional tools are required to identify pre-frail and frail older adults. The fact that our recruitment method was successful in recruiting pre-frail but not frail participants requires reflection, and future studies should allow more time and resources for recruitment, target more frail individuals (with scores of ≥2 on the FRAIL Screen) and explore why GPs are reluctant to give advice or refer patients to exercise programmes. Conclusion The results of this study suggest that GPs should capitalize on older people’s positive attitudes towards exercise and their readiness to follow the GP’s advice by promoting exercise and available physical activity programmes more actively. Additionally, tertiary healthcare services, like rehabilitation and therapy services, should also be seen as an opportunity where pre-frail and frail older people could be linked long term into exercise programmes. Retirement villages with health promotion strategies are a resource from which the wider community can learn. Further research into health promotion strategies used in retirement villages would be of interest, as well as to explore why GPs are reluctant to give advice or refer patients to exercise programmes. The findings suggest that to facilitate exercise in pre-frail and frail older people further effort to inform and promote exercise is vital. Declaration Funding: ADJ is a recipient of the Beacon PhD Scholarship from the University of Adelaide Ethics: This study received ethics approval from the University of Adelaide Human Research Ethics Committee (HREC reference number H-2015–161). Conflict of interest: The authors report no conflict of interests. References 1. Clegg A , Young J , Iliffe S et al. Frailty in elderly people . Lancet 2013 ; 381 : 752 – 62 . Google Scholar CrossRef Search ADS PubMed 2. Xue QL . The frailty syndrome: definition and natural history . Clin Geriatr Med 2011 ; 27 : 1 – 15 . Google Scholar CrossRef Search ADS PubMed 3. Cesari M , Landi F , Vellas B et al. Sarcopenia and physical frailty: two sides of the same coin . Front Aging Neurosci 2014 ; 6 : 192 . Google Scholar PubMed 4. Clark BC , Manini TM . Functional consequences of sarcopenia and dynapenia in the elderly . Curr Opin Clin Nutr Metab Care 2010 ; 13 : 271 – 6 . Google Scholar CrossRef Search ADS PubMed 5. Weiss CO . Frailty and chronic diseases in older adults . Clin Geriatr Med 2011 ; 27 : 39 – 52 . Google Scholar CrossRef Search ADS PubMed 6. Fried LP , Tangen CM , Walston J et al. ; Cardiovascular Health Study Collaborative Research Group . Frailty in older adults: evidence for a phenotype . J Gerontol A Biol Sci Med Sci 2001 ; 56 : M146 – 56 . Google Scholar CrossRef Search ADS PubMed 7. Collard RM , Boter H , Schoevers RA et al. Prevalence of frailty in community-dwelling older persons: a systematic review . J Am Geriatr Soc 2012 ; 60 : 1487 – 92 . Google Scholar CrossRef Search ADS PubMed 8. Gill TM , Gahbauer EA , Allore HG et al. Transitions between frailty states among community-living older persons . Arch Intern Med 2006 ; 166 : 418 – 23 . Google Scholar CrossRef Search ADS PubMed 9. Taylor D . Physical activity is medicine for older adults . Postgrad Med J 2014 ; 90 : 26 – 32 . Google Scholar CrossRef Search ADS PubMed 10. Hallal PC , Andersen LB , Bull FC et al. ; Lancet Physical Activity Series Working Group . Global physical activity levels: surveillance progress, pitfalls, and prospects . Lancet 2012 ; 380 : 247 – 57 . Google Scholar CrossRef Search ADS PubMed 11. Australian Institute of Health and Welfare . 2014 . Australia’s Health 2014. Australia’s Health Series no. 14. Cat. no. AUS 178 . Canberra : AIHW . 12. Franco MR , Tong A , Howard K et al. Older people’s perspectives on participation in physical activity: a systematic review and thematic synthesis of qualitative literature . Br J Sports Med 2015 ; 49 : 1268 – 76 . Google Scholar CrossRef Search ADS PubMed 13. Rantanen T . Promoting mobility in older people . J Prev Med Public Health 2013 ; 46 ( Suppl 1 ): S50 – 4 . Google Scholar CrossRef Search ADS PubMed 14. Schutzer KA , Graves BS . Barriers and motivations to exercise in older adults . Prev Med 2004 ; 39 : 1056 – 61 . Google Scholar CrossRef Search ADS PubMed 15. Costello E , Leone JE , Ellzy M et al. Older adult perceptions of the physicians’ role in promoting physical activity . Disabil Rehabil 2013 ; 35 : 1191 – 8 . Google Scholar CrossRef Search ADS PubMed 16. Broderick L , McCullagh R , White EB et al. Perceptions, expectations, and informal supports influence exercise activity in frail older adults . SAGE Open 2015 ; 5 : 1 – 10 . Google Scholar CrossRef Search ADS 17. Morley JE , Vellas B , van Kan GA et al. Frailty consensus: a call to action . J Am Med Dir Assoc 2013 ; 14 : 392 – 7 . Google Scholar CrossRef Search ADS PubMed 18. Charlson ME , Pompei P , Ales KL et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation . J Chronic Dis 1987 ; 40 : 373 – 83 . Google Scholar CrossRef Search ADS PubMed 19. Washburn RA , Smith KW , Jette AM et al. The Physical Activity Scale for the Elderly (PASE): development and evaluation . J Clin Epidemiol 1993 ; 46 : 153 – 62 . Google Scholar CrossRef Search ADS PubMed 20. Joffe H , Yardley L . Content and thematic analysis . In: Marks DF & Yardley L (eds). Research Methods for Clinical and Health Psychology . London : Sage Publications Ltd , 2004 , pp. 56 – 68 . 21. Braun V , Clarke V . Using thematic analysis in psychology . Qual Res Psychol 2006 ; 3 : 77 – 101 . Google Scholar CrossRef Search ADS 22. Fusch PI , Ness LR . Are we there yet? Datasaturation in qualitative research . QualRep 2015 ; 20 : 1408 – 1416 . 23. Archer T , Fredriksson A , Schütz E et al. Influence of physical exercise on neuroimmunological functioning and health: aging and stress . Neurotox Res 2011 ; 20 : 69 – 83 . Google Scholar CrossRef Search ADS PubMed 24. Lee C . Attitudes, knowledge, and stages of change: a survey of exercise patterns in older Australian women . Health Psychol 1993 ; 12 : 476 – 80 . Google Scholar CrossRef Search ADS PubMed 25. Rich SC , Rogers ME . Stage of exercise change model and attitudes toward exercise in older adults . Percept Mot Skills 2001 ; 93 : 141 – 4 . Google Scholar CrossRef Search ADS PubMed 26. Courneya KS . Understanding readiness for regular physical activity in older individuals: an application of the theory of planned behavior . Health Psychol 1995 ; 14 : 80 – 7 . Google Scholar CrossRef Search ADS PubMed 27. Whitlock EP , Orleans CT , Pender N et al. Evaluating primary care behavioral counseling interventions: an evidence-based approach . Am J Prev Med 2002 ; 22 : 267 – 84 . Google Scholar CrossRef Search ADS PubMed 28. Charles C , Gafni A , Whelan T . Decision-making in the physician-patient encounter: revisiting the shared treatment decision-making model . Soc Sci Med 1999 ; 49 : 651 – 61 . Google Scholar CrossRef Search ADS PubMed 29. Glasgow RE , Eakin EG , Fisher EB et al. Physician advice and support for physical activity: results from a national survey . Am J Prev Med 2001 ; 21 : 189 – 96 . Google Scholar CrossRef Search ADS PubMed 30. Kerse N , Elley CR , Robinson E et al. Is physical activity counseling effective for older people? A cluster randomized, controlled trial in primary care . J Am Geriatr Soc 2005 ; 53 : 1951 – 6 . Google Scholar CrossRef Search ADS PubMed 31. Forman-Hoffman V , Little A , Wahls T . Barriers to obesity management: a pilot study of primary care clinicians . BMC Fam Pract 2006 ; 7 : 35 . Google Scholar CrossRef Search ADS PubMed 32. Bethancourt HJ , Rosenberg DE , Beatty T et al. Barriers to and facilitators of physical activity program use among older adults . Clin Med Res 2014 ; 12 : 10 – 20 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Thai, Thuy N;Dale, Ariella P;Ebell, Mark H

2017 Family Practice

doi: 10.1093/fampra/cmx072pmid: 29045629

Abstract Introduction Both non-Group A streptococcal (non-GAS) pharyngitis and Group A streptococcal (GAS) pharyngitis are commonly found in patients with sore throat. It is not known whether or not they present with similar signs and symptoms compared to patients with non-streptococcal pharyngitis. Methods MEDLINE was searched for prospective studies that reported throat culture for both GAS and non-GAS as a reference standard, and reported at least one sign, symptom, or the Centor score. Summary estimates of sensitivity, specificity, likelihood ratios (LR+ and LR−), and diagnostic odds ratios (DOR) were calculated using a bivariate random effects model. Summary receiver operating characteristic (ROC) curves were created for key signs and symptoms. Results Eight studies met our inclusion criteria. Tonsillar exudate had the highest LR+ for both GAS and non-GAS pharyngitis (1.53 versus 1.71). The confidence intervals of sensitivity, LR+, LR−, and DOR for all signs, symptoms, and the Centor score between two groups overlapped, with the relative difference between sensitivities within 15% for arthralgia or myalgia, fever, injected throat, tonsillar enlargement, and tonsillar exudate. Larger differences in sensitivities were observed for sore throat, cervical adenopathy, and lack of a cough, although the difference for lack of a cough largely due to a single outlier. Discussion Signs and symptoms of patients with GAS and non-GAS pharyngitis are generally similar. No signs or symptoms clearly distinguish GAS from non-GAS infection. Further work is needed to determine whether Group C streptococcus is a pathogen that should be treated. Group A, non-Group A, pharyngitis, signs, streptococcus, symptoms Introduction Sore throat is the second most common acute infection seen by family physicians, and Group A β-hemolytic streptococcus (GAS) is the most common bacterial pathogen causing acute pharyngitis (1). In 2010–2011, pharyngitis was one of the three most common diagnoses leading to an antibiotic prescription, with 56% of children and 72% of adults receiving one (2). However, the prevalence of GAS in this study was only 37% in children and 18% in adults (2). Non-group A streptococcal pharyngitis (non-GAS), most commonly Group C streptococcus (GCS) or Group G streptococcus (GGS), is also observed in patients with sore throat. According to a recent meta-analysis, the prevalence of GCS was 6.6% in laboratory-based studies of throat cultures and 6.1% in outpatients with sore throat (3). Nevertheless, whether GCS and GGS are pathogenic for pharyngitis is controversial. Many virulence properties found in GCS and GGS are similar to those of GAS (4). GCS and GGS infections have been associated with severe or recurrent pharyngitis, reactive arthritis, and other adverse outcomes such as acute glomerulonephritis, streptococcal toxic shock-like syndrome, and subdural empyema in case reports (5–12). Antibiotics have been recommended by some authors for non-GAS pharyngitis, in particular for Group C strep (13,14). However, the potential benefits of treatment for non-GAS pharyngitis have not been well studied. A double blind randomized controlled trial showed that in patients with high colony counts of non-GAS infection, penicillin resolved symptoms 1.3 days earlier as compared with the placebo (14). Currently, there is insufficient evidence to support antibiotic treatment for patients with GCS or GGS infections (15–17). Rapid antigen detection tests (RADTs) are not available for non-GAS (18), so initial diagnosis of non-GAS pharyngitis is mainly based on clinical signs and symptoms and then confirmed by culture. One unanswered question in the literature is whether GAS and non-GAS infections present similarly. Some researchers have found that GAS and non-GAS pharyngitis present with similarly (18,19), while another found six clinical distinctions between GAS and GCS infection (20). The current study will be the first meta-analysis comparing the accuracy of signs and symptoms for distinguishing GAS and non-GAS pharyngitis from non-streptococcal pharyngitis. Put another way, do GAS and non-GAS present similarly or differently? Methods Inclusion and exclusion criteria We included studies that reported at least one sign, symptom, or the Centor score for diagnosis of both GAS and non-GAS (Group B, C, D, F, or G) streptococcus in the same population. The Centor score assigns one point for each of the following clinical criteria: tonsillar exudates, swollen tender anterior cervical nodes, lack of a cough, and fever, with higher scores associated with a higher likelihood of GAS (21). Our target population was children or adults presenting with pharyngitis or sore throat as the chief complaint. The reference standard had to be throat culture reporting the number with both GAS and non-GAS infection. Only studies that had a cohort design with prospective data collection of signs and symptoms were included in our review. We excluded studies that included a specialized population (e.g. HIV positive, immunosuppressed, undergoing surgery), case-control studies, retrospective studies (‘chart reviews’) or studies without adequate data to calculate test accuracy. There were no language limits in our study. Search strategy and data abstraction We searched MEDLINE using two search strategies shown in Supplementary Search Strategy. First, we searched for previous published systematic reviews of the diagnosis of GAS or any non-GAS infection from 01 January 2000 to the current time (26 March 2016). We then compiled a list of all studies that had been included in at least one of the systematic reviews. These articles were all reviewed in full. Second, a bridge search was performed to find more recent studies of GAS and any non-GAS clinical diagnosis that were published since the year of the most recent systematic review in the first search strategy (from 01 January 2010 to 26 March 2016). Two investigators reviewed those abstracts in the bridge search in parallel, and the full articles of any abstract identified by either investigator as potentially relevant were pulled to review in full. Every full article from both search strategies was carefully reviewed by two investigators to evaluate for the inclusion and exclusion criteria. At least two investigators abstracted data about study design, study quality, and accuracy of signs, symptoms, and the Centor score. Any conflict in the review process was resolved by consensus discussion, generally involving all three researchers. Once data were abstracted, we combined similar variables. Notably, adenitis, adenopathy, cervical adenitis, tender cervical lymph nodes, and cervical adenopathy were combined as ‘cervical adenopathy’; enlarged tonsils or tonsillar hypertrophy as ‘tonsillar enlargement’; and tonsillar exudate, follicular tonsils, and purulent tonsils as ‘tonsillar exudate’. Quality assessment The QUADAS-2 framework was adapted for the purpose of this study (22) and is shown in Supplementary QUADAS-2 Instrument. Each study was assessed for patient selection, index test, reference standard test, and flow and timing with total 16 criteria. If a study met a criteria, then yes (Y) was assigned, if it did not meet a criteria no (N) was assigned, and if it was not possible to evaluate a criteria based on the study’s information, uncertain (U) was assigned. Overall risk of bias was low if all domains were at low risk of bias; moderate if one domain was at high risk of bias; and high if two or more domains were at high risk of bias. Based on these criteria, the quality of the included studies was assessed independently by two investigators. A consensus discussion resolved any disagreements between them. Analytic strategy R version 3.2.4 and the mada (Meta-Analysis of Diagnostic Accuracy) package was used to perform a bivariate random effects meta-analysis for each test using the Reitsma procedure (23,24). For signs, symptoms and the Centor score for low and high risk groups, summary statistics for accuracy including sensitivity, specificity, likelihood ratios (LR+, LR−), area under receiver operating characteristic curves (AUROCC), and the diagnostic odds ratio (DOR) were calculated. The calibrate package was used to plot the calibration curve of sensitivities of signs, symptoms, and the Centor score (25). For individual studies with sensitivities of 0 or 1, we added 0.5 to each cell of the 2 × 2 table to calculate LR+ or LR- (26).We assessed heterogeneity using visual inspection of the summary ROC curves and by calculating 95% confidence intervals. Where possible, we evaluated accuracy separately for studies of children only. We also performed a statistical test to compare sensitivity of each sign, symptom, and the Centor score between GAS and non-GAS pharyngitis. Specificities were the same for GAS and non-GAS