Abstract Background Primary care test requests for serum immunoglobulins are rising rapidly, with concerns that many requests may be unnecessary. Evidence suggests some characteristics of general practitioners (GPs) and practices are associated with higher test ordering. Objective To identify the physician and practice characteristics associated with immunoglobulin test ordering. Methods Retrospective, cross-sectional study using routine laboratory data on primary care serum immunoglobulin requests. Data were linked with GP patient list size data. The primary outcome measure was the count of test requests per GP. Predictor variables were physician gender, years experience, practice region and type (number of GPs), GP patient list size and composition. Mixed-effects multilevel regression models were used to calculate incidence rate ratios (IRRs) with 95% confidence intervals (CIs) for the associations between physician and practice characteristics and GP requesting. Sensitivity analysis was performed by limiting the model to the more than 70 years age category. Results In total, 5990 immunoglobulin tests were ordered by 481 GPs in the South of Ireland during 2013. The number of tests ordered by individual GPs varied from one to 377. In the final fully adjusted Poisson regression analysis, female gender (IRR: 1.81; 95% CI: 1.45–2.26) and less experience (IRR: 2.27; 95% CI: 1.47–3.51) were associated with higher requesting (P < 0.001). None of the practice factors were associated with test ordering. Sensitivity analysis on the 70 years or more age category found similar results. Conclusion Further research is required to explore the potential reasons for higher requesting among GPs with fewer years of experience and also among female GPs. Blood tests, general practice, general practitioners, inter-doctor variation, physician’s practice patterns, test ordering Introduction Laboratory testing is the single highest volume medical activity in healthcare and demand is rising year on year in many countries (1–3). General Practitioners (GPs) initiate an estimated 50% of all requests (4), while approximately 30% of all primary care patient visits in the US result in a laboratory test order (5). Over the past 20 years, the number of new laboratory tests available to GPs has increased rapidly (5). The volume of orders from GPs has also risen significantly for the most commonly ordered tests (3). Given the increasing financial pressure on health systems, judicious laboratory testing is imperative. A better understanding of the determinants of GP test ordering is necessary, as evidence suggests that between 25% and 40% of all test orders are unnecessary (6,7). Differences in test ordering patterns of GPs has been linked to both physician and practice level characteristics (8). At the physician level, higher test ordering has been reported in female GPs (9,10) and GPs with less medical experience (11). At the practice level, both practice type and practice setting have been linked to GP test ordering patterns. Those in group practices have been shown to order significantly fewer tests than GPs in single handed or two-person practices (12). Working in an urban practice has been positively linked to higher test ordering (10). However, previous research has not considered the interaction of physician and practice factors in a multilevel model. Existing literature also does not account for individual GP patient list size or the composition of lists. This study focuses on a low volume test, serum immunoglobulins, which poses a significant challenge to GPs (13). Moreover, a local audit of one year of serum immunoglobulin test requests found that these tests were ordered for screening in many instances rather than for suspected plasma cell dyscrasias (myeloma, lymphoma, chronic lymphatic leukaemia, heavy chain disease, and amyloidosis) or investigation of causes of immune deficiency (14). Depending on the condition (e.g. myeloma), serum immunoglobulin tests were also ordered periodically to monitor disease progression (14). In the UK, one of the largest increases in GP test use between 2005 and 2009 were for serum immunoglobulins with a relative increase of 73.4%, from 61 tests per GP in 2005 to 106 tests per GP in 2009 (3). Immunoglobulins are particularly problematic in primary care (13). High levels are a feature of many clinical