pharyngitis, since in each case these groups were compared within a study to the same group of patients with non-streptococcal pharyngitis. We did a search for studies reporting the prevalence of GAS and non-GAS infections in both symptomatic patients and controls. Pooled estimates of prevalences were calculated by using metaprop procedure of meta package (27). Results Study characteristics The result of the two search strategies is presented in the PRISMA diagram in Supplementary Prisma Flow S1. One study was excluded from our review because of the small number of non-GAS infections identified (1 case of GCS, 1 case of GGS) (28). A total of eight studies met our inclusion criteria; their characteristics are summarized in Table 1. Four studies enrolled only children (19,29–31), one enrolled only adults (32) and three enrolled both children and adults (18,33,34). Settings included a hospital emergency department (30), an outpatient department (19,31), a general practice (18,33), or a healthcare centre (29,32,34). All eight studies used a prospective cohort design with throat culture as the reference standard. Four studies reported signs or symptoms (18,29–31), one study reported only the Centor score (34), and three studies reporting both signs, symptoms, and the Centor score (19,32,33). One study was an extreme outlier with regards to absence of cough (reportedly present in all 589 patients with non-streptococcal pharyngitis) and cervical adenopathy (present in only 6% with GAS pharyngitis, as opposed to 58% to 100% in the other studies) (31). Table 1. Characteristics of included studies of children and children or adults, with data collected between 1964 and 2008 Study Patient population Setting Sample size Sign/ symptom included Sample size of GAS Sample size of non-GAS Sample size of non streptococcus Mean age and/or age range Country Year(s) of data collection Children  Kaplan, 1971 Children under 15 years with symptoms of uncomplicated pharyngitis. Excluded if AOMa or LRTIb. Hospital emergency room 624 Adenitis, adenopathy, coryza, lack of a cough, exudate, fever, injected throat, previous URTIc, sick siblings, sore throat 218 38 368 Mean 6.8 years USA 1964 to 1967  Principi, 1990 Consecutive children with symptomatic pharyngitis and at least one of the following: hyperemia, fever, sore throat, exudate, or palatine petechiae. Excluded if recent antibiotic or peritonsillar abscess. Outpatient Department 865 Adenitis, lack of a cough, exudate, fever, headache, hyperemia, petechiae, sore throat 230 46 589 Range 5 months to 14 years; 394/852 were 3 to 5 years. Italy 1988 to 1989  Bassili, 2002 Children 1 to 15 years with uncomplicated tonsillopharyngitis and either pharyngeal erythema or exudate, or tonsillar enlargement and redness. Excluded if recent antibiotics, AOMa, or other infections. Private health services and public facilities 578 Abdominal pain, arthralgia or myalgia, lack of cough, dysphagia, fever, hoarseness, injected throat, petechiae, pharyngeal exudate, scarlatiniform rash, sore throat, throat congestion, tonsillar enlargement, tonsillar exudate, vomiting, watery eyes 98 69 411 Mean age 6.3 years, range 1 to 15 years Egypt 2000  Fretzayas, 2009 Children age 4 to 14 years with pharyngitis. Excluded if recent antibiotics. Outpatient department 144 Adenitis, exudate, sore throat, tonsillar enlargement, Centor score of 0 to 4 57 13 74 Mean 6.5 years Greece 2006 Children and adults  Seppala, 1993 Adults with sore throat. Excluded if received antibiotics last month Private health centre 106 Centor score of 3 to 4, adenitis, fever, rhinorrhea and/or cough, tonsillar exudate 5 19 82 Mean 30.1 years, range 15 to 62 years Finland 1986  Lindbaek, 2005 Adults and children presenting with sore throat. Excluded if recent antibiotics. General practices 306 Centor score of 0 to 4, adenitis, lack of a cough, dysphagia, fever, injected throat, tonsillar exudate 127 33 146 Mean 23.9 years Norway 2000 to 2002  Llor, 2008 Consecutive patients 14 years and older with sore throat and at least Centor score 2. Excluded if recent antibiotics. Primary care centre 182 Centor score of 2 to 4 40 42 100 Mean 30.6 years Spain 2007 to 2008  Little, 2012 Adults and children at least 5 years of age with chief complaint of sore throat of less than 2 week duration or physician diagnosed pharyngitis as the primary symptom. Excluded if judged to have non-infectious cause of sore throat. General practice 592 Adenopathy, age less than 10 years, arthralgia or myalgia, absence of coryza, lack of a cough, fever, headache, prior duration ≤3 days, sore throat, tonsillar exudate 136 45 410 Range 5 years and older; 67 of 605 were age < 10 years. England 2007 to 2008 Study Patient population Setting Sample size Sign/ symptom included Sample size of GAS Sample size of non-GAS Sample size of non streptococcus Mean age and/or age range Country Year(s) of data collection Children  Kaplan, 1971 Children under 15 years with symptoms of uncomplicated pharyngitis. Excluded if AOMa or LRTIb. Hospital emergency room 624 Adenitis, adenopathy, coryza, lack of a cough, exudate, fever, injected throat, previous URTIc, sick siblings, sore throat 218 38 368 Mean 6.8 years USA 1964 to 1967  Principi, 1990 Consecutive children with symptomatic pharyngitis and at least one of the following: hyperemia, fever, sore throat, exudate, or palatine petechiae. Excluded if recent antibiotic or peritonsillar abscess. Outpatient Department 865 Adenitis, lack of a cough, exudate, fever, headache, hyperemia, petechiae, sore throat 230 46 589 Range 5 months to 14 years; 394/852 were 3 to 5 years. Italy 1988 to 1989  Bassili, 2002 Children 1 to 15 years with uncomplicated tonsillopharyngitis and either pharyngeal erythema or exudate, or tonsillar enlargement and redness. Excluded if recent antibiotics, AOMa, or other infections. Private health services and public facilities 578 Abdominal pain, arthralgia or myalgia, lack of cough, dysphagia, fever, hoarseness, injected throat, petechiae, pharyngeal exudate, scarlatiniform rash, sore throat, throat congestion, tonsillar enlargement, tonsillar exudate, vomiting, watery eyes 98 69 411 Mean age 6.3 years, range 1 to 15 years Egypt 2000  Fretzayas, 2009 Children age 4 to 14 years with pharyngitis. Excluded if recent antibiotics. Outpatient department 144 Adenitis, exudate, sore throat, tonsillar enlargement, Centor score of 0 to 4 57 13 74 Mean 6.5 years Greece 2006 Children and adults  Seppala, 1993 Adults with sore throat. Excluded if received antibiotics last month Private health centre 106 Centor score of 3 to 4, adenitis, fever, rhinorrhea and/or cough, tonsillar exudate 5 19 82 Mean 30.1 years, range 15 to 62 years Finland 1986  Lindbaek, 2005 Adults and children presenting with sore throat. Excluded if recent antibiotics. General practices 306 Centor score of 0 to 4, adenitis, lack of a cough, dysphagia, fever, injected throat, tonsillar exudate 127 33 146 Mean 23.9 years Norway 2000 to 2002  Llor, 2008 Consecutive patients 14 years and older with sore throat and at least Centor score 2. Excluded if recent antibiotics. Primary care centre 182 Centor score of 2 to 4 40 42 100 Mean 30.6 years Spain 2007 to 2008  Little, 2012 Adults and children at least 5 years of age with chief complaint of sore throat of less than 2 week duration or physician diagnosed pharyngitis as the primary symptom. Excluded if judged to have non-infectious cause of sore throat. General practice 592 Adenopathy, age less than 10 years, arthralgia or myalgia, absence of coryza, lack of a cough, fever, headache, prior duration ≤3 days, sore throat, tonsillar exudate 136 45 410 Range 5 years and older; 67 of 605 were age < 10 years. England 2007 to 2008 aAcute otitis media; bLower respiratory tract infection; cUpper respiratory tract infection. View Large Table 1. Characteristics of included studies of children and children or adults, with data collected between 1964 and 2008 Study Patient population Setting Sample size Sign/ symptom included Sample size of GAS Sample size of non-GAS Sample size of non streptococcus Mean age and/or age range Country Year(s) of data collection Children  Kaplan, 1971 Children under 15 years with symptoms of uncomplicated pharyngitis. Excluded if AOMa or LRTIb. Hospital emergency room 624 Adenitis, adenopathy, coryza, lack of a cough, exudate, fever, injected throat, previous URTIc, sick siblings, sore throat 218 38 368 Mean 6.8 years USA 1964 to 1967  Principi, 1990 Consecutive children with symptomatic pharyngitis and at least one of the following: hyperemia, fever, sore throat, exudate, or palatine petechiae. Excluded if recent antibiotic or peritonsillar abscess. Outpatient Department 865 Adenitis, lack of a cough, exudate, fever, headache, hyperemia, petechiae, sore throat 230 46 589 Range 5 months to 14 years; 394/852 were 3 to 5 years. Italy 1988 to 1989  Bassili, 2002 Children 1 to 15 years with uncomplicated tonsillopharyngitis and either pharyngeal erythema or exudate, or tonsillar enlargement and redness. Excluded if recent antibiotics, AOMa, or other infections. Private health services and public facilities 578 Abdominal pain, arthralgia or myalgia, lack of cough, dysphagia, fever, hoarseness, injected throat, petechiae, pharyngeal exudate, scarlatiniform rash, sore throat, throat congestion, tonsillar enlargement, tonsillar exudate, vomiting, watery eyes 98 69 411 Mean age 6.3 years, range 1 to 15 years Egypt 2000  Fretzayas, 2009 Children age 4 to 14 years with pharyngitis. Excluded if recent antibiotics. Outpatient department 144 Adenitis, exudate, sore throat, tonsillar enlargement, Centor score of 0 to 4 57 13 74 Mean 6.5 years Greece 2006 Children and adults  Seppala, 1993 Adults with sore throat. Excluded if received antibiotics last month Private health centre 106 Centor score of 3 to 4, adenitis, fever, rhinorrhea and/or cough, tonsillar exudate 5 19 82 Mean 30.1 years, range 15 to 62 years Finland 1986  Lindbaek, 2005 Adults and children presenting with sore throat. Excluded if recent antibiotics. General practices 306 Centor score of 0 to 4, adenitis, lack of a cough, dysphagia, fever, injected throat, tonsillar exudate 127 33 146 Mean 23.9 years Norway 2000 to 2002  Llor, 2008 Consecutive patients 14 years and older with sore throat and at least Centor score 2. Excluded if recent antibiotics. Primary care centre 182 Centor score of 2 to 4 40 42 100 Mean 30.6 years Spain 2007 to 2008  Little, 2012 Adults and children at least 5 years of age with chief complaint of sore throat of less than 2 week duration or physician diagnosed pharyngitis as the primary symptom. Excluded if judged to have non-infectious cause of sore throat. General practice 592 Adenopathy, age less than 10 years, arthralgia or myalgia, absence of coryza, lack of a cough, fever, headache, prior duration ≤3 days, sore throat, tonsillar exudate 136 45 410 Range 5 years and older; 67 of 605 were age < 10 years. England 2007 to 2008 Study Patient population Setting Sample size Sign/ symptom included Sample size of GAS Sample size of non-GAS Sample size of non streptococcus Mean age and/or age range Country Year(s) of data collection Children  Kaplan, 1971 Children under 15 years with symptoms of uncomplicated pharyngitis. Excluded if AOMa or LRTIb. Hospital emergency room 624 Adenitis, adenopathy, coryza, lack of a cough, exudate, fever, injected throat, previous URTIc, sick siblings, sore throat 218 38 368 Mean 6.8 years USA 1964 to 1967  Principi, 1990 Consecutive children with symptomatic pharyngitis and at least one of the following: hyperemia, fever, sore throat, exudate, or palatine petechiae. Excluded if recent antibiotic or peritonsillar abscess. Outpatient Department 865 Adenitis, lack of a cough, exudate, fever, headache, hyperemia, petechiae, sore throat 230 46 589 Range 5 months to 14 years; 394/852 were 3 to 5 years. Italy 1988 to 1989  Bassili, 2002 Children 1 to 15 years with uncomplicated tonsillopharyngitis and either pharyngeal erythema or exudate, or tonsillar enlargement and redness. Excluded if recent antibiotics, AOMa, or other infections. Private health services and public facilities 578 Abdominal pain, arthralgia or myalgia, lack of cough, dysphagia, fever, hoarseness, injected throat, petechiae, pharyngeal exudate, scarlatiniform rash, sore throat, throat congestion, tonsillar enlargement, tonsillar exudate, vomiting, watery eyes 98 69 411 Mean age 6.3 years, range 1 to 15 years Egypt 2000  Fretzayas, 2009 Children age 4 to 14 years with pharyngitis. Excluded if recent antibiotics. Outpatient department 144 Adenitis, exudate, sore throat, tonsillar enlargement, Centor score of 0 to 4 57 13 74 Mean 6.5 years Greece 2006 Children and adults  Seppala, 1993 Adults with sore throat. Excluded if received antibiotics last month Private health centre 106 Centor score of 3 to 4, adenitis, fever, rhinorrhea and/or cough, tonsillar exudate 5 19 82 Mean 30.1 years, range 15 to 62 years Finland 1986  Lindbaek, 2005 Adults and children presenting with sore throat. Excluded if recent antibiotics. General practices 306 Centor score of 0 to 4, adenitis, lack of a cough, dysphagia, fever, injected throat, tonsillar exudate 127 33 146 Mean 23.9 years Norway 2000 to 2002  Llor, 2008 Consecutive patients 14 years and older with sore throat and at least Centor score 2. Excluded if recent antibiotics. Primary care centre 182 Centor score of 2 to 4 40 42 100 Mean 30.6 years Spain 2007 to 2008  Little, 2012 Adults and children at least 5 years of age with chief complaint of sore throat of less than 2 week duration or physician diagnosed pharyngitis as the primary symptom. Excluded if judged to have non-infectious cause of sore throat. General practice 592 Adenopathy, age less than 10 years, arthralgia or myalgia, absence of coryza, lack of a cough, fever, headache, prior duration ≤3 days, sore throat, tonsillar exudate 136 45 410 Range 5 years and older; 67 of 605 were age < 10 years. England 2007 to 2008 aAcute otitis media; bLower respiratory tract infection; cUpper respiratory tract infection. View Large Quality of included studies The results of the quality assessment are summarized in Supplementary Table S1, including our adaptation of the QUADAS-2 framework for the current study. Because the individual domains were at low risk of bias for all eight studies, the overall risk of bias of all eight studies was assessed to be low. Accuracy of signs and symptoms For each sign or symptom, we evaluated its accuracy as a test for distinguishing GAS from non-streptococcal infection, and for distinguishing non-GAS infection (i.e. Group B, C, D, F or G streptococcal infection) from non-streptococcal infection. We selected common signs and symptoms among the included studies for our pooled analysis: arthralgia or myalgia (2 studies), cervical adenopathy (6 studies), lack of a cough (5 studies), fever (6 studies), injected throat (3 studies), sore throat (4 studies), tonsillar enlargement (2 studies), tonsillar exudate (4 studies), Centor score of 2 or higher (3 studies), Centor sore of 3 or higher (4 studies). The accuracy of signs and symptoms for both GAS and non-GAS infections is summarized in Table 2. A detailed table that shows data for individual studies is shown in Supplementary Table S2. Table 2. Accuracy of signs, symptoms, and Centor score for the diagnosis of Group A strep versus non-streptococcal pharyngitis in all included studies Sign/ Symptom/ Centor Score GAS or Non-GAS Sensitivity (95% CI) % Diff GAS versus Non-GAS Specificity (95% CI) LR+ (95% CI) % Diff GAS versus Non-GAS LR- (95% CI) % Diff GAS versus Non-GAS AUROCC DOR (95% CI) % Diff GAS versus Non-GAS Signs and symptoms  Arthralgia or myalgia GAS 0.18 (0.06–0.44) −10% p = .885 0.87 (0.70–0.95) 1.42 (1.00–1.91) −13% 0.93 (0.78–1.00) +2% 0.60 1.55 (1.01–2.27) −15% Non-GAS 0.20 (0.08–0.42) 1.64 (1.10–2.38) 0.91 (0.78–0.98) 0.57 1.83 (1.13–2.81)  Cervical adenopathya GAS 0.82 (0.71–0.89) +17% p = .192 0.40 (0.23–0.61) 1.40 (1.12–1.89) +18% 0.46 (0.32–0.66) −39% 0.72 3.17 (1.74–5.32) +85% Non-GAS 0.70 (0.48–0.86) 1.19 (0.91–1.57) 0.76 (0.45–1.15) 0.57 1.71 (0.79–3.22)  Cervical adenopathyb GAS 0.72 (0.40–0.91) +7% p = .835 0.41 (0.26–0.58) 1.20 (0.74–1.64) +3% 0.72 (0.25–1.38) −9% 0.52 2.16 (0.56–5.94) +39% Non-GAS 0.67 (0.48–0.82) 1.16 (0.93–1.45) 0.79 (0.52–1.11) 0.56 1.55 (0.83–2.67)  Fever GAS 0.58 (0.42–0.73) −15% p = 0.433 0.46 (0.30–0.63) 1.09 (0.84–1.40) −13% 0.93 (0.67–1.26) +33% 0.53 1.22 (0.68–2.05) −35% Non-GAS 0.68 (0.44–0.85) 1.25 (1.06–1.44) 0.70 (0.46–0.94) 0.57 1.87 (1.16–2.91)  Injected throat GAS 0.86 (0.63–0.95) +2% p = .995 0.19 (0.06–0.43) 1.1 (0.93–1.26) +4% 0.80 (0.43–1.35) −8% 0.55 1.47 (0.69–2.74) +10% Non-GAS 