conditions in older patients, but their greatest diagnostic value is in specific haematological disorders such as myeloma and lymphoma. Unfortunately, clinical features of these disorders can be vague and non-specific and overlap with the symptoms of a wide range of other conditions. Thus, in older patients, immunoglobulin testing is probably best undertaken as a second-line investigation where there are other tests (such as a full blood count) which indicate the possibility of a blood dyscrasia. Knowing when to order immunoglobulins can be challenging for GPs and may require a clinical judgement in the context of non-specific clinical features. Interpretation of test results is also difficult and often requires specialist input (13). In particular, the rarity of the conditions that immunoglobulins are aimed at, suggests that factors such as GP gender, medical experience and practice setting or number of GPs at a practice are all likely to be important drivers of using this test (13). To date, no research has been carried out on physician and practice level predictors of serum immunoglobulin testing. The aim of this study was to identify the relationship between physician and practice characteristics and the volume of serum immunoglobulins requested among GPs in the South of Ireland, adjusting for the size and composition of GP public patient lists. Methods Study design and population We conducted a cross-sectional study using one year (2013) of routine laboratory data to analyse the determinants of serum immunoglobulin test requests. The study was conducted in 2 adjacent counties (Cork and Kerry) in the South of Ireland which have a combined population of 664534. All serum immunoglobulin analyses for these counties are processed at Cork University Hospital (CUH) and all GP test requests for the region are captured in our study. Variables The dependent variable was the 2013 count of immunoglobulin test requests per GP. Information about the number of publically funded patients cared for by each GP was also included in the model. This was used as a proxy for clinician workload and was used to make the request volumes of different GPs comparable. We also included the age and gender composition of publically funded patient lists in our model as the conditions associated with serum immunoglobulins are more common in older and male patients (15). Physician characteristics At the physician level, we collected data on the following predictors of test ordering, the GPs’ gender and their medical experience (years since graduation). The number of years since the GP graduated with their medical degree was used as a proxy for the GPs clinical experience as per previous research (11,16,17). Practice characteristics Practice level variables were the practice setting (urban, rural) and type of practice (single handed, 2–4 GPs or >4 GPs). Rural is defined according to Irish census data as areas with no population centre above 1500 people, with a population density below 150 per square km, and which are not part of an urban district (18). The number of GPs practising at a given practice was used to describe the practice type as per previous research (12). Data collection and cleaning Immunoglobulin test order counts and data about GPs and their practices were extracted for 2013 using Cognos Impromptu software to interrogate the hospital’s APEX laboratory system. The data fields requested were: request date, specimen number, test code, patient age, and gender, requesting GP, the location of test request, GP gender, surgery name, and address. The extracted data were exported from Cognos using Excel and imported into Stata v12 for analysis. Data on GP list size and the gender and age composition of these lists were obtained from the Health Service Executive Primary Care Reimbursement Scheme (HSE-PCRS). In Ireland, patients under a certain income level and more than 70 years of age, are entitled to free health care through the General Medical Services (GMS) scheme. Those participating in this scheme are reimbursed by the HSE-PCRS for a range of services they provide to these patients. Information held in the HSE-PCRS database is retrievable at individual patient level using the GPs unique Medical Council record number. This number was used to merge HSE-PCRS data with the immunoglobulin laboratory data extracted at CUH. In 2013, 498 GPs were contracted to the GMS in the Cork-Kerry region and were registered to receive reimbursement for providing services to 265000 eligible GMS patients (19). Of these patients, 54988 were over 70 years of age, representing approximately 97% of all those over the age of 70 who live in Cork and Kerry. Data on each GP’s medical experience was obtained from the Medical Council, the statutory, regulatory and registration body for doctors in Ireland. Medical experience was defined as years since GPs qualified with their medical degree and was coded to ‘<10 years’, ‘10–20 years’ and ‘>20 years’ as per previous research (11). Data analysis Mixed-effects multilevel regression models were used to calculate incidence rate ratios (IRRs) with 95% confidence intervals (CI) for the associations between GP and practice characteristics and requesting patterns for serum immunoglobulins. Multilevel analyses were used to take into account the structure of the data: physicians were nested within practices. The model, therefore, consists of 2 levels: the physician level (level 1) and the practice level (level 2). The outcome (test requests) was modelled as a count variable with a conditional Poisson distribution (20). Calculating variance estimates from random effects Poisson regressions rather than directly from the observed rates is a more statistically rigorous approach, which appropriately adjusts for lack of independence in requesting rates for GP in the same practice. Estimated regression coefficients were expressed as IRRs with 95% confidence intervals (CI). To analyse the extent to which the final model was sensitive to the impact of not including non-GMS patients in the study a sub-analysis was performed by limiting the model to include the 70 years of age or more category only. Data were analysed using Stata v12. Results Sample characteristics In total, 481 primary care physicians requested 5990 serum immunoglobulins during 2013. This represents 96.6% of all GPs registered in the GMS scheme in the Cork-Kerry region in 2013. The remaining GPs may not have requested any immunoglobulins or were not active at the time. Table 1 provides the GP requesting patterns by physician and practice characteristics (both crude and adjusted for their standardised patient list sizes). Table 1. General Practitioner immunoglobulin test ordering patterns in the South of Ireland by physician and practice characteristics, 2013 (crude and adjusted for patient list sizes) Total GPs Total IG counts Mean per GPa (crude) Standard deviation Mean per 1000 GMS patientsb Totals N = 481 N = 5990 12.5 27.5 43.9 Physician level GP Gender Males 230 3206 13.0 28.6 35.5 Females 251 2784 12.1 26.6 60.9 Experience <10 years 68 519 7.6 9.2 49.5 10–20 years 125 1108 9.1 10.1 36.9 >20 years 266 4132 15.6 35.2 45.8 List size <500 299 3649 8.9 19.9 44.0 500–1000 166 1875 31.7 50.4 43.7 >1000 16 466 51.7 34.1 44.2 Practice level Location Urban 316 3858 12.3 24.9 44.7 Rural 165 2132 13.0 32.1 42.3 Practice type >4 GPs 201 2239 11.3 28.9 43.9 2–4 GPs 178 2278 12.8 18.2 41.9 Single handed 98 1435 14.9 37.7 47.9 Total GPs Total IG counts Mean per GPa (crude) Standard deviation Mean per 1000 GMS patientsb Totals N = 481 N = 5990 12.5 27.5 43.9 Physician level GP Gender Males 230 3206 13.0 28.6 35.5 Females 251 2784 12.1 26.6 60.9 Experience <10 years 68 519 7.6 9.2 49.5 10–20 years 125 1108 9.1 10.1 36.9 >20 years 266 4132 15.6 35.2 45.8 List size <500 299 3649 8.9 19.9 44.0 500–1000 166 1875 31.7 50.4 43.7 >1000 16 466 51.7 34.1 44.2 Practice level Location Urban 316 3858 12.3 24.9 44.7 Rural 165 2132 13.0 32.1 42.3 Practice type >4 GPs 201 2239 11.3 28.9 43.9 2–4 GPs 178 2278 12.8 18.2 41.9 Single handed 98 1435 14.9 37.7 47.9 *Maximum volume of immunoglobulin requests per GP per year. aMean volume of immunoglobulin requests per GP per year. bStandardised for GP GMS list sizes and composition (patient age and gender adjusted)—GMS patients are public patients receiving free GP care, IG: immunoglobulins. View Large Overview of GP requesting patterns The crude mean count of test requests per GP for 2013 was 12.5 (SD: 25.5; range 1 to 377). The mean number of test orders was 43.9 per 1000 public patients. Allowing for GP public list size had the greatest impact on requesting patterns by gender of the GP. Crude count data indicate female GPs requested fewer immunoglobulin