0.84 (0.70–0.93) 1.06 (0.95–1.32) 0.87 (0.51–1.48) 0.66 1.34 (0.64–2.49)  Lack of cougha GAS 0.77 (0.63–0.87) +15% p = .266 0.40 (0.30–0.51) 1.28 (1.02–1.61) +13% 0.60 (0.33–0.97) −29% 0.58 2.35 (1.04–4.60) +68% Non-GAS 0.67 (0.60–0.74) 1.13 (0.93–1.38) 0.84 (0.61–1.15) 0.65 1.40 (0.82–2.25)  Lack of coughb GAS 0.75 (0.64–0.84) +27% p = 0.109 0.19 (0.03–0.67) 1.08 (0.70–2.53) +24% 2.39 (0.29–10.90) −54% 0.67 1.47 (0.06–7.51) +53% Non-GAS 0.59 (0.41–0.75) 0.87 (0.42–2.27) 5.25 (0.39–27.3) 0.47 0.96 (0.02–5.73)  Sore throat GAS 0.64 (0.46–0.79) +19% p = 0.522 0.45 (0.29–0.62) 1.17 (1.05–1.32) +18% 0.80 (0.65–0.94) −18% 0.55 1.47 (1.13–1.89) +41% Non-GAS 0.54 (0.29–0.78) 0.99 (0.74–1.17) 0.98 (0.70–1.20) 0.49 1.04 (0.62–1.63)  Tonsillar enlargement GAS 0.62 (0.49–0.74) +2% p = 0.823 0.52 (0.20–0.83) 1.47 (0.87–3.07) +18% 0.80 (0.52–1.50) +8% 0.62 2.08 (0.58–5.35) +14% Non-GAS 0.61 (0.20–0.91) 1.25 (0.94–1.65) 0.74 (0.39–1.03) 0.58 1.82 (0.91–3.28)  Tonsillar exudate GAS 0.38 (0.27–0.51) −12% p = 0.578 0.74 (0.64–0.83) 1.53 (1.00–2.24) −11% 0.83 (0.67–1.00) +8% 0.59 1.89 (0.99–3.27) −17% Non-GAS 0.43 (0.33–0.53) 1.71 (1.13–2.51) 0.77 (0.63–0.94) 0.57 2.27 (1.22–3.86) Centor score  Centor score 2 or higher GAS 0.93 (0.72–0.99) +3% p = 0.652 0.11 (0.01–0.63) 1.15 (0.99–1.93) +10% 0.70 (0.33–1.60) −36% 0.74 2.05 (0.62–5.06) +77% Non-GAS 0.90 (0.51–0.99) 1.05 (0.92–1.50) 1.09 (0.50–2.41) 0.59 1.16 (0.40–2.65)  Centor score 3 or higher GAS 0.65 (0.41–0.83) +44% p = 0.191 0.71 (0.48–0.86) 2.30 (1.35–4.05) +40% 0.51 (0.28–0.77) −35% 0.73 4.92 (1.87–10.60) +129% Non-GAS 0.45 (0.32–0.59) 1.64 (1.00–2.80) 0.79 (0.64–1.00) 0.54 2.15 (1.00–4.05) Sign/ Symptom/ Centor Score GAS or Non-GAS Sensitivity (95% CI) % Diff GAS versus Non-GAS Specificity (95% CI) LR+ (95% CI) % Diff GAS versus Non-GAS LR- (95% CI) % Diff GAS versus Non-GAS AUROCC DOR (95% CI) % Diff GAS versus Non-GAS Signs and symptoms  Arthralgia or myalgia GAS 0.18 (0.06–0.44) −10% p = .885 0.87 (0.70–0.95) 1.42 (1.00–1.91) −13% 0.93 (0.78–1.00) +2% 0.60 1.55 (1.01–2.27) −15% Non-GAS 0.20 (0.08–0.42) 1.64 (1.10–2.38) 0.91 (0.78–0.98) 0.57 1.83 (1.13–2.81)  Cervical adenopathya GAS 0.82 (0.71–0.89) +17% p = .192 0.40 (0.23–0.61) 1.40 (1.12–1.89) +18% 0.46 (0.32–0.66) −39% 0.72 3.17 (1.74–5.32) +85% Non-GAS 0.70 (0.48–0.86) 1.19 (0.91–1.57) 0.76 (0.45–1.15) 0.57 1.71 (0.79–3.22)  Cervical adenopathyb GAS 0.72 (0.40–0.91) +7% p = .835 0.41 (0.26–0.58) 1.20 (0.74–1.64) +3% 0.72 (0.25–1.38) −9% 0.52 2.16 (0.56–5.94) +39% Non-GAS 0.67 (0.48–0.82) 1.16 (0.93–1.45) 0.79 (0.52–1.11) 0.56 1.55 (0.83–2.67)  Fever GAS 0.58 (0.42–0.73) −15% p = 0.433 0.46 (0.30–0.63) 1.09 (0.84–1.40) −13% 0.93 (0.67–1.26) +33% 0.53 1.22 (0.68–2.05) −35% Non-GAS 0.68 (0.44–0.85) 1.25 (1.06–1.44) 0.70 (0.46–0.94) 0.57 1.87 (1.16–2.91)  Injected throat GAS 0.86 (0.63–0.95) +2% p = .995 0.19 (0.06–0.43) 1.1 (0.93–1.26) +4% 0.80 (0.43–1.35) −8% 0.55 1.47 (0.69–2.74) +10% Non-GAS 0.84 (0.70–0.93) 1.06 (0.95–1.32) 0.87 (0.51–1.48) 0.66 1.34 (0.64–2.49)  Lack of cougha GAS 0.77 (0.63–0.87) +15% p = .266 0.40 (0.30–0.51) 1.28 (1.02–1.61) +13% 0.60 (0.33–0.97) −29% 0.58 2.35 (1.04–4.60) +68% Non-GAS 0.67 (0.60–0.74) 1.13 (0.93–1.38) 0.84 (0.61–1.15) 0.65 1.40 (0.82–2.25)  Lack of coughb GAS 0.75 (0.64–0.84) +27% p = 0.109 0.19 (0.03–0.67) 1.08 (0.70–2.53) +24% 2.39 (0.29–10.90) −54% 0.67 1.47 (0.06–7.51) +53% Non-GAS 0.59 (0.41–0.75) 0.87 (0.42–2.27) 5.25 (0.39–27.3) 0.47 0.96 (0.02–5.73)  Sore throat GAS 0.64 (0.46–0.79) +19% p = 0.522 0.45 (0.29–0.62) 1.17 (1.05–1.32) +18% 0.80 (0.65–0.94) −18% 0.55 1.47 (1.13–1.89) +41% Non-GAS 0.54 (0.29–0.78) 0.99 (0.74–1.17) 0.98 (0.70–1.20) 0.49 1.04 (0.62–1.63)  Tonsillar enlargement GAS 0.62 (0.49–0.74) +2% p = 0.823 0.52 (0.20–0.83) 1.47 (0.87–3.07) +18% 0.80 (0.52–1.50) +8% 0.62 2.08 (0.58–5.35) +14% Non-GAS 0.61 (0.20–0.91) 1.25 (0.94–1.65) 0.74 (0.39–1.03) 0.58 1.82 (0.91–3.28)  Tonsillar exudate GAS 0.38 (0.27–0.51) −12% p = 0.578 0.74 (0.64–0.83) 1.53 (1.00–2.24) −11% 0.83 (0.67–1.00) +8% 0.59 1.89 (0.99–3.27) −17% Non-GAS 0.43 (0.33–0.53) 1.71 (1.13–2.51) 0.77 (0.63–0.94) 0.57 2.27 (1.22–3.86) Centor score  Centor score 2 or higher GAS 0.93 (0.72–0.99) +3% p = 0.652 0.11 (0.01–0.63) 1.15 (0.99–1.93) +10% 0.70 (0.33–1.60) −36% 0.74 2.05 (0.62–5.06) +77% Non-GAS 0.90 (0.51–0.99) 1.05 (0.92–1.50) 1.09 (0.50–2.41) 0.59 1.16 (0.40–2.65)  Centor score 3 or higher GAS 0.65 (0.41–0.83) +44% p = 0.191 0.71 (0.48–0.86) 2.30 (1.35–4.05) +40% 0.51 (0.28–0.77) −35% 0.73 4.92 (1.87–10.60) +129% Non-GAS 0.45 (0.32–0.59) 1.64 (1.00–2.80) 0.79 (0.64–1.00) 0.54 2.15 (1.00–4.05) aExcluding Principi’s study as an outlier due to its very low sensitivity in cervical adenopathy in Group A strep (6%) and its very low specificity in lack of cough (0%). bIncluding Principi’s study. View Large Table 2. Accuracy of signs, symptoms, and Centor score for the diagnosis of Group A strep versus non-streptococcal pharyngitis in all included studies Sign/ Symptom/ Centor Score GAS or Non-GAS Sensitivity (95% CI) % Diff GAS versus Non-GAS Specificity (95% CI) LR+ (95% CI) % Diff GAS versus Non-GAS LR- (95% CI) % Diff GAS versus Non-GAS AUROCC DOR (95% CI) % Diff GAS versus Non-GAS Signs and symptoms  Arthralgia or myalgia GAS 0.18 (0.06–0.44) −10% p = .885 0.87 (0.70–0.95) 1.42 (1.00–1.91) −13% 0.93 (0.78–1.00) +2% 0.60 1.55 (1.01–2.27) −15% Non-GAS 0.20 (0.08–0.42) 1.64 (1.10–2.38) 0.91 (0.78–0.98) 0.57 1.83 (1.13–2.81)  Cervical adenopathya GAS 0.82 (0.71–0.89) +17% p = .192 0.40 (0.23–0.61) 1.40 (1.12–1.89) +18% 0.46 (0.32–0.66) −39% 0.72 3.17 (1.74–5.32) +85% Non-GAS 0.70 (0.48–0.86) 1.19 (0.91–1.57) 0.76 (0.45–1.15) 0.57 1.71 (0.79–3.22)  Cervical adenopathyb GAS 0.72 (0.40–0.91) +7% p = .835 0.41 (0.26–0.58) 1.20 (0.74–1.64) +3% 0.72 (0.25–1.38) −9% 0.52 2.16 (0.56–5.94) +39% Non-GAS 0.67 (0.48–0.82) 1.16 (0.93–1.45) 0.79 (0.52–1.11) 0.56 1.55 (0.83–2.67)  Fever GAS 0.58 (0.42–0.73) −15% p = 0.433 0.46 (0.30–0.63) 1.09 (0.84–1.40) −13% 0.93 (0.67–1.26) +33% 0.53 1.22 (0.68–2.05) −35% Non-GAS 0.68 (0.44–0.85) 1.25 (1.06–1.44) 0.70 (0.46–0.94) 0.57 1.87 (1.16–2.91)  Injected throat GAS 0.86 (0.63–0.95) +2% p = .995 0.19 (0.06–0.43) 1.1 (0.93–1.26) +4% 0.80 (0.43–1.35) −8% 0.55 1.47 (0.69–2.74) +10% Non-GAS 0.84 (0.70–0.93) 1.06 (0.95–1.32) 0.87 (0.51–1.48) 0.66 1.34 (0.64–2.49)  Lack of cougha GAS 0.77 (0.63–0.87) +15% p = .266 0.40 (0.30–0.51) 1.28 (1.02–1.61) +13% 0.60 (0.33–0.97) −29% 0.58 2.35 (1.04–4.60) +68% Non-GAS 0.67 (0.60–0.74) 1.13 (0.93–1.38) 0.84 (0.61–1.15) 0.65 1.40 (0.82–2.25)  Lack of coughb GAS 0.75 (0.64–0.84) +27% p = 0.109 0.19 (0.03–0.67) 1.08 (0.70–2.53) +24% 2.39 (0.29–10.90) −54% 0.67 1.47 (0.06–7.51) +53% Non-GAS 0.59 (0.41–0.75) 0.87 (0.42–2.27) 5.25 (0.39–27.3) 0.47 0.96 (0.02–5.73)  Sore throat GAS 0.64 (0.46–0.79) +19% p = 0.522 0.45 (0.29–0.62) 1.17 (1.05–1.32) +18% 0.80 (0.65–0.94) −18% 0.55 1.47 (1.13–1.89) +41% Non-GAS 0.54 (0.29–0.78) 0.99 (0.74–1.17) 0.98 (0.70–1.20) 0.49 1.04 (0.62–1.63)  Tonsillar enlargement GAS 0.62 (0.49–0.74) +2% p = 0.823 0.52 (0.20–0.83) 1.47 (0.87–3.07) +18% 0.80 (0.52–1.50) +8% 0.62 2.08 (0.58–5.35) +14% Non-GAS 0.61 (0.20–0.91) 1.25 (0.94–1.65) 0.74 (0.39–1.03) 0.58 1.82 (0.91–3.28)  Tonsillar exudate GAS 0.38 (0.27–0.51) −12% p = 0.578 0.74 (0.64–0.83) 1.53 (1.00–2.24) −11% 0.83 (0.67–1.00) +8% 0.59 1.89 (0.99–3.27) −17% Non-GAS 0.43 (0.33–0.53) 1.71 (1.13–2.51) 0.77 (0.63–0.94) 0.57 2.27 (1.22–3.86) Centor score  Centor score 2 or higher GAS 0.93 (0.72–0.99) +3% p = 0.652 0.11 (0.01–0.63) 1.15 (0.99–1.93) +10% 0.70 (0.33–1.60) −36% 0.74 2.05 (0.62–5.06) +77% Non-GAS 0.90 (0.51–0.99) 1.05 (0.92–1.50) 1.09 (0.50–2.41) 0.59 1.16 (0.40–2.65)  Centor score 3 or higher GAS 0.65 (0.41–0.83) +44% p = 0.191 0.71 (0.48–0.86) 2.30 (1.35–4.05) +40% 0.51 (0.28–0.77) −35% 0.73 4.92 (1.87–10.60) +129% Non-GAS 0.45 (0.32–0.59) 1.64 (1.00–2.80) 0.79 (0.64–1.00) 0.54 2.15 (1.00–4.05) Sign/ Symptom/ Centor Score GAS or Non-GAS Sensitivity (95% CI) % Diff GAS versus Non-GAS Specificity (95% CI) LR+ (95% CI) % Diff GAS versus Non-GAS LR- (95% CI) % Diff GAS versus Non-GAS AUROCC DOR (95% CI) % Diff GAS versus Non-GAS Signs and symptoms  Arthralgia or myalgia GAS 0.18 (0.06–0.44) −10% p = .885 0.87 (0.70–0.95) 1.42 (1.00–1.91) −13% 0.93 (0.78–1.00) +2% 0.60 1.55 (1.01–2.27) −15% Non-GAS 0.20 (0.08–0.42) 1.64 (1.10–2.38) 0.91 (0.78–0.98) 0.57 1.83 (1.13–2.81)  Cervical adenopathya GAS 0.82 (0.71–0.89) +17% p = .192 0.40 (0.23–0.61) 1.40 (1.12–1.89) +18% 0.46 (0.32–0.66) −39% 0.72 3.17 (1.74–5.32) +85% Non-GAS 0.70 (0.48–0.86) 1.19 (0.91–1.57) 0.76 (0.45–1.15) 0.57 1.71 (0.79–3.22)  Cervical adenopathyb GAS 0.72 (0.40–0.91) +7% p = .835 0.41 (0.26–0.58) 1.20 (0.74–1.64) +3% 0.72 (0.25–1.38) −9% 0.52 2.16 (0.56–5.94) +39% Non-GAS 0.67 (0.48–0.82) 1.16 (0.93–1.45) 0.79 (0.52–1.11) 0.56 1.55 (0.83–2.67)  Fever GAS 0.58 (0.42–0.73) −15% p = 0.433 0.46 (0.30–0.63) 1.09 (0.84–1.40) −13% 0.93 (0.67–1.26) +33% 0.53 1.22 (0.68–2.05) −35% Non-GAS 0.68 (0.44–0.85) 1.25 (1.06–1.44) 0.70 (0.46–0.94) 0.57 1.87 (1.16–2.91)  Injected throat GAS 0.86 (0.63–0.95) +2% p = .995 0.19 (0.06–0.43) 1.1 (0.93–1.26) +4% 0.80 (0.43–1.35) −8% 0.55 1.47 (0.69–2.74) +10% Non-GAS 0.84 (0.70–0.93) 1.06 (0.95–1.32) 0.87 (0.51–1.48) 0.66 1.34 (0.64–2.49)  Lack of cougha GAS 0.77 (0.63–0.87) +15% p = .266 0.40 (0.30–0.51) 1.28 (1.02–1.61) +13% 0.60 (0.33–0.97) −29% 0.58 2.35 (1.04–4.60) +68% Non-GAS 0.67 (0.60–0.74) 1.13 (0.93–1.38) 0.84 (0.61–1.15) 0.65 1.40 (0.82–2.25)  Lack of coughb GAS 0.75 (0.64–0.84) +27% p = 0.109 0.19 (0.03–0.67) 1.08 (0.70–2.53) +24% 2.39 (0.29–10.90) −54% 0.67 1.47 (0.06–7.51) +53% Non-GAS 0.59 (0.41–0.75) 0.87 (0.42–2.27) 5.25 (0.39–27.3) 0.47 0.96 (0.02–5.73)  Sore throat GAS 0.64 (0.46–0.79) +19% p = 0.522 0.45 (0.29–0.62) 1.17 (1.05–1.32) +18% 0.80 (0.65–0.94) −18% 0.55 1.47 (1.13–1.89) +41% Non-GAS 0.54 (0.29–0.78) 0.99 (0.74–1.17) 0.98 (0.70–1.20) 0.49 1.04 (0.62–1.63)  Tonsillar enlargement GAS 0.62 (0.49–0.74) +2% p = 0.823 0.52 (0.20–0.83) 1.47 (0.87–3.07) +18% 0.80 (0.52–1.50) +8% 0.62 2.08 (0.58–5.35) +14% Non-GAS 0.61 (0.20–0.91) 1.25 (0.94–1.65) 0.74 (0.39–1.03) 0.58 1.82 (0.91–3.28)  Tonsillar exudate GAS 0.38 (0.27–0.51) −12% p = 0.578 0.74 (0.64–0.83) 1.53 (1.00–2.24) −11% 0.83 (0.67–1.00) +8% 0.59 1.89 (0.99–3.27) −17% Non-GAS 0.43 (0.33–0.53) 1.71 (1.13–2.51) 0.77 (0.63–0.94) 0.57 2.27 (1.22–3.86) Centor score  Centor score 2 or higher GAS 0.93 (0.72–0.99) +3% p = 0.652 0.11 (0.01–0.63) 1.15 (0.99–1.93) +10% 0.70 (0.33–1.60) −36% 0.74 2.05 (0.62–5.06) +77% Non-GAS 0.90 (0.51–0.99) 1.05 (0.92–1.50) 1.09 (0.50–2.41) 0.59 1.16 (0.40–2.65)  Centor score 3 or higher GAS 0.65 (0.41–0.83) +44% p = 0.191 0.71 (0.48–0.86) 2.30 (1.35–4.05) +40% 0.51 (0.28–0.77) −35% 0.73 4.92 (1.87–10.60) +129% Non-GAS 0.45 (0.32–0.59) 1.64 (1.00–2.80) 0.79 (0.64–1.00) 0.54 2.15 (1.00–4.05) aExcluding Principi’s study as an outlier due to its very low sensitivity in cervical adenopathy in Group A strep (6%) and its very low specificity in lack of cough (0%). bIncluding Principi’s study. View Large Summary sensitivities of signs and symptoms for GAS pharyngitis and non-GAS pharyngitis are presented in Figure 1. In both GAS and non-GAS infections, the two most sensitive signs and symptoms were injected throat (0.86 versus 0.84) and cervical adenopathy (0.82 versus 0.70). The sensitivities of injected throat (0.86 versus 0.84) and tonsillar enlargement (0.62 versus 0.61) were nearly identical. All but three signs and symptoms had point estimates for sensitivity within 15% of each other, and the 95% confidence intervals for all signs and symptoms overlapped between GAS and non-GAS pharyngitis vs non-streptococcal pharyngitis. A larger difference was seen for lack of cough, but this was largely due to a single outlier study (31). When the outlier was excluded, the relative difference in sensitivities was 15%. Figure 1. View largeDownload slide Summary ROC curve sensitivities of signs, symptoms, the Centor score 2 or higher (CS 2+), and Centor score 3 or higher (CS 3+) in Group A strep and non-Group A strep (two boundary lines show the limit of 15% difference among two groups). Figure 1. View largeDownload slide Summary ROC curve sensitivities of signs, symptoms, the Centor score 2 or higher (CS 2+), and Centor score 3 or higher (CS 3+) in Group A strep and non-Group A strep (two boundary lines show the limit of 15% difference among two groups). Because we compared signs and symptoms of GAS and non-GAS pharyngitis in the same population against non-streptococcal infections, the specificities were the same for both GAS and non-GAS pharyngitis. The sign or symptom having the highest specificity was arthralgia or myalgia (0.87), while the one with lowest specificity was injected throat (0.19). No individual sign or symptom for either GAS or non-GAS pharyngitis had a LR+ greater than 2.0. Tonsillar exudate had the highest LR+ in both GAS pharyngitis (1.5) and in non-GAS pharyngitis (1.7). Only cervical adenopathy, sore throat, and tonsillar enlargement had more than a 15% difference in LR+. Five out eight signs and symptoms had small differences in LR− between GAS and non-GAS pharyngitis; larger relative differences were seen for cervical adenopathy, fever, and lack of cough with the percentage of difference −39%, +33%, and −29%, respectively. However, the confidence intervals of LR+ and LR- for all signs and symptoms between GAS and non-GAS infections largely overlapped. Cervical adenopathy was fairly good at discriminating GAS infection from non-streptococcal infection (AUROCC 0.72), but it was much less accurate in predicting non-GAS infection (AUROCC 0.57). Injected throat had the highest AUROCC (0.66) in non-GAS infection, but its AUROCC was the second lowest (0.55) in GAS infection. Beside cervical adenopathy, AUROCC of other signs and symptoms between GAS and non-GAS infection had limited differences. Summary ROC curves for selected signs and symptoms are shown in Figure 2 and in Supplementary Figures S2–S6. The sensitivities were similar for arthralgia or myalgia, while the sensitivities of cervical adenopathy, lack of a cough, and sore throat for GAS infection were consistently higher than for non-GAS infection. Sensitivities of arthralgia or myalgia, cervical adenopathy, and lack of cough in children were generally lower than those observed in studies of both children and adult patients. Figure 2. View largeDownload slide Summary ROC curve for cervical adenopathy without outlier (a), fever (b), lack of a cough with outlier (c), and tonsillar exudate (d) in Group A strep and non-Group A strep. Figure 2. View largeDownload slide Summary ROC curve for cervical adenopathy without outlier (a), fever (b), lack of a cough with outlier (c), and tonsillar exudate (d) in Group A strep and non-Group A strep. Accuracy of Centor score The accuracy of the Centor score for both GAS and non-GAS infections is shown in Table 2 (a detailed table with data of individual studies is shown in Supplementary Table S3). For a Centor score of 2 or higher versus score less than 2, sensitivity and LR+ were similar between GAS and non-GAS infections, while LR− and DOR had some small differences. For a Centor score of 3 or higher versus less than 3, there were larger differences between GAS and non-GAS regarding sensitivity, LR+, LR−, and DOR. Again, though, the confidence intervals for these parameters for GAS pharyngitis largely overlapped those for non-GAS infections. The area under the ROC curve was greater for GAS pharyngitis than for non-GAS pharyngitis for both the Centor score of 2 or higher and the Centor score of 3 or higher. Summary ROC curves for the Centor score 2 or higher and Centor score 3 or higher are shown in Supplementary Figures S7 and S8. In each individual study, the sensitivity of the Centor score was higher for GAS infection than for non-GAS infection. Discussion In general, individual signs and symptoms have limited accuracy for the diagnosis of both GAS and non-GAS pharyngitis. As seen most clearly in Figure 1, the sensitivity of each sign and symptoms was generally similar for patients with GAS and non-GAS pharyngitis, with the largest differences in sensitivity seen for cervical adenopathy, lack of cough, and fever. The sensitivity of signs and symptoms was higher for the diagnosis of GAS pharyngitis than for non-GAS pharyngitis. The general similarity in clinical presentation between GAS and non-GAS pharyngitis provides some support for the idea that non-GAS may be a pathogen in patients with pharyngitis. As shown in Table 3, GAS is significantly more common in symptomatic persons than in asymptomatic controls (35–39). While non-GAS was also present more often in symptomatic patients than in asymptomatic