tests (mean: 12.1) compared to male GPs (mean: 13.0). However, when we adjust for GP public patient list size and demographics, female GP test ordering rates are higher than males (60.9/1000 patients versus 35.5/1000 patients, respectively). Similarly, crude count data alone indicate the GPs with less than 10 years’ experience (mean: 7.6) order fewer tests than those with 10–20 years (mean: 9.1) or those with more than 20 years’ experience (mean: 15.6). GPs with less than 10 years’ experience had fewer patients, and when GMS list sizes and demographics were adjusted for, they had higher test ordering rates (49.5/1000 patients). Requesting patterns by patient age and gender In total, 30% (n = 1809) of immunoglobulin tests were requested for patients in the 70 years or more age category, 25% (n = 1490) for patients in the 60–70 years age category, 23% (n = 1377) for 45–60 years age category, 16% (n = 929) for 30–45 years age category and 6% (n = 385) for the under 30 years age category. 61% (n = 3675) of test requests were for female patients. Association between GP/practice characteristics and immunoglobulin test ordering Table 2 provides the results of the mixed-effects multilevel Poisson regression analysis. In model 1 (adjusted for GP gender and medical experience), all of the physician factors were associated with immunoglobulin requesting rates. Model 2 (adjusted for the physician factors plus practice location and practice type) also found all of the physician factors to be associated with GP requesting rates for immunoglobulins. In particular, GP gender and medical experience were positively associated with immunoglobulin test ordering rates. Female GP requesting rates were 81% greater than male GPs (IRR: 1.81; 95% CI: 1.45–2.26, P < 0.001). GPs with less than 10 years medical experience were also more likely to request immunoglobulin tests (IRR: 2.72; 95% CI: 1.47–3.51, P < 0.001). There was a correlation between gender and years of experience, whereby more females were less experienced compared to males. However, there was no significant interaction between gender and years of experience in the final model. Table 2. Physician and patient level factors associated with General Practitioner immunoglobulin test ordering patterns in the South of Ireland, 2013 Model 1 (physician only) Model 2 (physician + practice) Variables* IRRa (95% CI) P valueb IRRa (95% CI) P valueb Physician level GP gender <0.001 <0.001 Female 1.80 (1.45–2.34) 1.81 (1.45–2.26) Male 1 1 Experience 0.003 0.001 <10 years 2.12 (1.39–3.23) 2.27 (1.47–3.51) 10–20 years 1.09 (0.86–1.39) 1.17 (0.80–1.42) >20 years 1 1 Practice level Location 0.26 Rural 0.88 (0.70–1.10) Urban 1 Practice type 0.52 >4 GPs 1.17 (0.88–1.56) 2–4 GPs 1.07 (0.80–1.42) Single handed 1 Model 1 (physician only) Model 2 (physician + practice) Variables* IRRa (95% CI) P valueb IRRa (95% CI) P valueb Physician level GP gender <0.001 <0.001 Female 1.80 (1.45–2.34) 1.81 (1.45–2.26) Male 1 1 Experience 0.003 0.001 <10 years 2.12 (1.39–3.23) 2.27 (1.47–3.51) 10–20 years 1.09 (0.86–1.39) 1.17 (0.80–1.42) >20 years 1 1 Practice level Location 0.26 Rural 0.88 (0.70–1.10) Urban 1 Practice type 0.52 >4 GPs 1.17 (0.88–1.56) 2–4 GPs 1.07 (0.80–1.42) Single handed 1 *Final fully adjusted model. aCluster grouping: primary care practices, total groups: 214. bP values based on likelihood test ratio. View Large There were no statistically significant associations between the practice-level factors (location or practice type) and GP immunoglobulin test ordering rates. Sensitivity analysis Table 3 presents the results of the sensitivity analysis on the 70 years or more age category (fully adjusted model). Results were similar to that of the full model including all patient age groups. The strongest difference was among females, where the strength of the association was greater in the subgroup analysis compared with the full dataset. Based on the fully adjusted model, female GPs requested over twice as many tests compared to males (IRR: 2.37; 95% CI: 1.76–3.18). Experience also remained strongly associated with test ordering among patients over the age of 70. Those with less than 10 years’ experience were more than twice as likely to order a test compared to those with more than 20 years’ experience (IRR: 2.55; 95% CI: 1.43–4.52). Table 3. Physician and patient level factors associated with General Practitioner immunoglobulin test ordering patterns among patients over the age of 70 in the South of Ireland, 2013 Model 2 (physician + practice) Variables* IRRa (95% CI) P valueb Physician level GP gender <0.001 Female 2.37 (1.76–3.18) Male 1 Experience <0.001 <10 years 2.55 (1.43–4.52) 10–20 years 1.34 (0.97–1.86) >20 years 1 Location 0.17 Rural 0.81 (0.60–1.10) Urban 1 Practice level Practice type 0.18 >4 GPs 1.37 (0.93–2.00) 2–4 GPs 1.07 (0.73–1.58) Single handed 1 Model 2 (physician + practice) Variables* IRRa (95% CI) P valueb Physician level GP gender <0.001 Female 2.37 (1.76–3.18) Male 1 Experience <0.001 <10 years 2.55 (1.43–4.52) 10–20 years 1.34 (0.97–1.86) >20 years 1 Location 0.17 Rural 0.81 (0.60–1.10) Urban 1 Practice level Practice type 0.18 >4 GPs 1.37 (0.93–2.00) 2–4 GPs 1.07 (0.73–1.58) Single handed 1 *Final fully adjusted model. aCluster grouping: primary care practices. bP values based on likelihood test ratio Groups: 174. View Large Discussion Summary of main findings This study evaluated the relationship between certain physician and practice characteristics and the use of serum immunoglobulin tests over 1 year. The strongest predictors of test requests were physicians’ gender and medical experience (years since graduating with a medical degree). No associations were found between practice-level factors and immunoglobulin test ordering rates. At the physician level, female gender was associated with higher serum immunoglobulin test ordering rates. This is consistent with previous studies examining the characteristics of GPs and test ordering behaviour (10,21). Further, research has found that female GPs also have a higher level of referrals to secondary care compared to male GPs (22). With the proportion of female physicians rising in most European countries (23), is important to investigate why females may be ordering more tests referring more patients than their male counterparts. Previous research suggests that higher test ordering rates of female GPs may relate to differences in practice styles (24) and patient demographics (25). For example, 61% of test orders were for female patients, and evidence suggests that women prefer to be seen by female GPs (25). In Ireland, a recent national survey also reported that female GPs see fewer patients on average per session than males (26). The same report found less female GPs practice full time, compared to males (70% versus 100% in the under 40 age group) and also retire at a younger age (26). It is possible that the experience of part-time GPs with certain low volume tests falls below a critical threshold, which leads them to increased ordering. GPs with less than 10 years’ experience were also significantly higher requesters in this study. Previous research also found a statistically significant inverse association between years of experience and test ordering (11). While this study found more females were less experienced, this did not explain the relationship between experience and test ordering in this sample. In a previous qualitative study with GPs from this sample, it was suggested that younger GPs may be more likely to request serum immunoglobulins (13). Potential reasons for this include a lack of confidence in their clinical judgement at this stage of their career, recent clinical experience in a hospital setting with unrestricted readily available laboratory access and a fear of missing serious cancers such as myeloma (13). The fear of missing a diagnosis or the impact of test ordering on patient care may also play an important role in test ordering behaviour. Unfortunately, this is more difficult to study and outside the scope of this paper. However, using data from this paper, along with a previous qualitative paper and findings of an audit on the clinical impact of IG testing in this sample of GPs, we have implemented a lab-based guideline and education based strategy which has shown to be effective at reducing test ordering 1 year after implementation. Once enough time passes, we aim to reassess the clinical impact of the intervention through further audit. This study found no significant association between practice location (urban/rural), practice type (single handed, 2–4, or 4 or more GPs) and GP test ordering patterns. This is in contrast with previous research reporting 18% fewer tests among GPs in group