controls in four of five studies, this difference was statistically significant in only one of these studies (35–39). Pooling the data from the five studies shown in Table 3, non-GAS had a prevalence of 5% in symptomatic patients and 2% in asymptomatic controls (P = 0.38). However, these results were dominated by a very large study that found no difference in prevalence, but that only enrolled patients retrospectively who had at least four of the following: sore throat, erythema of tonsils, exudate on tonsils, painful cervical lymph nodes, and fever, creating an important selection bias (37). Table 3. Prevalence of Group A strep and non-Group A strep in symptomatic and asymptomatic groups Study GAS Non-GAS Symptomatic group Asymptomatic group p value Symptomatic group Asymptomatic group p value Hedin, 2015 (35) 0.30 (66/220) 0.02 (3/128) <0.001 0.04 (8/220) 0.01 (1/128) 0.16 Centor, 2015 (36) 0.07 (21/312) 0.01 (2/180) 0.009a 0.06 (19/312) 0.03 (6/180) 0.26a Begovac, 1993 (37) 0.45 (281/629) 0.06 (107/1796) <0.001 0.01 (9/629) 0.01 (16/1796) 0.36a Hayden, 1989 (38) 0.39 (58/150) 0.16 (24/150) <0.001 0.17 (25/150) 0.21 (32/150) >0.05 Hofkosh, 1988 (39) 0.20 (189/929) 0.05 (19/414) <0.01 0.06 (60/929) 0.01 (4/414) <0.01 Pooled estimate (95% CI) 0.25 (0.15–0.40) 0.05 (0.03–0.10) <0.001 0.05 (0.03–0.10) 0.02 (0.00–0.13) 0.38 Study GAS Non-GAS Symptomatic group Asymptomatic group p value Symptomatic group Asymptomatic group p value Hedin, 2015 (35) 0.30 (66/220) 0.02 (3/128) <0.001 0.04 (8/220) 0.01 (1/128) 0.16 Centor, 2015 (36) 0.07 (21/312) 0.01 (2/180) 0.009a 0.06 (19/312) 0.03 (6/180) 0.26a Begovac, 1993 (37) 0.45 (281/629) 0.06 (107/1796) <0.001 0.01 (9/629) 0.01 (16/1796) 0.36a Hayden, 1989 (38) 0.39 (58/150) 0.16 (24/150) <0.001 0.17 (25/150) 0.21 (32/150) >0.05 Hofkosh, 1988 (39) 0.20 (189/929) 0.05 (19/414) <0.01 0.06 (60/929) 0.01 (4/414) <0.01 Pooled estimate (95% CI) 0.25 (0.15–0.40) 0.05 (0.03–0.10) <0.001 0.05 (0.03–0.10) 0.02 (0.00–0.13) 0.38 aNot reported by the original study View Large Table 3. Prevalence of Group A strep and non-Group A strep in symptomatic and asymptomatic groups Study GAS Non-GAS Symptomatic group Asymptomatic group p value Symptomatic group Asymptomatic group p value Hedin, 2015 (35) 0.30 (66/220) 0.02 (3/128) <0.001 0.04 (8/220) 0.01 (1/128) 0.16 Centor, 2015 (36) 0.07 (21/312) 0.01 (2/180) 0.009a 0.06 (19/312) 0.03 (6/180) 0.26a Begovac, 1993 (37) 0.45 (281/629) 0.06 (107/1796) <0.001 0.01 (9/629) 0.01 (16/1796) 0.36a Hayden, 1989 (38) 0.39 (58/150) 0.16 (24/150) <0.001 0.17 (25/150) 0.21 (32/150) >0.05 Hofkosh, 1988 (39) 0.20 (189/929) 0.05 (19/414) <0.01 0.06 (60/929) 0.01 (4/414) <0.01 Pooled estimate (95% CI) 0.25 (0.15–0.40) 0.05 (0.03–0.10) <0.001 0.05 (0.03–0.10) 0.02 (0.00–0.13) 0.38 Study GAS Non-GAS Symptomatic group Asymptomatic group p value Symptomatic group Asymptomatic group p value Hedin, 2015 (35) 0.30 (66/220) 0.02 (3/128) <0.001 0.04 (8/220) 0.01 (1/128) 0.16 Centor, 2015 (36) 0.07 (21/312) 0.01 (2/180) 0.009a 0.06 (19/312) 0.03 (6/180) 0.26a Begovac, 1993 (37) 0.45 (281/629) 0.06 (107/1796) <0.001 0.01 (9/629) 0.01 (16/1796) 0.36a Hayden, 1989 (38) 0.39 (58/150) 0.16 (24/150) <0.001 0.17 (25/150) 0.21 (32/150) >0.05 Hofkosh, 1988 (39) 0.20 (189/929) 0.05 (19/414) <0.01 0.06 (60/929) 0.01 (4/414) <0.01 Pooled estimate (95% CI) 0.25 (0.15–0.40) 0.05 (0.03–0.10) <0.001 0.05 (0.03–0.10) 0.02 (0.00–0.13) 0.38 aNot reported by the original study View Large Antibiotic treatment for non-GAS pharyngitis is currently not recommended (15,40). However, some have argued that because the symptoms are similar, that patients with non-GAS pharyngitis (in particular group C strep) would benefit from antibiotic treatment (3,13,14,18). In fact, a clinical decision rule was recently developed that identifies patients with both Group A and Group C streptococcal pharyngitis to guide antibiotic treatment (41). At a minimum, clinicians should be aware of the possibility that a patient with a negative rapid strep test may have non-GAS pharyngitis and may benefit from treatment, especially if symptoms otherwise resemble GAS pharyngitis. Limitations There are several limitations in this review. Only high quality studies reporting signs, symptoms, or the Centor score for both GAS and non-GAS pharyngitis were selected, which improved validity but limited the number of studies included in our review. The comparison of signs, symptoms, and Centor score between GAS and non-GAS infections were not fully stratified and evaluated by age group, confidence intervals were relatively broad, and the summary ROC curves each had a relatively small number of studies. In addition, the definitions of signs and symptoms likely varied somewhat between the included studies. For example, a patient with temperature from 37.3 °C was considered as fever in one study (29), while another study used a cutoff of 38°C (19). There was significant heterogeneity for most of the signs, symptoms, and the Centor score, notably cervical adenopathy, fever, or the Centor score 3 or 4. Suggestions for future research First, the number of studies comparing signs and symptoms of GAS and non-GAS infections in the same population is very limited. More research should be conducted that compares clinical characteristics among GAS, GCS, GGS, and non-GAS infections. Second, there are many clinical decision rules for GAS pharyngitis or viral sore throat (21,42–46). Most of these rules were also validated in alternative studies (47–49). However, the development and validation of clinical decision rules for GCS is still limited (18,33), so further work is needed to prospectively validate them. Lastly, there is no clear evidence that treatment of non-GAS pharyngitis provides a net clinical benefit in terms of reduction in the duration of symptoms, decreased transmission, or fewer complications. Randomized controlled trials are needed to determine whether or not GCS pharyngitis should be treated. Supplementary Material Supplementary data are available at Family Practice online. Declaration Ethical approval: not necessary. Funding: departmental resources Conflict of interest: none. Reference 1. Worrall GJ . Acute sore throat . Can Fam Physician 2007 ; 53 : 1961 – 2 . Google Scholar PubMed 2. Fleming-Dutra KE , Hersh AL , Shapiro DJ et al. Prevalence of Inappropriate Antibiotic Prescriptions Among US Ambulatory Care Visits, 2010-2011 . JAMA 2016 ; 315 : 1864 – 73 . Google Scholar CrossRef Search ADS PubMed 3. Marchello C , Ebell MH . Prevalence of Group C Streptococcus and Fusobacterium Necrophorum in Patients With Sore Throat: A Meta-Analysis . Ann Fam Med 2016 ; 14 : 567 – 74 . Google Scholar CrossRef Search ADS PubMed 4. Efstratiou A . Pyogenic streptococci of Lancefield groups C and G as pathogens in man . J Appl Microbiol 1997 ; 83 ( S1 ): 72S – 9S . Google Scholar CrossRef Search ADS PubMed 5. Pelucchi C , Grigoryan L , Galeone C et al. Guideline for the management of acute sore throat . Clin Microbiol Infect 2012 ; 18 ( suppl 1 ): 1 – 28 . Google Scholar CrossRef Search ADS PubMed 6. Morgan MC , Rice LI . Recurrent group C streptococcal tonsillopharyngitis in an adolescent . J Adolesc Health Care 1989 ; 10 : 421 – 2 . Google Scholar CrossRef Search ADS PubMed 7. Fulginiti VA , Ey JL , Ryan KJ . Recurrent group C streptococcal tonsillitis in an adolescent male requiring tonsillectomy . Clin Pediatr (Phila) 1980 ; 19 : 829 – 30 . Google Scholar CrossRef Search ADS PubMed 8. Jansen TL , Janssen M , de Jong AJ . Reactive arthritis associated with group C and group G beta-hemolytic streptococci . J Rheumatol 1998 ; 25 : 1126 – 30 . Google Scholar PubMed 9. Young L , Deighton CM , Chuck AJ , Galloway A . Reactive arthritis and group G streptococcal pharyngitis . Ann Rheum Dis 1992 ; 51 : 1268 . Google Scholar CrossRef Search ADS PubMed 10. Almroth G , Lindell A , Aselius H et al. Acute glomerulonephritis associated with streptococcus pyogenes with concomitant spread of streptococcus constellatus in four rural families . Ups J Med Sci 2005 ; 110 : 217 – 31 . Google Scholar CrossRef Search ADS PubMed 11. Natoli S , Fimiani C , Faglieri N et al. Toxic shock syndrome due to group C streptococci. A case report . Intensive Care Med 1996 ; 22 : 985 – 9 . Google Scholar CrossRef Search ADS PubMed 12. Gettler JF , el-Sadr W . Group C streptococcal subdural empyema in a healthy man: possible complication of pharyngitis . Clin Infect Dis 1993 ; 16 : 726 – 7 . Google Scholar CrossRef Search ADS PubMed 13. Shah M , Centor RM , Jennings M . Severe acute pharyngitis caused by group C streptococcus . J Gen Intern Med 2007 ; 22 : 272 – 4 . Google Scholar CrossRef Search ADS PubMed 14. Zwart S , Sachs AP , Ruijs GJ , Gubbels JW , Hoes AW , de Melker RA . Penicillin for acute sore throat: randomised double blind trial of seven days versus three days treatment or placebo in adults . BMJ 2000 ; 320 : 150 – 4 . Google Scholar CrossRef Search ADS PubMed 15. Shulman ST , Bisno AL , Clegg HW et al. Clinical practice guideline for the diagnosis and management of group A streptococcal pharyngitis: 2012 update by the Infectious Diseases Society of America . Clin Infect Dis 2012 ; 55 : 1279 – 82 . Google Scholar CrossRef Search ADS PubMed 16. Cooper RJ , Hoffman JR , Bartlett JG et al. ; American Academy of Family Physicians; American College of Physicians-American Society of Internal Medicine; Centers for Disease Control . Principles of appropriate antibiotic use for acute pharyngitis in adults: background . Ann Intern Med 2001 ; 134 : 509 – 17 . Google Scholar CrossRef Search ADS PubMed 17. Matthys J , De Meyere M , van Driel ML , De Sutter A . Differences among international pharyngitis guidelines: not just academic . Ann Fam Med 2007 ; 5 : 436 – 43 . Google Scholar CrossRef Search ADS PubMed 18. Little P , Hobbs FD , Mant D , McNulty CA , Mullee M ; PRISM investigators . Incidence and clinical variables associated with streptococcal throat infections: a prospective diagnostic cohort study . Br J Gen Pract 2012 ; 62 : e787 – 94 . Google Scholar CrossRef Search ADS PubMed 19. Fretzayas A , Moustaki M , Kitsiou S , Nychtari G , Nicolaidou P . The clinical pattern of group C streptococcal pharyngitis in children . J Infect Chemother 2009 ; 15 : 228 – 32 . Google Scholar CrossRef Search ADS PubMed 20. Meier FA , Centor RM , Graham L Jr , Dalton HP . Clinical and microbiological evidence for endemic pharyngitis among adults due to group C streptococci . Arch Intern Med 1990 ; 150 : 825 – 9 . Google Scholar CrossRef Search ADS PubMed 21. Centor RM , Witherspoon JM , Dalton HP , Brody CE , Link K . The diagnosis of strep throat in adults in the emergency room . Med Decis Making 1981 ; 1 : 239 – 46 . Google Scholar CrossRef Search ADS PubMed 22. Whiting PF , Rutjes AW , Westwood ME et al. ; QUADAS-2 Group . QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies . Ann Intern Med 2011 ; 155 : 529 – 36 . Google Scholar CrossRef Search ADS PubMed 23. R Development Core Team . R: A language and environment for statistical computing . 2016 . R Foundation for Statistical Computing : Vienna, Austria . 24. Doebler P . Mada (R package): Meta-analysis of Diagnostic Accuracy, version 0.5.7. 2015. https://www.r-project.org/ (accessed on 15 June 2017) . 25. Graffelman J . Calibrate (R package): Calibration of Scatterplot and Biplot Axes, version 1.7.2. 2013. https://www.r-project.org/ (accessed on 15 June 2017) . 26. Irwig L , Macaskill P , Glasziou P , Fahey M . Meta-analytic methods for diagnostic test accuracy . J Clin Epidemiol 1995 ; 48 : 119 – 30; discussion 131 –2. Google Scholar CrossRef Search ADS PubMed 27. Schwarzer G . Meta (R package): General Package for Meta-Analysis, version 4.5-0. 2016. https://www.r-project.org/ (accessed on 15 June 2017) . 28. Atlas SJ , McDermott SM , Mannone C , Barry MJ . BRIEF REPORT: The role of point of care testing for patients with acute pharyngitis . J Gen Intern Med 2005 ; 20 : 759 – 61 . Google Scholar CrossRef Search ADS PubMed 29. Bassili A , Barakat S , Sawaf GE , Zaher S , Zaki A , Din Saleh EE . Identification of clinical criteria for group A-beta hemolytic streptococcal pharyngitis in children living in a rheumatic fever endemic area . J Trop Pediatr 2002 ; 48 : 285 – 93 . Google Scholar CrossRef Search ADS PubMed 30. Kaplan EL , Top FH Jr, Dudding BA , Wannamaker LW . Diagnosis of streptococcal pharyngitis: differentiation of active infection from the carrier state in the symptomatic child . J Infect Dis 1971 ; 123 : 490 – 501 . Google Scholar CrossRef Search ADS PubMed 31. Principi N , Marchisio P , Calanchi A et al. Streptococcal pharyngitis in Italian children: epidemiology and treatment with miocamycin . Drugs Exp Clin Res 1990 ; 16 : 639 – 47 . Google Scholar PubMed 32. Seppälä H , Lahtonen R , Ziegler T et al. Clinical scoring system in the evaluation of adult pharyngitis . Arch Otolaryngol Head Neck Surg 1993 ; 119 : 288 – 91 . Google Scholar CrossRef Search ADS PubMed 33. Lindbaek M , Høiby EA , Lermark G et al. Clinical symptoms and signs in sore throat patients with large colony variant β-haemolytic streptococci groups C or G versus group A . The British Journal of General Practice 2005 ; 55 : 615 – 19 . Google Scholar PubMed 34. Llor C , Hernández Anadón S , Gómez Bertomeu FF , Santamaria Puig JM , Calviño Domínguez O , Fernández Pagés Y . [Validation of a rapid antigenic test in the diagnosis of pharyngitis caused by group a beta-haemolytic Streptococcus] . Aten Primaria 2008 ; 40 : 489 – 94 . Google Scholar CrossRef Search ADS PubMed 35. Hedin K , Bieber L , Lindh M , Sundqvist M . The aetiology of pharyngotonsillitis in adolescents and adults - Fusobacterium necrophorum is commonly found . Clin Microbiol Infect 2015 ; 21 : 263.e1 – 7 . Google Scholar CrossRef Search ADS 36. Centor RM , Atkinson TP , Ratliff AE et al. The clinical presentation of Fusobacterium-positive and streptococcal-positive pharyngitis in a university health clinic: a cross-sectional study . Ann Intern Med 2015 ; 162 : 241 – 7 . Google Scholar CrossRef Search ADS PubMed 37. Begovac J , Bobinac E , Benic B et al. Asymptomatic pharyngeal carriage of beta-haemolytic streptococci and streptococcal pharyngitis among patients at an urban hospital in Croatia . Eur J Epidemiol 1993 ; 9 : 405 – 10 . Google Scholar CrossRef Search ADS PubMed 38. Hayden GF , Murphy TF , Hendley JO . Non-group A streptococci in the pharynx. Pathogens or innocent bystanders ? Am J Dis Child 1989 ; 143 : 794 – 7 . Google Scholar CrossRef Search ADS PubMed 39. Hofkosh D , Wald ER , Chiponis DM . Prevalence of non-group-A beta-hemolytic streptococci in childhood pharyngitis . South Med J 1988 ; 81 : 329 – 31 . Google Scholar CrossRef Search ADS PubMed 40. National Institute for Health and Care Excellence . NICE guidelines: respiratory tract infections (self-limiting): prescribing antibiotics . 2008 [cited 2017 May 31]; Available from: https://www.nice.org.uk/guidance/cg69. 41. Little P , Hobbs FD , Mant D , McNulty CA , Mullee M ; PRISM investigators . Incidence and clinical variables associated with streptococcal throat infections: a prospective diagnostic cohort study . Br J Gen Pract 2012 ; 62 : e787 – 94 . Google Scholar CrossRef Search ADS PubMed 42. Breese BB . A simple scorecard for the tentative diagnosis of streptococcal pharyngitis . Am J Dis Child 1977 ; 131 : 514 – 7 . Google Scholar PubMed 43. McIsaac WJ , White D , Tannenbaum D , Low DE . A clinical score to reduce unnecessary antibiotic use in patients with sore throat . CMAJ 1998 ; 158 : 75 – 83 . Google Scholar PubMed 44. Attia MW , Zaoutis T , Klein JD , Meier FA . Performance of a predictive model for streptococcal pharyngitis in children . Arch Pediatr Adolesc Med 2001 ; 155 : 687 – 91 . Google Scholar CrossRef Search ADS PubMed 45. Joachim L , Campos D Jr , Smeesters PR . Pragmatic scoring system for pharyngitis in low-resource settings . Pediatrics 2010 ; 126 : e608 – 14 . Google Scholar CrossRef Search ADS PubMed 46. Mistik S , Gokahmetoglu S , Balci E , Onuk FA . Sore throat in primary care project: a clinical score to diagnose viral sore throat . Fam Pract 2015 ; 32 : 263 – 8 . Google Scholar CrossRef Search ADS PubMed 47. Chiappini E , Principi N , Mansi N et al. Management of acute pharyngitis in children: summary of the Italian National Institute of Health guidelines . Clin Ther 2012 ; 34 : 1442 – 1458.e2 . Google Scholar CrossRef Search ADS PubMed 48. Karacan M , Karakelleoğlu C , Orbak Z . Diagnosis of group A beta-hemolytic Streptococcus using the Breese clinical scoring system . South Med J 2007 ; 100 : 1192 – 7 . Google Scholar CrossRef Search ADS PubMed 49. Tanz RR , Gerber MA , Kabat W , Rippe J , Seshadri R , Shulman ST . Performance of a rapid antigen-detection test and throat culture in community pediatric offices: implications for management of pharyngitis . Pediatrics 2009 ; 123 : 437 – 44 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Herrmann, Wolfram J;Haarmann, Alexander;Bærheim, Anders