practices compared to those in single handed or two-person practices (12). Working in an urban practice rather than a rural practice has also previously been linked to higher test ordering patterns (10). It is possible that the practice level effects seen in previous studies were driven by physician level effects such as the gender and experience of GPs: these effects have been accounted for in our study. For example, it is possible that the higher test ordering rates by urban practices was driven by the tendency for younger or less experienced GPs to work in these settings (26). Strengths and limitations Unique strengths of this study are the inclusion of all GP requests for immunoglobulin tests in 2 large adjacent regions with a population in excess of 650000, the use of patient list size and composition to allow for the fair comparison of different GPs and the use of multilevel modelling to separate GP and practice effects. An important limitation of our study is the use of HSE-PCRS data to characterise GP list size and composition. This data excludes private patients for which no accurate data is available in Ireland. According to 2014 figures, 58% of the Irish population were private patients and 42% were on the GMS scheme (free GP care) (19). However, a sensitivity analysis, which included only patients over the age of 70, 97% of whom are covered by the GMS scheme, found the same physician factors to be associated with immunoglobulin test ordering rates. A further limitation is our exclusion of other predictor variables which may influence test requesting patterns such as the presence of an electronic ordering system in the practice (27). Finally, a limitation of using routine medical record data is the possibility that unknown or unmeasured confounding may have affected our results. Conclusions Our research suggests that further qualitative research is required to explore why female GPs and less experienced GPs order greater volumes of serum immunoglobulins. In addition, potential strategies identified in our previous systematic review (28) targeting test use in primary care should consider the differences in test ordering patterns amongst these 2 groups. It must be emphasised that our study was not designed to detect inappropriate test ordering. It is possible that female GPs and those with less medical experience are behaving with an appropriate level of caution. The consequences of finding (or missing) pathology are very significant for the doctor and patient. Further study of the clinical utility of serum immunoglobulin test orders is required and currently ongoing in our study region. Declaration Ethical approval: the clinical research ethics committee of the Cork University Teaching Hospitals (ref: ECM (ii) 07/01/14). Funding: the first author (SLC) is funded by the Health Research Board in Ireland under the Scholars Programme in Health Services Research Grant No. PHD/2007/16. Conflicts of interest: none. Acknowledgements We would like to thank the laboratory staff at Cork University Hospital, in particular, Mr Aidan Kelleher for facilitating the data extraction and Mr Brendan O’Reilly for providing Cognos Impromptu software training for SLC. References 1. Britt H, Miller GC, Henderson Jet al. A decade of Australian general practice activity 2003–04 to 2012–13: General practice series no. 34 . Sydney: Sydney University Press; 2013. 2. McGregor MJ, Martin D. Testing 1, 2, 3: is overtesting undermining patient and system health? Can Fam Physician 2012; 58: 1191– 3, e615–7. Google Scholar PubMed 3. Busby J, Schroeder K, Woltersdorf Wet al. Temporal growth and geographic variation in the use of laboratory tests by NHS general practices: using routine data to identify research priorities. Br J Gen Pract 2013; 63: e256– 66. Google Scholar CrossRef Search ADS PubMed 4. Bird C. Laboratory tests: guidance for general practitioners. Biomed Sci 2008; 52: 42. 