2017 Family Practice

doi: 10.1093/fampra/cmx092pmid: 28973219

Abstract Background Copayments are implemented in many health care systems. The effect of copayments differs between countries. Up to now, patients’ attitudes regarding copayments are mainly unknown. Objectives Thus, the goal of our analysis was to explore adult patients’ attitudes in Germany and Norway towards copayments as a steering tool. Methods We conducted a qualitative comparative study. Episodic interviews were conducted with 40 patients in Germany and Norway. The interviews were analysed by thematic coding in the framework of grounded theory. All text segments related to copayments were analysed in depth for emerging topics and types. Results We found three dimensions of patients’ attitudes towards copayments: the perceived steering effect, the comprehensibility, and the assessment of copayments. The perceived steering effect consists of three types: having been influenced by copayments, not having experienced any influence and the experience of other persons to be influenced. The category comprehensibility describes that not all patients understand rules and regulations of copayments and its caps. The assessment of copayments consists of nine subcategories, three of which are rather negative and six of which are rather positive. In all three dimensions the patterns between the German and Norwegian sub-samples differ considerably. Conclusions The results of our study point at the importance of communicating clear rules for copayments which are easily comprehensible. Comprehensibility, copayments, Germany, Norway, patients’ perspective, qualitative research Background Copayments are an important feature in many health care systems (1). They serve two main functions, first, as a steering tool and, second, as a source of funding. There are three types of copayments: deductibles, copayments in their narrower sense and maximum coverage plans. In this article, we focus on copayments in their narrower sense, also called out-of-pocket-payments. They are a share of the actual costs for seeing a doctor or of pharmaceuticals which the patient has to pay. These out-of-pocket-payments are common in many countries no matter which health care type they belong to (1). There are two concepts underlying the assumed steering effect of copayments. The first idea is that copayments are an instrument to raise the cost-awareness of patients. The second idea is that they make health care services less like a flat rate service (2). This, in turn, assumes, that patients are capable to discern ‘useful’ from ‘superfluous’ services and that copayments stop them from falling for moral hazards (3). Moral hazard means the temptation to use services to a greater extent than one usually would for what has already been paid for, for example by an insurance premium or taxes. However, evidence tends to show copayments are more likely to have a blunt effect because patients are not capable to distinguish between necessary and unnecessary services (4). Best-known evidence for the steering effect of copayments is the ‘Health Insurance Experiment’ conducted by the RAND Corporation in the late 1970s and early 1980s. The experiment showed a clear relation between copayments and utilization, however, it has a several shortcomings. Next to ethical issues no consideration was paid to the cultural, temporal, and methodological implications of the study (5). This makes it difficult to draw valid conclusions for outside the USA. For Belgium and France, studies showed no relevant effect of copayments on health care utilization (6,7). However, poorer groups in society or chronically ill with a higher need of utilizing services tend to refrain from or even delay necessary visits to the doctor because of copayments (8–11). A measure to reduce these unintended, negative health and social effects are caps to copayments. The differing effects of copayments in the different studies have not yet been sufficiently explained. Skriabikova et al. (12) suggest that ‘consumer attitudes, experience and culture’ play a role in the effects of copayments, however, measures of these factors are usually not included in the according studies. O’Reilly et al. (13) conducted a survey regarding patients’ attitudes towards copayments for GP services in Ireland. They see a need for more research on the underlying motives for support for or opposition to copayments. Benedetti et al. (14) found limited knowledge of patients regarding copayments in a US study. One of the few qualitative studies was conducted by Doran et al. (15) who researched Australian patients’ attitudes towards copayments for medication. They found costs not to play a fundamental role. An Irish study researched patients’ attitudes towards a 50 cent copayment for medication (16): patients’ attitudes were somehow accepting but sceptical. Another study conducted by Schafheutle et al. (17) shows that patients follow several strategies to reduce their medication cost. In summary, it can be stated that regarding the widespread use of copayments there is little research on a more comprehensive perspective of patients regarding copayments. Especially, we could not find any qualitative study including patients from more than one health care system. There are several comparative studies from an economical health care system perspective, however, they do not include the patients’ perspective (18–20). Thus, there is a need for more knowledge on the patients’ attitudes towards copayments. Knowledge on patients’ attitudes might allow to explain the differences in effects of copayments. Copayments in Norway and Germany We chose to compare Norway and Germany, because they are similar regarding social structure, morbidity and mortality. Norway has a tax-based health care system while Germany has a social insurance system of the Bismarck type. However, general practitioners (GPs) are working mainly self-employed in both countries. The gross national income per capita is higher in Norway than in Germany [in 2011: 88430$ versus 47360$ (21)] mainly due to oil production. The Gini-Index as a parameter for inequity is in Norway slightly lower than in Germany [in 2011: 25.5 versus 30.1 (22)]. Norway and Germany differ greatly in the use of health services as well as regarding the amount and type of copayments. In both countries, there are or were copayments for visiting the doctor and for medication, but the approaches differ. Germany had opted for a quarterly lump-sum service charge of 10 euros (Praxisgebühr, lit. ‘practice charge’) for each adult patient for any number of consultations and referrals within the particular quarter (23). This practice charge did not prove to have an effect on health care utilization in the mid- and long-term (24). Hence, the practice charge was abolished in 2013. In contrast, out-of-pocket payments in Norway are, as a general rule, paid for every new consultation or service by every adult patient. Fees are graduated according to the kind and daytime of service and the qualification of the provider. Copayments for medication in Germany are in general 10% of the price with a minimum of 5 euros and a maximum of 10 euros per package. There are several additional rules, partly depending on individual contracts of public health insurances with specific pharmaceutical companies. In Norway, there is a distinction between medication for serious and chronic diseases and other medications. The public social insurance pays a larger or major part for medications for serious and chronic diseases. The patient just has to pay a copayment. Other medication usually has to be paid in full by the patient. Copayments for medications for chronic diseases have a cap. There is no such limit for other medications. Patients with low income or suffering from certain illnesses get a taxation reduction in case of high expenditure on nonaddictive medications. Norway has introduced caps for copayments in the early 1980s, Germany followed in 2004 (23,25). Germany limits private expenditures for all kinds of copayments on all reimbursable services, pharmaceuticals, and appliances to 2% of the gross household income (1% for chronically ill) per year (23). Norway, in contrast, applies two separate caps independent of the actual income: one for seeing doctors, psychologists, treatment in hospitals, doing X-rays, copayments for pharmaceuticals and patient transport (cap EUR 240), while the other cap limits copayments for physiotherapy, treatment by dentists, or treatment abroad [cap EUR 210; (26–28)]. In both countries, copayments are a relevant source of funding: in 2012, 12% of the expenditure on health in Germany was paid by private out-of-pocket payments and 15% of the expenditure on health in Norway (29). However, this figure includes more than just copayments and incorporates all privately paid health expenditures. Hence, the research question of this article reads: what are adult patients’ attitudes in Germany and Norway towards copayments as a steering tool? And, do these attitudes differ systematically between these two countries? Methods We conducted a comparative qualitative study in the methodological framework of grounded theory (30). In the study, there were three methodological approaches: qualitative interviews with 40 patients, participant observation in overall eight primary care practices—both equally divided between Germany and Norway—and a context analysis of health care system factors emerging during the analysis. Results from the concurrent participant observation in overall eight primary care practices were not entered in the analysis for this article. Field access was gained via eight primary care practices, four in Norway, four in Germany. In both countries, two practices were situated in a rural setting and two in an urban setting. On at least 2 days, all patients in these practices received a short questionnaire, informing them about the study and asking them to participate in a qualitative interview. Furthermore, these potential interviewees were asked for their age, gender, number of chronic conditions and number of visits to the doctor over the last 3 months. From this pool of 280 potential interviewees, we chose the Norwegian interviewees by theoretical sampling and matched the German interviewees. There were 40 participants, 20 in Germany and 20 in Norway born between 1931 and 1990. Eighteen of the participants were female and 22 male. Twenty-two of the interviewees went to a GP in an urban setting and 18 in a rural setting. The interviews conducted were episodic interviews. Episodic interviews are semi-structured qualitative interviews containing a narrative part relating to the interviewee’s experiences and abstract questions. The interviews were conducted in the years 2012 till 2014. Selected interviews were coded line-by-line to develop an initial thematic structure. Then, we coded all interviews and refined the codes and thematic structure. For a further analysis regarding the attitudes towards copayments as a steering tool, we chose all segments coded with ‘copayments’, ‘financing’, ‘costs’, or ‘self-pay’. These text-segments were analysed further for dimensions of attitudes towards copayments and recurring themes and categories as well as types. All participants gave their written informed consent to participate in the study. The study was approved by the local ethics committees in Germany and Norway. A more detailed description of the study design is available in the study protocol (31). All quotes in this article are followed by the study-ID of the interviewee, including a D for patients from the German sample and N for patients from the Norwegian sample. I denotes the interviewer and P the interviewee. Results We found three dimensions regarding attitudes towards copayments: the perceived steering effect, the comprehensibility, and the assessment of copayments. Perceived steering effect of copayments The dimension ‘Perceived Steering Effect of Copayments’ describes how the interviewees perceive the actual steering effect of copayments. We could find three types of interviewees regarding this aspect. The first type are respondents who perceived that copayments had a steering effect on themselves. They reported to have postponed consultations with a doctor for financial reasons. Typical is the following quote of a 40-year-old woman: It was annoying, because I postponed particular diseases, most often those which appeared at the end of the quarter year. Then I said to myself, let’s wait. Let’s wait, because otherwise, you’ve to pay twice. That’s how it is. (D06) To understand the interviewee’s explanation, it is important to know that the copayment for a visit to the physician in outpatient care had to be paid only once in each quarter in Germany. Another type of interview partner were those stating that they themselves had not been influenced, but knew others who were. This type can be summarized by the following quote: We can still afford that [, the copayments], but many others are not able. Then, there was this thing with the ten euros [practice charge]. Many people didn’t see the doctor because they didn’t have these ten euros. And people say, this girl here [in the neighbourhood] died because she did not have these ten euros. If she had gone to see the doctor, she would still be alive. (D03) A third type consists of those who neither see any influence on themselves nor report any influence on people they know of. A typical example is the following: Money doesn’t play a role. And it’s also so convenient, it’s so little money [we have to pay], that has never played a role for me. (N11) However, it is important to notice that whereas interviewees stated that copayments did not influence them personally they mention that copayments might influence peoples’ behaviour in general. For example people stated sometimes that Germans visit the doctor so often because no copayment is necessary. These statements highlight that they assume a general steering effect of copayments on health care utilization. Table 1 shows the distribution of the three types in the qualitative sample. While the Norwegian interviewees experienced nearly uniformly no steering effect of copayments, the types were more equally distributed in the German sample. Table 1. Distribution of three types of perceived steering effect of copayments in the qualitative interview material of 20 German and 20 Norwegian interviewees (n = 40) Type Norway Germany Influenced me 1 4 Influenced others 1 3 Did not influence 18 12 NA 0 1 Type Norway Germany Influenced me 1 4 Influenced others 1 3 Did not influence 18 12 NA 0 1 NA, not applicable. View Large Table 1. Distribution of three types of perceived steering effect of copayments in the qualitative interview material of 20 German and 20 Norwegian interviewees (n = 40) Type Norway Germany Influenced me 1 4 Influenced others 1 3 Did not influence 18 12 NA 0 1 Type Norway Germany Influenced me 1 4 Influenced others 1 3 Did not influence 18 12 NA 0 1 NA, not applicable. View Large Comprehensibility of copayments In the German sample of the interviewees, for some participants the rules regarding copayments appeared not to be comprehensible: That’s when you say, they are not quite right in the head: For [medications produced by] one manufacturer I need to pay co-payments, for those by another manufacturer I do not need to pay. And I cannot comprehend all this. (D14) However, not only the rules for in which cases copayments have to be paid but also the rules how to get a remission from copayments once you have exceeded the maximum amount seem to be difficult to understand. One example is the following patient who thinks to be eligible for such a waiver and tried to get one: P: However, I have to pay copayments for the medications I get prescribed. I do not have a remission. We wanted to get one, but they didn’t do it. I always have to pay five euros or seven euros or so. I: So, you don’t have a remission? So, you do not exceed the maximum? P: Oh, I always have to pay. I’ve broached that. Wouldn’t it be better, if I do not have to pay for it with the minimal pension I get? However, I always have to pay for it. And there was this almoner here, and I addressed this topic. But she didn’t do anything. […] Well, but I always have to pay copayments. I haven’t got a waiver yet. […] I broached that topic several times. I said it, but nothing has changed. (D18) This interviewee seems not to know who is responsible to waive the copayments, which requirements have to be fulfilled, and what he has to do in order to get such a waiver. This can be highlighted also by an interviewee who has experience with health care services in both Germany and Norway: P1: And how many copayments I’ve paid, that’s registered at the Norwegian social insurance office. And as soon as I’ve reached the limit, I get the waiver, and everything I’ve paid too much gets re-transferred to my bank account. P2: I do not need to care. P1: And I do not need to care about it my own. I do not need to think about it. I do not have to collect receipts as I had to in Germany. (N13) Accordingly, while in the German sub-sample some interviewees report problems in the comprehensibility of the rules, the Norwegian interviewees did not mention such difficulties. Assessment of copayments Regarding the general assessment of copayments, we could find nine topics. Three with a negative tendency towards copayments: disgrace, commerce and burden. These rather negative assessment types can be contrasted to six with a more positive tendency: the importance of the cap, the perception of copayments as undramatic and fair, health care services as privilege, the educational effect of copayments and health as worth it. A rather negative assessment of copayments The most negative assessment of copayments is the view some German interviewees took. They dispraised copayments as a cruel disgrace: This form of financing the healthcare system is a disgrace. (D14) They highlight that they perceive copayments as inhuman, as in general not acceptable in a social welfare state. This relates to another group of interviewees who regard copayments as being indicative for a negative commerce orientation of the health care system: If they [the healthcare insurances] impel it like the pharmacies do; that’s all just commerce, nothing else. (D19) Thus, interviewees arguing in these two categories regard an economical perspective on health care as generally not acceptable. Hence, the third negative category is more strongly related to the individual perspective. Interviewees of this category highlight that copayments can be a burden for individuals: Well, the financial burden for some of these things is enormous. (D02) Thus, negative assessments of copayments can be more generalized opposing a financial dimension of health care in general or more devoted to the impact of the individual burden of copayments. A more positive assessment of copayments In contrast to these negative stances, several more positive aspects of copayments are mentioned by other interviewees. Instead of perceiving copayments as a burden, some interviewees highlight that the amount of copayments is not dramatic: They [copayments] are nothing dramatic at all. (D11) Some even see it as just a small amount to pay: That is not really a price to pay if you go to such a public doctor. (N20) Connected to assessing the financial burden of copayments as negligible is the role of the cap of copayments. Thus, some of the interviewees highlight the importance of the limit of copayments for their assessment: Costs with visiting the doctor? No, I think that in particular that little bit is absolutely all right. You shall stand for this copayment, and then you get the healthcare exemption card. (N17) The cap is perceived as an important factor to assess copayments and is considered as fair: I: That means you assess these 200, 250 [euros of copayments] as … P: Yes, as fair. (D16) Some interviewees, in particular those, who experienced having a severe disease resulting in high treatment costs highlighted that it is a privilege to get health care paid and how grateful they are. And beyond this gratefulness they assess copayments as a minor amount of money. Other interviewees highlight that health is worth paying what needs to be paid: If health is at stake and I need to see the doctor, then money doesn’t play a role. (N13) Finally, several interviewees highlight that copayments fulfil an educational goal and steer on an abstract level: Well, that’s at least something the state has achieved: Because [of copayments] people didn’t see the doctor so often anymore. (D06) Thus, the positive views of copayments are mainly related to assessing the burden by copayments as small or at least acceptable. This is either due to the relatively small amount of the single copayments or the relatively small overall amount of copayments because of the cap. Table 2 shows the distribution of the different categories in assessing copayments in the qualitative sample. It shows that the assessment of copayments is different between the German and the Norwegian sub-sample. In the German sub-sample, all categories are mentioned, especially also the more negative ones. In the Norwegian sub-sample, the negative categories have not been mentioned and the limit of copayments it is described as undramatic. Table 2. Frequency of mentioned categories regarding the assessment of copayments in the qualitative interview material of 20 German and 20 Norwegian interviewees (participants can count multiple times, n = 40) Category Norway Germany Disgrace 0 2 Burden 0 5 Commerce 0 3 Importance of limit 8 1 Undramatic 10 3 Fair 3 1 Privilege 3 3 Educational effect 3 4 Health is worth it 3 3 NA 5 3 Category Norway Germany Disgrace 0 2 Burden 0 5 Commerce 0 3 Importance of limit 8 1 Undramatic 10 3 Fair 3 1 Privilege 3 3 Educational effect 3 4 Health is worth it 3 3 NA 5 3 NA, not applicable. View Large Table 2. Frequency of mentioned categories regarding the assessment of copayments in the qualitative interview material of 20 German and 20 Norwegian interviewees (participants can count multiple times, n = 40) Category Norway Germany Disgrace 0 2 Burden 0 5 Commerce 0 3 Importance of limit 8 1 Undramatic 10 3 Fair 3 1 Privilege 3 3 Educational effect 3 4 Health is worth it 3 3 NA 5 3 Category Norway Germany Disgrace 0 2 Burden 0 5 Commerce 0 3 Importance of limit 8 1 Undramatic 10 3 Fair 3 1 Privilege 3 3 Educational effect 3 4 Health is worth it 3 3 NA 5 3 NA, not applicable. View Large Discussion In this study, we researched patients’ attitudes towards copayments as a steering tool in Norway and Germany. We found three dimensions of patients’ attitudes towards copayments as a steering tool: the perceived steering effect, the comprehensibility and the general assessment of copayments. A limitation of our study is that there might be a selection bias: because we recruited participants from patients in health care centres, only persons who were willing and able to pay copayments were included in the pool of potential participants. However, for the German sub-sample we could show that there is only a small difference of those interested in taking part in such an interview and the general population. This potential selection bias should be addressed in quantitative follow-up studies by including the socio-economic status of the participants. Another limitation is that copayments for visiting a physician in outpatient care have been abolished in Germany during the course of our study. However, the interviewees still vividly remembered the times of the copayments. The aspect of the comprehensibility of copayment regulations is rarely discussed in literature: Benedetti et al. (14) showed for respondents from a private insurance in the US health care system that 4.4% of the respondents did not know correctly if they had to pay copayments for a consultation and only half of the respondents knew the amount of copayments they had to pay correctly. This is in line with the findings of Hsu et al. (32), who found in another US survey that only half of the respondents knew the amount of their copayments correctly. In a focus group with elderly in the USA, Cline et al. (33) found the aspect of drug benefit comprehensibility which is similar to our dimension of comprehensibility: the regulations were often not comprehensible to the elder patients taking part in Cline’s study. In our qualitative study, there seems to be a difference in comprehensibility between Norway and Germany, comprehensibility being no issue in Norway and a problem for some patients in Germany. This difference in the perceived comprehensibility of copayments might relate to the different rules in the two countries: in Norway, all payments are registered automatically in a central register so that the waiver is sent automatically to the patient in question once the individual copayments have exceeded the fixed amount. In Germany, in contrast, patients need to collect receipts from physician, pharmacies, hospitals, etc. to apply for a waiver to their sickness fund once the amount relative to their income is exceeded. Therefore, patients in Germany need to have organizational skills. This might be inappropriate especially for the ill, disabled, less educated and vulnerable patients and might explain the difference in reported comprehensibility. The divergent assessment of copayments in our study fits to the findings of Sinnott et al. (16) who found diverging assessment of a 50 cent copayment on drugs. They found varying perceptions of copayments from a relevant burden to nothing dramatic. The different patterns of distribution in both samples of our study point towards a different view of copayments in Norway and Germany. In Norway, the function as a financing instrument of copayments seems to dominate. The participants usually saw copayments as a normal part of using the health care system and did mainly not attribute any individual influences on their behaviour. This might be connected to the long existence of copayments in Norway, and their origin in the rather small amount that was initially covered by the health care system. In Germany, patients found the role of copayments less clear, and there was a relevant opposition against copayments. This more critical view is in line with the findings of a survey conducted by Wendt et al. (34) who found an opposition in the German population against higher payments or less service. The different assessment of copayments might be also have been influenced by the different way of paying for the health care system. Paying for health care indirectly through taxes might not as much lead to the feeling of being entitled to a full-scale coverage as paying an individual health care insurance fee. Additionally, our results reveal that an important factor to perceive copayments as fair is a cap for copayments. That this cap is perceived less important for fairness in Germany than in Norway should be further explored. Possible reasons in Germany are that the height of the cap depends on the individual disease status and the individual income and that patients have to collect their receipts actively and apply for a remission reaching the cap. This is contrasted by the Norwegian system with a high degree of automatism. In Norway, the administration and not the patient have to know the rules. Thus, the procedures for copayment caps might influence their perception implicitly. Conclusions Our study highlights that the procedures around and comprehensibility of copayments might be of importance for the perception and understanding of copayments in the population and thereby influence their acceptance. In future, studies on the steering effect of copayments, the individual concepts of copayments should be taken into account. Our results may help to operationalize these concepts for quantitative measurements of such factors. Future studies should explore the interplay of the socio-economic status, health literacy, the perceived steering effect, individual concepts of copayments and the comprehensibility in a quantitative way. The results emphasize that when communicating about copayments one should distinguish the steering function of copayments from the financing function. To relieve patients from the burden to administrate their own copayments, automated caps might be an important measure to increase the acceptance of copayments. Declaration Funding: the study has been funded by the Deutsche Forschungsgemeinschaft (DFG—HE6399/1-1 and HE6399/1–2). Ethics: the study was approved by the local ethics committees in Germany and Norway. Conflict of interest: none. Acknowledgements We thank all researchers and students who contributed to the study, namely Anne-Katharina Koch, Kevin Schröder, Jona Ober, Christina Wagenknecht, Anne Bretschneider, Stephan Bilkenroth, Yvonne Marx, Sigurd Skrondal, Sabina Sendrowicz, Sebastian Huter, Jan Islei, Thomas Lichte, Markus Herrmann and Uwe Flick. Especially, we thank all GPs who allowed us to do participant observation in their practices and all interviewees. References 1. Bíró A . Copayments, gatekeeping, and the utilization of outpatient public and private care at age 50 and above in Europe . Health Policy 2013 ; 111 : 24 – 33 . Google Scholar CrossRef Search ADS PubMed 2. Drummond M , Towse A . Is it time to reconsider the role of patient co-payments for pharmaceuticals in Europe ? Eur J Health Econ 2012 ; 13 : 1 – 5 . Google Scholar CrossRef Search ADS PubMed 3. Holst J. Modalitäten von Zuzahlungen im Gesundheitswesen . Eschborn : gtz , 2004 . 4. Remler DK , Greene J . Cost-sharing: a blunt instrument . Annu Rev Public Health 2009 ; 30 : 293 – 311 . Google Scholar CrossRef Search ADS PubMed 5. Brook RH , Keeler EB , Lohr KL et al. The Health Insurance Experiment: A Classic RAND Study Speaks to the Current Health Care Reform Debate . Santa Monica, CA : RAND Corporation , 2006 . http://www.rand.org/pubs/research_briefs/RB9174.html (accessed on 23 August 2016). 6. Chiappori PA . Moral hazard and the demand for physician services: first lessons from a French natural experiment . Eur Econ Rev 1998 ; 42 : 499 – 511 . Google Scholar CrossRef Search ADS 7. Cockx B , Brasseur C . The demand for physician services. Evidence from a natural experiment . J Health Econ 2003 ; 22 : 881 – 913 . Google Scholar CrossRef Search ADS PubMed 8. Elofsson S , Undén AL , Krakau I . Patient charges—a hindrance to financially and psychosocially disadvantage groups seeking care . Soc Sci Med 1998 ; 46 : 1375 – 80 . Google Scholar CrossRef Search ADS PubMed 9. Lostao L , Regidor E , Geyer S , Aïach P . Patient cost sharing and social inequalities in access to health care in three western European countries . Soc Sci Med 2007 ; 65 : 367 – 76 . Google Scholar CrossRef Search ADS PubMed 10. Tamblyn R , Laprise R , Hanley JA et al. Adverse events associated with prescription drug cost-sharing among poor and elderly persons . JAMA 2001 ; 285 : 421 – 9 . Google Scholar CrossRef Search ADS PubMed 11. Rückert IM , Böcken J , Mielck A . Are German patients burdened by the practice charge for physician visits (‘Praxisgebuehr’)? A cross sectional analysis of socio-economic and health related factors . BMC Health Serv Res 2008 ; 8 : 232 . Google Scholar CrossRef Search ADS PubMed 12. Skriabikova O , Pavlova M , Groot W . Empirical models of demand for out-patient physician services and their relevance to the assessment of patient payment policies: a critical review of the literature . Int J Environ Res Public Health 2010 ; 7 : 2708 – 25 . Google Scholar CrossRef Search ADS PubMed 13. O’Reilly D , O’Neill C , O’Dowd T et al. Patients’ attitudes to co-payments for general practitioner services: do they reflect the prevailing system ? J Health Serv Res Policy 2007 ; 12 : 197 – 201 . Google Scholar CrossRef Search ADS PubMed 14. Benedetti NJ , Fung V , Reed M et al. Office visit copayments: patient knowledge, response, and communication with providers . Med Care 2008 ; 46 : 403 – 9 . Google Scholar CrossRef Search ADS PubMed 15. Doran E , Robertson J , Henry D . Moral hazard and prescription medicine use in Australia—the patient perspective . Soc Sci Med 2005 ; 60 : 1437 – 43 . Google Scholar CrossRef Search ADS PubMed 16. Sinnott SJ , Guinane M , Whelton H , Byrne S . Is 50 cent the price of the optimal copayment?—A qualitative study of patient opinions and attitudes in response to a 50 cent charge on prescription drugs in a publicly funded health system in Ireland . BMC Health Serv Res 2013 ; 13 : 16 . Google Scholar CrossRef Search ADS PubMed 17. Schafheutle EI , Hassell K , Noyce PR , Weiss MC . Access to medicines: cost as an influence on the views and behaviour of patients . Health Soc Care Community 2002 ; 10 : 187 – 95 . Google Scholar CrossRef Search ADS PubMed 18. Fuchs VR , Hahn JS . How does Canada do it? A comparison of expenditures for physicians’ services in the United States and Canada . N Engl J Med 1990 ; 323 : 884 – 90 . Google Scholar CrossRef Search ADS PubMed 19. Henke K-D , Schreyögg J. Towards Sustainable Health Care Systems—Strategies in Health Insurance Schemes in France, Germany, Japan and the Netherlands . Geneva : International Social Security Association , 2004 . 20. Ettelt S , Thomson S , Nolte E et al. Reimbursing highly specialised hospital services: the experience of activity-based funding in eight countries. A report commissioned by the Department of Health and prepared by the London School of Hygiene and Tropical Medicine . London School of Hygiene and Tropical Medicine , 2006 . http://researchonline.lshtm.ac.uk/id/eprint/5510 (accessed 1 August 2017). 21. The World Bank . GNI per Capita, Atlas Method (Current US$) . http://data.worldbank.org/indicator/NY.GNP.PCAP.CD?locations=NO-DE (accessed 1 August 2017). 22. The World Bank . GINI Index (World Bank Estimate) . http://data.worldbank.org/indicator/SI.POV.GINI?locations=DE-NO (accessed 1 August 2017). 23. Busse R , Riesberg A. Gesundheitssysteme im Wandel: Deutschland . Copenhagen : WHO Regionalbüro im Auftrag des Europäischen Oberservatoriums für Gesundheitssysteme und Gesundheitspolitik , 2005 . 24. Schreyögg J , Grabka MM . Copayments for ambulatory care in Germany: A natural experiment using a difference-in-difference approach . Eur J Health Econ 2010 ; 11 : 331 – 41 . Google Scholar CrossRef Search ADS PubMed 25. Johnsen JR. Health Systems in Transition: Norway . Copenhagen : WHO Regional Office for Europe on behalf of the European Observatory on Health Systems and Policies , 2006 . 26. HELFO . Frikort for helsetjenester egenandelstak 1 . https://helfo.no/frikort-for-helsetjenester-egenandelstak-1 (accessed 25 October 2015). 27. HELFO . Frikort for helsetjenester egenandelstak 2 .https://helfo.no/frikort-for-helsetjenester-egenandelstak-2 (accessed 25 October 2015). 28. HELFO . Frikort for helsetjenester . https://helsenorge.no/betaling-for-helsetjenester/frikort-for-helsetjenester (accessed 26 October 2015). 29. OECD . Health at a Glance: Europe 2014 . http://dx.doi.org/10.1787/health_glance_eur-2014-en (accessed 23 August 2016). 30. Glaser BG , Strauss AL. The Discovery of Grounded Theory. Strategies for Qualitative Research . Chicago : Aldine , 1967 . 31. Herrmann WJ , Haarmann A , Flick U et al. Patients’ subjective concepts about primary healthcare utilisation: the study protocol of a qualitative comparative study between Norway and Germany . BMJ Open 2013 ; 3 : e002952 . Google Scholar CrossRef Search ADS PubMed 32. Hsu J , Reed M , Brand R , Fireman B , Newhouse JP , Selby JV . Cost-sharing: patient knowledge and effects on seeking emergency department care . Med Care 2004 ; 42 : 290 – 6 . Google Scholar CrossRef Search ADS PubMed 33. Cline RR , Gupta K , Singh RL , Schommer JC . Older adults’ drug benefit beliefs: a focus group study . J Manag Care Pharm 2005 ; 11 : 77 – 85 . Google Scholar PubMed 34. Wendt C , Naumann E , Klitzke J . Reformbedarf im deutschen Gesundheitssystem aus Sicht der Bevölkerung . Zeitschrift für Sozialreform 2014 ; 60 : 333 – 48 . Google Scholar CrossRef Search ADS © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Toussaint, Anne;Riedl, Bernhard;Kehrer, Simon;Schneider, Antonius;Löwe, Bernd;Linde, Klaus

2017 Family Practice

doi: 10.1093/fampra/cmx116pmid: 29145575

Abstract Aim The Somatic Symptom Disorder–B Criteria Scale (SSD-12) assesses the psychological features of DSM-5 somatic symptom disorder. The purpose of the current study was to investigate the psychometric characteristics and validity of the 12-item instrument to demonstrate its suitability in primary care. Method The study was designed as a cross-sectional survey set in five primary care practices from Munich, Germany (n = 501, 52.0% female, mean age 47 ± 16 years). Item and scale characteristics, as well as measures of reliability and validity, were determined. Results The SSD-12 has good item characteristics and excellent reliability (Cronbach’s α = 0.92). Confirmatory factor analyses provided evidence to support a general factor model of the SSD-12 in primary care (comparative fit index > 0.98, Tucker–Lewis index > 0.98, root mean square error of approximation = 0.090, 90% confidence interval: 0.078–0.102). SSD-12 total sum-score was significantly associated with somatic symptom burden (r = 0.48, P < 0.001), general anxiety (r = 0.54, P < 0.001) and depressive symptoms (r = 0.60, P < 0.001). At the group level, SSD-12 scores could differentiate between different patient groups (e.g. with and without chronic illness). Conclusions The SSD-12 appears to be a reliable, valid and time-efficient self-report measure of the psychological characteristics related to the experience of somatic symptoms which is suitable for primary care. Future research should evaluate its responsiveness to treatment and feasibility as a screening tool in different clinical settings. Diagnosis, medically unexplained symptoms, psychological factors, psychometrics, questionnaires, somatoform disorders Introduction With the release of the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, the diagnostic category of Somatoform Disorders was changed to Somatic Symptom and Related Disorders (SSD) (1). By taking the accumulated knowledge of the last 20 years on pathogenesis, maintenance and prognosis of distressing somatic symptoms into account, DSM-5 has fundamentally shifted the way somatoform disorders are defined, aiming to increase their relevance in the primary care setting (2). While medically unexplained symptoms were a key feature for many of the disorders in DSM-IV (3) and ICD-10 (4), a SSD diagnosis does not require that the patients’ somatic symptoms are medically unexplained. Regardless of their etiology, SSD is characterized by somatic symptoms that are either very distressing or result in significant disruption of functioning (A criterion). In addition, to be diagnosed with SSD, the individual must experience excessive and disproportionate thoughts, feelings and behaviours regarding those symptoms (B criteria) which typically persist at least for 6 months (C criterion). It is believed that ICD-11 will follow a similar approach by introducing the so called ‘bodily distress syndrome’ when it is released (5). Primary care physicians often treat patients with distressing symptoms, such as pain, digestive, cardiovascular or sensory complaints for which there are no biological causes. The symptoms may result from a heightened awareness of certain bodily sensations, combined with a tendency to interpret these sensations as indicators of a medical illness (6). An estimated 20–25% of patients who present with acute somatic symptoms go on to develop a chronic illness. The prevalence of SSD in the general population is estimated to be 5–7% (1), making it one of the most common categories of patient complaints in the primary care setting (7). Many of these patients are significantly impaired by their symptoms and may be subjected to unnecessary testing and procedures, leading to high socio-economic costs (8). Thus, appropriate diagnosis is essential. Self-report questionnaires may assist with the frequent challenge of making a precise assessment in a limited period of time. Although questionnaires alone are insufficient to form the basis of a diagnosis and should always be used in conjunction with a comprehensive clinical evaluation, they are useful to capture the patients’ perception and overall experience of their symptoms. In addition, responses may help to guide discussions about goals and expectations for symptom management (9). There are several well-validated screening questionnaires to determine the presence and severity of a patients’ somatic symptoms (SSD criterion A) [e.g. PHQ-15 (10)]. The Somatic Symptom Disorder–B Criteria Scale (SSD-12) was developed to additionally measure the psychological distress associated with bothersome somatic symptoms (SSD B criteria). Previous studies have shown that the SSD-12 has good reliability and content validity, and is suitable for screening and evaluating psychological aspects of SSD in specialized clinical settings (11). German population-based norms are also available (12). Since the SSD-12 has not yet been applied to general practice, the current study investigated the psychometric characteristics and validity of the instrument within a primary care sample. Methods Participants In a cross-sectional study, the SSD-12 was administered to a convenience primary care sample of 501 patients (52.0% female, mean age 47 ± 16 years) at five primary care practices from Munich, Germany, affiliated to the German Association of Statutory Health Insurance Doctors. Data were collected between October 2015 and April 2016. All consecutive patients who met the inclusion criteria were invited to participate. Inclusion criteria were an age at or above 18 years and the ability to read and understand the German language. Patients who did not attend the practice for a personal consultation or required emergency care were excluded from participation. Fifty of the approached patients did not want to participate. After providing written informed consent, 501 patients completed the presented self-report questionnaires. Table 1 provides characteristics of the sample. Table 1. Demographic characteristics of the primary care sample (2016: Munich, Germany; N = 501) Total (n = 501) Age in years: mean (SD) 47 (16) Affiliation to primary care practice in years: mean (SD) 8 (6) n % Sex (female) 259 52.0 Marital status  Married 284 57.1  Unmarried couple (living together) 59 11.9  Single 133 26.8  Widowed 21 4.2 Minimum of one child 332 66.7 Education  No formal qualifications 21 4.2  Up to 10 years of education 321 64.1  More than 10 years of education 145 28.9  Students 14 2.8 Private health insurance 53 10.7 ICPC-2 categories  Respiratory (R) 109 21.8  Musculoskeletal (L) 107 21.4  Process codes (−) 77 15.4  Digestive (D) 39 7.8  Cardiovascular (K) 29 5.8  General and unspecified (A) 28 5.6  Psychological (P) 27 5.4  Other 76 15.1 Most frequent single ICPC-2 codes  Upper respiratory infection acute (R74) 57 11.4  Back syndrome w/o radiating pain (L84) 23 4.6  Blood test (−34) 19 3.8  Gastroenteritis (D73) 17 3.4  Preventive immunization/medication (−44) 15 3.0  Therapeutic counselling (−58) 15 3.0  Neck syndrome (L83) 14 2.8  General symptom/complaint other (A29) 12 2.4  Sinusitis acute/chronic (R75) 10 2.0  Acute bronchitis/bronchiolitis (R78) 9 1.8  Bursitis/tendinitis/synovitis (L87) 8 1.6 At least one self-reported chronic disease (e.g. diabetes or hypertension) 184 36.7 High risk for any psychological disorder (PHQ-15 ≥ 15 or PHQ-9 ≥ 10 or GAD-7 ≥ 10) 123 24.6 High risk for somatization (PHQ-15 ≥ 15) 68 13.6 Total (n = 501) Age in years: mean (SD) 47 (16) Affiliation to primary care practice in years: mean (SD) 8 (6) n % Sex (female) 259 52.0 Marital status  Married 284 57.1  Unmarried couple (living together) 59 11.9  Single 133 26.8  Widowed 21 4.2 Minimum of one child 332 66.7 Education  No formal qualifications 21 4.2  Up to 10 years of education 321 64.1  More than 10 years of education 145 28.9  Students 14 2.8 Private health insurance 53 10.7 ICPC-2 categories  Respiratory (R) 109 21.8  Musculoskeletal (L) 107 21.4  Process codes (−) 77 15.4  Digestive (D) 39 7.8  Cardiovascular (K) 29 5.8  General and unspecified (A) 28 5.6  Psychological (P) 27 5.4  Other 76 15.1 Most frequent single ICPC-2 codes  Upper respiratory infection acute (R74) 57 11.4  Back syndrome w/o radiating pain (L84) 23 4.6  Blood test (−34) 19 3.8  Gastroenteritis (D73) 17 3.4  Preventive immunization/medication (−44) 15 3.0  Therapeutic counselling (−58) 15 3.0  Neck syndrome (L83) 14 2.8  General symptom/complaint other (A29) 12 2.4  Sinusitis acute/chronic (R75) 10 2.0  Acute bronchitis/bronchiolitis (R78) 9 1.8  Bursitis/tendinitis/synovitis (L87) 8 1.6 At least one self-reported chronic disease (e.g. diabetes or hypertension) 184 36.7 High risk for any psychological disorder (PHQ-15 ≥ 15 or PHQ-9 ≥ 10 or GAD-7 ≥ 10) 123 24.6 High risk for somatization (PHQ-15 ≥ 15) 68 13.6 Values are absolute frequencies (percentages) or means (SD). View Large Table 1. Demographic characteristics of the primary care sample (2016: Munich, Germany; N = 501) Total (n = 501) Age in years: mean (SD) 47 (16) Affiliation to primary care practice in years: mean (SD) 8 (6) n % Sex (female) 259 52.0 Marital status  Married 284 57.1  Unmarried couple (living together) 59 11.9  Single 133 26.8  Widowed 21 4.2 Minimum of one child 332 66.7 Education  No formal qualifications 21 4.2  Up to 10 years of education 321 64.1  More than 10 years of education 145 28.9  Students 14 2.8 Private health insurance 53 10.7 ICPC-2 categories  Respiratory (R) 109 21.8  Musculoskeletal (L) 107 21.4  Process codes (−) 77 15.4  Digestive (D) 39 7.8  Cardiovascular (K) 29 5.8  General and unspecified (A) 28 5.6  Psychological (P) 27 5.4  Other 76 15.1 Most frequent single ICPC-2 codes  Upper respiratory infection acute (R74) 57 11.4  Back syndrome w/o radiating pain (L84) 23 4.6  Blood test (−34) 19 3.8  Gastroenteritis (D73) 17 3.4  Preventive immunization/medication (−44) 15 3.0  Therapeutic counselling (−58) 15 3.0  Neck syndrome (L83) 14 2.8  General symptom/complaint other (A29) 12 2.4  Sinusitis acute/chronic (R75) 10 2.0  Acute bronchitis/bronchiolitis (R78) 9 1.8  Bursitis/tendinitis/synovitis (L87) 8 1.6 At least one self-reported chronic disease (e.g. diabetes or hypertension) 184 36.7 High risk for any psychological disorder (PHQ-15 ≥ 15 or PHQ-9 ≥ 10 or GAD-7 ≥ 10) 123 24.6 High risk for somatization (PHQ-15 ≥ 15) 68 13.6 Total (n = 501) Age in years: mean (SD) 47 (16) Affiliation to primary care practice in years: mean (SD) 8 (6) n % Sex (female) 259 52.0 Marital status  Married 284 57.1  Unmarried couple (living together) 59 11.9  Single 133 26.8  Widowed 21 4.2 Minimum of one child 332 66.7 Education  No formal qualifications 21 4.2  Up to 10 years of education 321 64.1  More than 10 years of education 145 28.9  Students 14 2.8 Private health insurance 53 10.7 ICPC-2 categories  Respiratory (R) 109 21.8  Musculoskeletal (L) 107 21.4  Process codes (−) 77 15.4  Digestive (D) 39 7.8  Cardiovascular (K) 29 5.8  General and unspecified (A) 28 5.6  Psychological (P) 27 5.4  Other 76 15.1 Most frequent single ICPC-2 codes  Upper respiratory infection acute (R74) 57 11.4  Back syndrome w/o radiating pain (L84) 23 4.6  Blood test (−34) 19 3.8  Gastroenteritis (D73) 17 3.4  Preventive immunization/medication (−44) 15 3.0  Therapeutic counselling (−58) 15 3.0  Neck syndrome (L83) 14 2.8  General symptom/complaint other (A29) 12 2.4  Sinusitis acute/chronic (R75) 10 2.0  Acute bronchitis/bronchiolitis (R78) 9 1.8  Bursitis/tendinitis/synovitis (L87) 8 1.6 At least one self-reported chronic disease (e.g. diabetes or hypertension) 184 36.7 High risk for any psychological disorder (PHQ-15 ≥ 15 or PHQ-9 ≥ 10 or GAD-7 ≥ 10) 123 24.6 High risk for somatization (PHQ-15 ≥ 15) 68 13.6 Values are absolute frequencies (percentages) or means (SD). View Large Instruments The SSD-12 is composed of 12 items (Supplementary Material A). Each of the three psychological sub-criteria of SSD (cognitive, affective and behavioural aspects associated with bothersome somatic symptoms) is measured by four items with all item scores ranging between 0 (never) and 4 (very often). Previous studies from different settings could show that the SSD-12 has good item characteristics and excellent reliability (Cronbach’s α = 0.95) (11). Using confirmatory factor analyses, both a three-factorial structure which reflects the three psychological criteria of SSD, and a general one-factor model describing psychological distress associated with bothersome somatic symptoms fitted the data well. The SSD-12 total sum-score was closely associated with somatic symptom burden and health anxiety. SSD-12 scores were moderately associated with anxiety and depression. Patients with a higher SSD-12 score reported greater general physical and mental health impairment and significantly higher health care utilization which are key features of patients suffering from somatoform and related disorders (11). Norm values derived from a large sample of the German general population enable comparisons of individual SSD-12 sum-scores with representative data (12). In addition to questions regarding socio-demographic information, other well validated measures to examine the construct validity of the SSD-12 were included in the study. The Patient Health Questionnaire-15 (PHQ-15) (10) assesses the presence and severity of common somatic symptoms (SSD A criterion) within the last 4 weeks using 15 items. Sum-scores range from 0 (not at all) to 30 (very high) and indicate the self-rated symptom burden. We chose this measure because we expect patients with a higher somatic symptom load (PHQ-15) to report a higher psychological burden associated with the symptoms (SSD-12) which should translate into a substantial statistical correlation between the sum-scores of the two instruments. Furthermore, in terms of co-morbidity, somatic symptoms are highly associated with depression and anxiety. Therefore, participants also responded to the Patient Health Questionnaire-9 (PHQ-9) (13) which assesses the presence of the nine DSM criteria for major depression within the last 2 weeks. Scores range from 0 (not at all) to 27 (very high) and indicate the severity of depression. The General Anxiety Disorder-7 (GAD-7) (14) is a self-administered patient questionnaire which is used as a screening tool and severity measure of both generalized anxiety disorder as well as other common anxiety disorders. Scores range from 0 (minimal) to 21 (severe). However, since all three diagnostic concepts (somatoform disorders, depression, anxiety disorders) are considered to be distinct entities, the correlations between the SSD-12 and PHQ-9 and GAD-7 sum-scores were expected to be moderate at best and could then be interpreted in terms of divergent validity. Statistical analyses We computed means, standard deviations and corrected item-total correlations to reflect the psychometric properties of the items. Cronbach’s alpha was determined as a measure of the internal consistency of the scale. To investigate the factorial structure of the questionnaire, we conducted two confirmatory factor analyses (CFA) theoretically derived from a previous study in a clinical population (11): a simple general factor model with all items loading on one ‘general factor’ (the psychological distress associated with bothersome somatic symptoms) and a second ‘three-factor model’ to investigate whether three latent dimensions corresponding to the three sub-criteria of SSD (cognitive, affective, behavioural aspects) could explain the data. Due to the non-normal distribution of our data, robust weighted least squares estimation with mean and variance adjustment (WLSMV) was used to fit the model and missing cases were excluded listwise. The comparative fit index (CFI), Tucker–Lewis index (TLI) and the root mean square error of approximation (RMSEA) were used to test global model fits. Values of RMSEA < 0.60, TLI > 0.95, CFI > 0.95, and SRMR < 0.08 (15) indicate a good fit for continuous data. Construct validity was examined using bivariate correlations between SSD-12, PHQ-15, GAD-7 and PHQ-9. We only included participants who had answered at least 75 % of the items of the respective questionnaires. Mean responses were imputed for any missing data for included participants. To assess the suitability of the SSD-12 for different patient groups, we compared the means of the SSD-12 for patients with and without self-reported chronic disease, at high and low risk of a self-reported psychological disorder (PHQ-15 ≥ 15 or PHQ-9 ≥ 10 or GAD-7 ≥ 10) and at high and low risk of somatization (PHQ-15 ≥ 15). Since patients with either a chronic somatic or psychological disorder are likely to report more bothersome somatic symptoms associated with their disorder (10), we expected them to also score higher in terms of the psychological burden associated with these symptoms (SSD-12 sum-score). Respective frequencies are reported in Table 1. Group comparisons were performed using t-tests (mean) for continuous variables. Tests were considered statistically significant at a two-sided P-value <0.05. Data analyses were conducted using IBM SPSS Statistics 24 and the lavaan package from the statistical computing software R. Results We analyzed data of n = 465 (92.8%) participants who had answered a total of at least 9 of the 12 SSD-12 items (75%). Mean responses were imputed for any additional missing data. Descriptive item statistics The participants responded well to the questionnaire and there was no indication that particular items were skipped or neglected in a systematic way. Responses for every item covered the full range of response categories. Frequency distribution of responses, mean, standard deviation, skewness and item-total-correlations are given in Table 2. We acknowledge the marked skewness of all items which is expected given that the clinical aim is to assess psychological burden in SSD in a primary care setting. Item 10 showed somewhat problematic parameters. It specifies the need for ‘disproportionate’ thoughts about the seriousness of one’s symptoms which seems to be difficult for patients to judge. The internal consistency of the full scale was α = 0.92. Table 2. Frequency distribution of responses (%), means (SD), skewness and item-total correlations for the items of the SSD-12 in the primary care sample (2016: Munich, Germany; N = 501) Item Missing Never (0) Rarely (1) Sometimes (2) Often (3) Very often (4) Mean (SD) Skew. CoriT Cognitive 1 0.4 34.6 34.2 24.5 4.3 1.9 1.0 (1.0) 0.71 0.68 4 0.4 29.2 24.3 27.1 14.2 4.7 1.4 (1.2) 0.38 0.67 7 1.7 53.1 22.2 16.1 4.9 1.9 0.8 (1.0) 1.18 0.31 10 0.4 66.2 17.6 13.1 1.9 0.6 0.5 (0.8) 1.54 0.44 Affective 2 0.2 20.0 38.5 29.9 9.0 2.4 1.4 (1.0) 0.45 0.70 5 0.2 36.8 31.0 21.9 7.7 2.4 1.1 (1.1) 0.74 0.81 8 0.4 48.0 23.9 18.1 6.7 3.0 0.9 (1.1) 1.01 0.75 12 0.2 44.1 31.6 17.4 4.9 1.7 0.9 (1.1) 0.88 0.81 Behavioural 3 0.2 39.4 32.7 18.3 7.1 2.4 1.0 (1.0) 0.89 0.77 6 0.4 50.1 32.5 9.5 5.6 1.9 0.8 (1.0) 1.37 0.76 9 0.4 55.3 23.7 14.2 4.1 2.4 0.7 (1.0) 1.35 0.79 11 0.4 44.5 24.5 21.1 6.7 2.8 1.0 (1.0) 1.00 0.69 Item Missing Never (0) Rarely (1) Sometimes (2) Often (3) Very often (4) Mean (SD) Skew. CoriT Cognitive 1 0.4 34.6 34.2 24.5 4.3 1.9 1.0 (1.0) 0.71 0.68 4 0.4 29.2 24.3 27.1 14.2 4.7 1.4 (1.2) 0.38 0.67 7 1.7 53.1 22.2 16.1 4.9 1.9 0.8 (1.0) 1.18 0.31 10 0.4 66.2 17.6 13.1 1.9 0.6 0.5 (0.8) 1.54 0.44 Affective 2 0.2 20.0 38.5 29.9 9.0 2.4 1.4 (1.0) 0.45 0.70 5 0.2 36.8 31.0 21.9 7.7 2.4 1.1 (1.1) 0.74 0.81 8 0.4 48.0 23.9 18.1 6.7 3.0 0.9 (1.1) 1.01 0.75 12 0.2 44.1 31.6 17.4 4.9 1.7 0.9 (1.1) 0.88 0.81 Behavioural 3 0.2 39.4 32.7 18.3 7.1 2.4 1.0 (1.0) 0.89 0.77 6 0.4 50.1 32.5 9.5 5.6 1.9 0.8 (1.0) 1.37 0.76 9 0.4 55.3 23.7 14.2 4.1 2.4 0.7 (1.0) 1.35 0.79 11 0.4 44.5 24.5 21.1 6.7 2.8 1.0 (1.0) 1.00 0.69 Range for all items = 0–4 (with higher scores representing greater severity); range for complete 12-item scale = 0–48. Skew. = skewness; CoriT = corrected item-total correlations. View Large Table 2. Frequency distribution of responses (%), means (SD), skewness and item-total correlations for the items of the SSD-12 in the primary care sample (2016: Munich, Germany; N = 501) Item Missing Never (0) Rarely (1) Sometimes (2) Often (3) Very often (4) Mean (SD) Skew. CoriT Cognitive 1 0.4 34.6 34.2 24.5 4.3 1.9 1.0 (1.0) 0.71 0.68 4 0.4 29.2 24.3 27.1 14.2 4.7 1.4 (1.2) 0.38 0.67 7 1.7 53.1 22.2 16.1 4.9 1.9 0.8 (1.0) 1.18 0.31 10 0.4 66.2 17.6 13.1 1.9 0.6 0.5 (0.8) 1.54 0.44 Affective 2 0.2 20.0 38.5 29.9 9.0 2.4 1.4 (1.0) 0.45 0.70 5 0.2 36.8 31.0 21.9 7.7 2.4 1.1 (1.1) 0.74 0.81 8 0.4 48.0 23.9 18.1 6.7 3.0 0.9 (1.1) 1.01 0.75 12 0.2 44.1 31.6 17.4 4.9 1.7 0.9 (1.1) 0.88 0.81 Behavioural 3 0.2 39.4 32.7 18.3 7.1 2.4 1.0 (1.0) 0.89 0.77 6 0.4 50.1 32.5 9.5 5.6 1.9 0.8 (1.0) 1.37 0.76 9 0.4 55.3 23.7 14.2 4.1 2.4 0.7 (1.0) 1.35 0.79 11 0.4 44.5 24.5 21.1 6.7 2.8 1.0 (1.0) 1.00 0.69 Item Missing Never (0) Rarely (1) Sometimes (2) Often (3) Very often (4) Mean (SD) Skew. CoriT Cognitive 1 0.4 34.6 34.2 24.5 4.3 1.9 1.0 (1.0) 0.71 0.68 4 0.4 29.2 24.3 27.1 14.2 4.7 1.4 (1.2) 0.38 0.67 7 1.7 53.1 22.2 16.1 4.9 1.9 0.8 (1.0) 1.18 0.31 10 0.4 66.2 17.6 13.1 1.9 0.6 0.5 (0.8) 1.54 0.44 Affective 2 0.2 20.0 38.5 29.9 9.0 2.4 1.4 (1.0) 0.45 0.70 5 0.2 36.8 31.0 21.9 7.7 2.4 1.1 (1.1) 0.74 0.81 8 0.4 48.0 23.9 18.1 6.7 3.0 0.9 (1.1) 1.01 0.75 12 0.2 44.1 31.6 17.4 4.9 1.7 0.9 (1.1) 0.88 0.81 Behavioural 3 0.2 39.4 32.7 18.3 7.1 2.4 1.0 (1.0) 0.89 0.77 6 0.4 50.1 32.5 9.5 5.6 1.9 0.8 (1.0) 1.37 0.76 9 0.4 55.3 23.7 14.2 4.1 2.4 0.7 (1.0) 1.35 0.79 11 0.4 44.5 24.5 21.1 6.7 2.8 1.0 (1.0) 1.00 0.69 Range for all items = 0–4 (with higher scores representing greater severity); range for complete 12-item scale = 0–48. Skew. = skewness; CoriT = corrected item-total correlations. View Large Factorial validity Confirmatory factor analysis revealed good fit indices for a general-factor model [n = 465, CFI = 0.98, TLI = 0.97, RMSEA = 0.098, 90% confidence interval (CI): 0.087–0.109], but also for a model which includes the three latent dimensions as proposed by the theoretical conceptualization of SSD (n = 465, CFI > 0.98, TLI > 0.98, RMSEA = 0.090, 90% CI: 0.078–0.102). The three factor model is displayed in Figure 1. Given the high correlations between the three sub-criteria (r = 0.92–0.96), there appears to be substantial overlap in content between them. Altogether, the results support, therefore, a general factor model of the SSD-12 in primary care which is also in line with previous results from specialized care and the general population (11,12). Figure 1. View largeDownload slide Path diagram illustrating the three-factor-model estimates (n = 465). Figure 1. View largeDownload slide Path diagram illustrating the three-factor-model estimates (n = 465). Construct validity Given the evidence to suggest the general factor model of the SSD-12 is appropriate, an overall sum-score of the SSD-12 was calculated for further analyses. The SSD-12 sum score was significantly correlated with other well-established scales which are evidence of construct validity, that is the PHQ-15 (somatic symptom burden), the PHQ-9 depression scores and the GAD-7 anxiety scores. Note however, that the relatively high correlations with anxiety and depression can also suggest problems in terms of discriminative validity. The series of correlational analyses are shown in Table 3. Table 3. Descriptive characteristics and Pearson correlations between SSD-12 sum-score and other scales in the primary care sample (2016: Munich, Germany; N = 501) Descriptive characteristics SSD-12 PHQ-15 PHQ-9 GAD-7 N 465 424 462 464 Mean (SD) 11.43 (8.97) 4.74 (4.03) 5.10 (4.86) 3.91 (4.08) Range 0–47 0–26 0–27 0–21 Correlation with SSD-12 score  Correlation co-efficient (r) 0.48 0.60 0.54 Descriptive characteristics SSD-12 PHQ-15 PHQ-9 GAD-7 N 465 424 462 464 Mean (SD) 11.43 (8.97) 4.74 (4.03) 5.10 (4.86) 3.91 (4.08) Range 0–47 0–26 0–27 0–21 Correlation with SSD-12 score  Correlation co-efficient (r) 0.48 0.60 0.54 SSD-12, Somatic Symptom Disorder–B Criterion; PHQ-15, Patient Health Questionnaire Somatic Symptom Scale–15; PHQ-9, Patient Health Questionnaire Depression Scale–9; GAD-7, Generalized Anxiety Disorder Scale–7. All tests were significant after applying a Bonferroni adjustment for multiple testing. View Large Table 3. Descriptive characteristics and Pearson correlations between SSD-12 sum-score and other scales in the primary care sample (2016: Munich, Germany; N = 501) Descriptive characteristics SSD-12 PHQ-15 PHQ-9 GAD-7 N 465 424 462 464 Mean (SD) 11.43 (8.97) 4.74 (4.03) 5.10 (4.86) 3.91 (4.08) Range 0–47 0–26 0–27 0–21 Correlation with SSD-12 score  Correlation co-efficient (r) 0.48 0.60 0.54 Descriptive characteristics SSD-12 PHQ-15 PHQ-9 GAD-7 N 465 424 462 464 Mean (SD) 11.43 (8.97) 4.74 (4.03) 5.10 (4.86) 3.91 (4.08) Range 0–47 0–26 0–27 0–21 Correlation with SSD-12 score  Correlation co-efficient (r) 0.48 0.60 0.54 SSD-12, Somatic Symptom Disorder–B Criterion; PHQ-15, Patient Health Questionnaire Somatic Symptom Scale–15; PHQ-9, Patient Health Questionnaire Depression Scale–9; GAD-7, Generalized Anxiety Disorder Scale–7. All tests were significant after applying a Bonferroni adjustment for multiple testing. View Large Differential validity The comparison of the average SSD-12 scores from different patient groups (with and without chronic disease, with and without an indication of at least one psychological disorder and with and without indication of somatization) showed that the patients who suffer from a chronic disease (i.e. diabetes or hypertension) reported significantly higher SSD-12 scores than the patients without self-reported chronic disease. The same is true for patients with an indication of at least one psychological disorder or, more specifically, with somatization. Whereas the effect size measuring the differences in SSD-12 scores for patients with and without a chronic disease is rather small (d = 0.44), the differences in SSD-12 scores for patients with and without psychological disorders resulted in large effect sizes (d = 1.29 and 1.35, respectively). The comparison of the average SSD-12 scores from the different patient groups and the corresponding effect sizes are shown in Figure 2. Figure 2. View largeDownload slide Average SSD-12 sum-scores in different patient sub-groups from general practice. Figure 2. View largeDownload slide Average SSD-12 sum-scores in different patient sub-groups from general practice. Discussion The main aim of this study was to assess the usefulness and validity of the SSD-12 to measure the psychological characteristics of SSD in primary care. While there is already some evidence that supports the validity of the scale in specialized care and in the general population (11,12), it had not been applied to a primary care setting. Primary care practitioners represent an integral part of mental health care. As the patients’ usual first point of contact within the health care system, they not only connect patients to secondary and specialist services, but also have the unique opportunity to comprehensively evaluate the patients’ health and psychosocial context (16). The changes to the DSM-5 (and expected ICD-11) diagnostic criteria create a great opportunity to address the problem of under-diagnosis of somatoform disorders in the past. By emphasizing the psychological aspects associated with persistent somatic symptoms, it is possible that the new criteria may be more widely applied. This is especially likely as the diagnosis no longer relies on the idea that symptoms are ‘medically unexplained’. Previously, such uncertainty as to the cause of the symptoms often created unease in physicians who must balance the necessity of ruling out serious illness and increasing chronicity against the cost and distress of extensive testing (17). Primary care practitioners have the competing demands to treat and manage all or most of the patient’s acute and chronic conditions as well as providing preventive health care (17). Such competing demands create particular challenges for which the implementation of screening methods for mental disorders might be useful. This is particularly true as screening methods are time efficient and can open a discussion regarding the patient’s mental health. Since the vast majority of individuals with mental disorders never present to a mental health professional, mental health care is often undertaken at primary care (18). Therefore, there is a growing impetus to incorporate patient-reported outcome measures of symptoms and psychological distress into clinical practice (9). The SSD-12 could be a useful tool to assess psychological burden associated with bothersome somatic symptoms in primary care. Completion of the scale takes approximately 2–3 minutes, and the scoring is easily done within another minute. Its items are easy to understand and it measures a clear construct that patients feel is important to them (11). In the current study, the scale showed a high internal consistency and good item characteristics, apart from one item which showed somewhat problematic parameters. The disproportion of thoughts about the seriousness of one’s symptoms is very difficult for patients to judge themselves, but also for clinicians, especially since SSD can now also be accompanied by known diagnoses of somatic illness. Item 10 was explicitly included to reflect the opinions of physicians. This is in line with the Structured Clinical Interview for DSM-5 (SCID-5) (19), but its inclusion is at the expense of a higher heterogeneity of the scale. Confirmatory factor analyses performed on a ‘three latent dimensions’ model and on a ‘general-factor’ model both fit the data in an acceptable way. The calculation of three sub-scores is more in line with the structure of the DSM-5 criteria and this model might, therefore, be of greater importance for clinical purposes. However, since the three sub-criteria were demonstrably highly correlated in the current sample (Figure 1), a one-factor model suggestive of an overarching psychological burden factor should be favoured. This is in line with previous results from the general population (12). The high co