5. Hickner J, Thompson PJ, Wilkinson Tet al. Primary care physicians’ challenges in ordering clinical laboratory tests and interpreting results. J Am Board Fam Med 2014; 27: 268– 74. Google Scholar CrossRef Search ADS PubMed 6. Rao GG, Crook M, Tillyer ML. Pathology tests: is the time for demand management ripe at last? J Clin Pathol 2003; 56: 243– 8. Google Scholar CrossRef Search ADS PubMed 7. Zhi M, Ding EL, Theisen-Toupal J, Whelan J, Arnaout R. The landscape of inappropriate laboratory testing: a 15-year meta-analysis. PLoS One 2013; 8: e78962. Google Scholar CrossRef Search ADS PubMed 8. Whiting P, Toerien M, de Salis Iet al. A review identifies and classifies reasons for ordering diagnostic tests. J Clin Epidemiol 2007; 60: 981– 9. Google Scholar CrossRef Search ADS PubMed 9. Ferrier BM, Woodward CA, Cohen M, Goldsmith CH. Laboratory Tests: Which Physicians Order More?: How women’s practices differ from men’s. Can Fam Physician 1991; 37: 349– 52. Google Scholar PubMed 10. Vinker S, Kvint I, Erez R, Elhayany A, Kahan E. Effect of the characteristics of family physicians on their utilisation of laboratory tests. Br J Gen Pract 2007; 57: 377– 82. Google Scholar PubMed 11. Salloum S, Franssen E. Laboratory investigations in general practice. Can Fam Physician 1993; 39: 1055– 61. Google Scholar PubMed 12. Verstappen WH, ter Riet G, Dubois WI, Winkens R, Grol RP, van der Weijden T. Variation in test ordering behaviour of GPs: professional or context-related factors? Fam Pract 2004; 21: 387– 95. Google Scholar CrossRef Search ADS PubMed 13. Cadogan SL, McHugh SM, Bradley CP, Browne JP, Cahill MR. General practitioner views on the determinants of test ordering: a theory-based qualitative approach to the development of an intervention to improve immunoglobulin requests in primary care. Implement Sci 2016; 11: 102. Google Scholar CrossRef Search ADS PubMed 14. Smellie WS, Wilson D, McNulty CAet al. Best practice in primary care pathology: review 1. J Clin Pathol 2005; 58: 1016– 24. Google Scholar CrossRef Search ADS PubMed 15. Jagannath S, Richardson P, Munshi NC. Multiple myeloma and other plasma cell dyscrasias. In: Razdur R, Coia LR, Hoskins WJ, Wagman LD (eds). Cancer management: a multidisciplinary approach. Darien Conn: The Oncology Group, 2005, pp. 749– 67. 16. Bugter-Maessen AM, Winkens RA, Grol RPet al. Factors predicting differences among general practitioners in test ordering behaviour and in the response to feedback on test requests. Fam Pract 1996; 13: 254– 8. Google Scholar CrossRef Search ADS PubMed 17. Eisenberg JM, Nicklin D. Use of diagnostic services by physicians in community practice. Med Care 1981; 19: 297– 309. Google Scholar CrossRef Search ADS PubMed 18. Brady Shipman M. Irish Rural Structure and Gaeltacht Areas . Dublin: DoELG, 2000. 19. Burke P. Primary Care Reimbursement Service; Statistical Analysis of Claims and Payments 2010 . Dublin: Health Service Executive, 2014. 20. Cameron AC, Trivedi PK. Regression analysis of count data . Cambridge, UK: Cambridge university press, 2013. Google Scholar CrossRef Search ADS 21. Kristiansen IS, Hjortdahl P. The general practitioner and laboratory utilization: why does it vary? Fam Pract 1992; 9: 22– 7. Google Scholar CrossRef Search ADS PubMed 22. Ingram JC, Calnan MW, Greenwood RJ, Kemple T, Payne S, Rossdale M. Risk taking in general practice: GP out-of-hours referrals to hospital. Br J Gen Pract 2009; 59: e16– 24. Google Scholar CrossRef Search ADS PubMed 23. Mayorova T, Stevens F, Scherpbier A, van der Velden L, van der Zee J. Gender-related differences in general practice preferences: longitudinal evidence from the Netherlands 1982-2001. Health Policy 2005; 72: 73– 80. Google Scholar CrossRef Search ADS PubMed 24. Bensing JM, van den Brink-Muinen A, de Bakker DH. Gender differences in practice style: a Dutch study of general practitioners. Med Care 1993; 31: 219– 29. Google Scholar CrossRef Search ADS PubMed 25. Phillips D, Brooks F. Women patients’ preferences for female or male GPs. Fam Pract 1998; 15: 543– 7. Google Scholar CrossRef Search ADS PubMed 26. O’Dowd T, O’Kelly M, O’Kelly Fet al. . Structure of general practice in Ireland: 1982–2015. Dublin: Irish College of General Practitioners, 2016. 27. Vardy DA, Simon T, Limoni Yet al. The impact of structured laboratory routines in computerized medical records in a primary care service setting. J Med Syst 2005; 29: 619– 26. Google Scholar CrossRef Search ADS PubMed 28. Cadogan SL, Browne JP, Bradley CP, Cahill MR. The effectiveness of interventions to improve laboratory requesting patterns among primary care physicians: a systematic review. Implement Sci 2015; 10: 167. Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org.
Family Practice – Oxford University Press
Published: Feb 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud