Access the full text.
Sign up today, get DeepDyve free for 14 days.
Background: Objective: To evaluate the usefulness of a clinical scorecard in managing sore throat in general practice. Design: Validation study of scorecard for sore throat with a throat swab culture used as the ‘gold standard’. Setting: A solo family practice in rural New South Wales, Australia Participants: Patients attending with sore throat. Methods: Patients from the age of 5 years and above presenting with the main symptom of a sore throat, and who have not had any antibiotic treatment in the previous two weeks, were invited to participate in the study. The doctor completed a scorecard for each patient participating and took a throat swab for culture. Adult patients (> 16 yrs) were asked to complete a patient satisfaction questionnaire, while guardians accompanying children (5 yr to < 16 yrs old) were asked to complete a similar, guardian questionnaire. Main outcome measures: 1. Ability of a new scorecard to differentiate between bacterial and non-bacterial sore throat. 2. Patients’ trust in the scorecard. Results: The scorecard has a sensitivity of 93.33%, a specificity of 63.16%, a positive predictive value of 50% and a negative predictive value of 96%. The sensitivity is better than other sore throat scorecards that have been published but with a slightly lower specificity. There was a high level of patient trust in the scorecard was (85.8% agreement). Patients also trusted their doctor’s judgement based on the scorecard (90.6% agreement). Conclusions: As the scorecard has a high sensitivity but only a moderate specificity, this means that it is more reliable for negative results, i.e. when the result suggests a viral infection. When the result favours a bacterial sore throat, then a high sensitivity can mean that there are a number of false positives. GPs can be confident in withholding antibiotics when the scorecard indicates a viral infection. Introduction Clinical scoring systems have been developed to help The management of sore throat in general practice is recognise bacterial or non-bacterial sore throats, e.g. traditionally based on the doctor’s clinical judgment and Centor’s and Breese’s criteria, which are based on 4 and empirical treatment. However, as the rate of prescribing 9 items respectively, and which only use clinical vari- remains quite high for a condition mostly due to viral ables (see Table 1). However, there are several short- causes, distinguishing between non-bacterial and bacter- comings to existing systems such as limitations in their ial causes of sore throat is still important . sensitivity (Breese 68%; Centor 65% to 83%) and their specificity (Breese 85%; Centor 67%-91%) . An accu- rate scorecard will remain valuable whenever alternative techniques for identifying bacterial or non bacterial * Correspondence: email@example.com causes, such as rapid antigen testing and throat swab Department of General Practice, School of Primary Health Care, Bldg 1, 270 culture, are unaffordable, unavailable or impractical. Ferntree Gully Rd, Notting Hill, Melbourne, Victoria 3168, Australia © 2010 Bakare and Schattner; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 2 of 7 http://www.apfmj.com/content/9/1/9 Table 1 Comparison of new scorecard with other scoring systems used in diagnosing group A streptococcal (GAS) pharyngitis Study Features included in final algorithm Age Sample Sensitivity Specificity Positive author group size (%) (%) predictive (year) (years) value(%) The Ten criteria (See Figure 1) 5 yrs to 106 93.3 63.2 50 current 99 yrs scorecard Breese‡ Season (late winter or early spring) Age (5-10 years) Elevated white Children 670 68 85 84 (1977) blood cell count Temperature > 38°C Sore throat Absence of cough Headache Pharyngeal erythema or oedema or exudates Tender or enlarged cervical lymph nodes Centor† History of fever Tonsillar exudates Tender and enlarged anterior > 15 234 65-83 67-91 56 (1981) cervical lymph nodes Absence of cough This study sought to ascertain whether a new score- necessarily appeal to the subjects for which it was card shows benefits over other ones published in the intended, in this case, the patients and carers. Hence, medical literature. For example, the widely used Centor the importance of also studying the acceptability of the scorecard has less than optimal sensitivity and specifi- scorecard. city. When tested in an urban emergency department using the four criteria, throat culture was positive in Methods 56% of these patients and in patients meeting three cri- Setting teria, the positive predictive value of a positive culture The study was conducted by the investigator, a solo was only 30% to 34% . family physician working in Cobar, a rural town in New The new scorecard was developed from ten criteria South Wales, Australia, with a population of close to six which included 4 from Centor (i.e. tonsilar exudates; thousand people. The investigator worked full time in tender anterior cervical adenopathy; fever by history; his medical practice which was open for 9 -10 hours and absence of cough). One criterion was taken from daily, six days a week. There were on average 30 to 50 Breese’s study  (season of illness - late winter or early patients per day, with over two hundred per week. The spring), and one from Wald’s study  (age 5-15 years). research took place at the investigator’s practice between The four remaining criteria were selected by the inves- April and December 2006 following ethics committee tigator as they seemed relevant to sore throat and were approval from Monash University. considered useful in increasing the accuracy of the scor- ecard, based on the author’s experience. These criteria Scorecard development included: The Centor criteria have provided the most widely used and accepted scorecard in America and in many other � History of exposure to group A streptococcal sore countries. The proposed new scorecard of ten factors, throat infection . therefore, contains Centor’s four factors and six addi- � Purulent nasopharyngeal discharge associated with tional factors as explained in the previous section. The sore throat . final version of the scorecard is presented in Figure 1. � Duration of illness > 5 days; longer duration of symptoms or illness may indicate a Strep A sore Questionnaire development throat . A patient/guardian perception questionnaire on sore � History of hoarseness or changes in voice - this throat was developed for this study using a five point symptom has been associated with a high negative Likert scale (5 Strongly agree, 4 Agree, 3 Neutral, 4 Dis- predictive value [9,10]. agree and 1 Strongly disagree). This questionnaire examined attitudes to and beliefs about the use of a Theinvestigatorwas of theviewthatbyincluding scorecard and also collected additional information such additional clinical features to the 4 Centor’s criteria, the as the severity of symptoms of the current sore throat. accuracy of a scoring system could be improved. The aim of this study was therefore to evaluate the validity Subject recruitment of a newly developed clinical scorecard in managing Patients were made aware of the study via a poster sore throat and also to determine the acceptability of placed in the waiting room and at the reception area. this scorecard by patients and carers. While an instru- Patients from the age of 5 years and above presenting ment may be successfully validated, it may not predominantly with a sore throat and who had not had Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 3 of 7 http://www.apfmj.com/content/9/1/9 Figure 1 The study scorecard. Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 4 of 7 http://www.apfmj.com/content/9/1/9 any antibiotic treatment in the previous two weeks were were analysed using histograms to plot the frequencies sequentially invited to participate. If they or their guar- of responses to the questionnaire against the 5 Likert dian agreed, one or the other then signed a consent scale responses. The Pearson’s product moment test was form. used to confirm that the correlation between the score- card and the throat swab results were within an accepta- Procedure ble range. The analysis was repeated using the A throat swab was taken from all patients participating non-parametric Spearman’srho test. Cronbach’salpha in the sore throat study by rubbing vigorously against was used to verify that the internal consistency of the each tonsil as recommended by Brien and Bass . guardian and adult questionnaires was within an accep- After taking the swab, the doctor completed a scorecard table range. for each patient participating in the study. Adult patients (>16 yrs) were asked to complete a patient Results questionnaire, while guardians accompanying children Subject participation (5 yr to <16 yrs old) were asked to complete a similar, Over a thousand patients were seen between April and guardian questionnaire. Patients either completed the December 2006, out of which one hundred and eighteen questionnaire while still in the consulting room or in patients presented with sore throat. All of these were the waiting room. This was then placed in a secure box approached to enter the study, but 12 patients were in the reception area or handed back to the doctor by excluded either because they had refused a throat swab the receptionist. or refused to complete a questionnaire. Participants in the study included twenty-four children Handling of throat swab under the age of 16 years who were accompanied by par- The throat swab was transported in an individually cov- ents or guardians (22 mothers, 1 father and 1 grand- ered Stewart’s transport medium to the regional pathol- mother), with fourteen of the children being male and ogy service in Dubbo, New South Wales, Australia, for with a mean age was 8.71 years (standard deviation 3.76, microscopy, culture and sensitivity. The throat swabs range 5 to <16 yrs). There were 82 adult patients, with 30 were forwarded within 24 hours of collection. of them being male. The mean age in the adult group Cultures showing any group A beta haemolytic Strep- was 34.85 (standard deviation 15.25, range 15 to 81). tococci, Streptococci groups B, C and G, and staphylo- cocci were considered positive. Cultures showing mixed Child group respiratory flora, Candida and non-specific growths were In the child group, 7 patients had a true positive result for considered negative. Undertaking viral studies was not bacterial infection (i.e. scorecard and microbiological cul- feasible in this setting and so the absence of bacterial ture in agreement) and none had a false negative result, i. growth was assumed to reflect a viral infection or other e. a sensitivity or true positive rate of 7/7, or 100% (Table minor causes such as allergy. 2). Further, 9 patients had a false positive result for non- bacterial infection (positive on scorecard and negative on Data analysis microbiological culture) and 8 patients had a true negative Patients were assigned to one of two groups based on (negative on scorecard and negative microbiological cul- the results of the scorecard: ture). This means that the specificity of the scorecard for the child group was 8/17 or 47.1%, giving an accuracy for � Bacterial group A: patients who had a bacterial the child group which is also 47.1%. score of ≥ 5 and a non-bacterial score of ≤ 4. � Non-bacterial group B: patients with a non-bacter- Adult group ial score of ≥ 6and abacteriascore of ≤ 4on the In the adult group, 21 patients had a true positive result scorecard. for bacterial infection, while 2 had a false negative The data obtained from the scorecard, the throat swab Table 2 Aetiology of sore throat: child group pathology reports from the laboratory and the question- Swab Number of patients naires were checked and then tabulated using the Score Card Bacterial Non-bacterial Microsoft Excel program. They were subjected to a Positive 7 9 16 recheck and analysis by a statistician. The sensitivity, (True Positive) (False Positive) specificity, predictive values and the accuracy of the Negative 0 8 8 scorecard were determined by comparing the scorecard (False Negative) (True Negative) findings with the gold standard of the throat swab. Data N 7 17 24 from the completed patient and guardian questionnaires Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 5 of 7 http://www.apfmj.com/content/9/1/9 Table 3 Aetiology of sore throat: adult group Table 5 Aetiology of sore throat for all subjects (combined) Swab N Swab Number of patients Score Card Bacterial Viral Score Card Bacterial Non-bacterial Positive 21 19 40 (True Positive) (False Positive) Positive 28 28 56 (True Positive) (False Positive) Negative 2 40 42 Negative 2 48 50 (False Negative) (True Negative) (False Negative) (True Negative) N23 59 82 N 30 76 106 (negative on scorecard but positive on microbiological Validation of the scorecard culture) result (See Table 3). This gives a true positive The Pearson product-moment correlation between the rate or sensitivity for bacterial infection in the adult scorecard and the throat swab for the guardian group group of 21/23 or 91.3%. Similarly, 19 patients had a was -0.454 (significant at the .05 level), and for the adult false positive result for non-bacterial infection while 40 group was -0.531 (significant at the .01 level). In general, patients had a true negative result. This means that the correlation coefficients of at least 0.7 are considered specificity for the adult group equalled 40/59 or 67.8%. ‘strong’, although the above results are acceptable. The accuracy of the scorecard for the adult group was Cronbach’s alpha for the guardian (i.e. child) group therefore 91.3% × 67.8% or 61.9%. was 0.685 and for the adult group was 0.670, and these Streptococcal group A was only identified in 11/106 or are acceptable values for internal reliability. 10% of the patients with sore throat (Table 4). The scor- ecard’s positive predictive value for bacterial sore throat Discussion was 50% in both adult and child groups and the negative The sensitivity of the scorecard was 93.3% and the spe- predictive value was 96% in both groups. The sensitivity cificity was 63.2%, giving an accuracy of 58.9%. This in the adult group was 91.3% as compared with 100% means that the scorecard gave a reliable answer in just sensitivity in the child group, with an overall sensitivity over half (i.e. 58.9%) of patients who presented with a of 93.3%. The specificity in the adult group was 67.8% sore throat. Most other scoring systems have either a as compared with a specificity of 47.1% in the child lower or at least no better accuracy. Although more sen- group, and with an overall specificity of 63.2%. The sitive, the new scorecard has a lower specificity (63.2%) accuracy of the scorecard was 58.9%. Patient of age less than other scoring systems, e.g. Breese (85%) and Centor than 16 years was the only criterion out of the ten in (67%-91%) . The new scorecard is therefore, with a the scorecard that emerged as an independent predictor sensitivity of 93.3%, more likely to confirm a bacterial of sore throat streptococcal infection. The combined sore throat than other previous scorecards, but a little group bacterial and non-bacterial positive and negative less likely to identify a ‘true negative’, which is a viral ratios are listed in (Table 5). sore throat. The implication of this in clinical practice is A total of 94.3% of patients agreed that they would that doctors may be more confident when the scorecard rely on theirdoctortodecide whetherornot they indicates a viral sore throat, but when it predicts a bac- needed an antibiotic. If there was a reliable scorecard, terial sore throat, given that the scorecard has a low 85.8% of patients would trust such a scorecard. specificity, (i.e. a high false positive rate) we cannot be absolutely confident that it is bacterial. Follow up of the patient may be required to ensure that the sore throat Table 4 Aetiology of sore throat by group resolves as would be expected for a viral cause, and if it Children Adult TOTAL does not, it should be treated as bacterial. Aetiology Male Female Male Female Apart from age less than 16 years, none of the other Normal throat flora 11 6 19 40 76 ten criteria in the scorecard emerged as an independent predictor of bacterial infection for sore throat. Strep A 2234 11 There are several limitations to this study. First, the cur- Strep B 0110 2 rent study was conducted by only one investigator with a Strep C 0135 9 fairly small number of subjects (106). Hence further stu- Strep G 0022 4 dies may be necessary to confirm the findings. Compar- Strep pyogenes 0001 1 able studies had higher numbers of subjects, for example, Staph aureus 1020 3 there were 234 in Centor’s study , 670 in Breese’s , Total number of subjects 14 10 30 52 106 513 in McIsaac’s  and 418 in Walsh’s . Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 6 of 7 http://www.apfmj.com/content/9/1/9 No sample size calculation was undertaken to deter- Ideally, the scorecard would be both sensitive and mine how many patients with sore throat are required specific. However, given that the problem with antibio- to develop a more accurate scorecard than existing tic use in general practice tends to be overuse, it ones. In that sense, this study must be considered to be would be preferable if the scorecard identified few a pilot study. false positives. Unfortunately, given the relatively low Other researchers have used only bacterial throat swab specificity, “positive” results may give more false posi- cultures to investigate sore throats and have not done tives than ideal. On the other hand, a “negative” result viral cultures on pragmatic grounds [15,16]. The diagno- is clinically significant in that it tends to rule out sis of viral infection is therefore presumptive. The same bacterial sore throat because the new scorecard has a applies to this study. sufficiently high sensitivity (93.3%), i.e. low false nega- The original wording of two of the items in the score- tives. In summary, the scorecard is most useful when card could have been clearer. Questions 7 and 10 were the result is negative. stated negatively and were therefore reverse scored in Further research is required to study the attitudes of the analysis. They could have been positively stated in doctors to the scorecard at the point of care. However, the first place and then reversely scored if necessary. trust by patients was high. Further, there were limitations with criterion 4 which referred to a history of exposure to a person with bac- Acknowledgements terial sore throat infection (e.g. GAS, Group C, Strep Our gratitude to: All patients who participated in the study. The staff at pyogenes). It is of course hard to prove the cause of a Orana Pathology, Dubbo, NSW, for providing pathology support and advice. A/Prof Rosemary Clerehan, Dr Sheila Vance, Vicky Ryan and staff at the sore throat from the taking of a history. This criterion School of Primary Care, Monash University, who assisted with the research was often scored as a “No” in doubtful cases (i.e. non- program. Professor Leon Piterman for early support in developing the study. bacterial). This might have skewed the overall scorecard Authors’ contributions result toward the non-bacterial causes, but the extent of TB conceived, designed, and conducted the research and acquired the data such an effect is unknown. It should be noted that the for the study at his general practice in rural New South Wales, Australia. TB scorecard cannot differentiate between Streptococcal conducted the analysis and interpretation of data. PS further contributed to the design of the study, the analysis and interpretation of data and its groups A, B, C or G based on clinical presentation. clinical applications. Both TB and PS were involved in drafting the The usefulness in patients under 5 years of age is manuscript and revising it critically for important intellectual content and unknown as they were excluded from the study, given have given final approval of the version to be published. that it would have been too difficult to take throat Competing interests swabs from them. The principal author (TB) is the GP of the patients recruited into the study Although patient trust in the use of the scorecard is who were given written information about the project and were able to refuse the invitation. Otherwise, the authors declare that they have no high, there is some evidence that where laboratory tests competing interests. are available and practical, such as the rapid antigen test, patient trust in their doctor’sjudgement,withor Received: 14 September 2009 Accepted: 29 July 2010 Published: 29 July 2010 without a scorecard, is diminished . However, as the use of these tests remains severely limited, clinical scor- References ecards such as the new one can be helpful as a practical 1. Danchin MH, Curtis N, Nolan TN: Treatment of sore throat in light of the tool for decision making in general practice. Cochrane verdict: is the jury still out? Med J of Australia 2002, 177:512-515. 2. Graham TAD: Diagnosis and treatment of pharyngitis. Can J of Emergency Conclusions Medicine 2002, 4:429-430. The sore throat scorecard presented in the current 3. Bisno AL: Diagnosing strep throat in the adult patient: do clinical criteria really suffice? Ann Intern Med 2003, 139(2):150-151. study appears to be more sensitive than several others 4. Breese BB: A simple scorecard for the tentative diagnosis of published in the literature. However, its specificity is a streptococcal pharyngitis. Am J Dis Child 1977, 131:514-517. little less. While all the other scoring systems focused 5. Wald ER, Green MD, Schwartz B, Barbadora K: A streptococcal score card revisited. Ped Emerg Care 1998, 14:109-111. on Streptococcal group A infection only, the new score- 6. Kahan S, Smith EG: In A Page: Signs & Symptoms. Lippincott Williams and card also included other bacterial organisms as Wilkins Philadelphia 2004. pathogens. 7. Hernanadez CG, Singleton JK, Aronzon DZ: Primary care paediatrics. Lippincott Williams and Wilkins Philadelphia 2001. The scorecard could predict non-bacterial sore throat 8. Zwart S, Sachs APE, Ruijs GJH, Gubbels JW, Hoes AW, Meker RAD: Penicillin in 96% of cases when the bacterial score was less than for sore throat: randomised double blind trial of seven days versus three or equal to four (i.e. its negative predictive value). How- days treatment or placebo in adults. BMJ 2000, 320:150-154. 9. Boruchoff SE, Weinstein MP: Throat cultures and other tests for the ever, the positive predictive value of the scorecard was diagnosis of pharyngitis. 2006 [http://www.uptodate.com]. 50%, meaning that the scorecard could predict bacterial 10. Guidelines & Protocols Advisory Committee: Diagnosis and Management sore throat infection in about 50% of the cases when the of Sore Throat. British Columbia Medical Association. British Columbia 2003 [http://www.health.gov.bc.ca/gpac/guideline_throat.html]. bacterial score was greater than or equal to five. Bakare and Schattner Asia Pacific Family Medicine 2010, 9:9 Page 7 of 7 http://www.apfmj.com/content/9/1/9 11. Brien JH, Bass JW: Streptococcal pharyngitis: optimal site for throat culture. J Pediatr 1985, 106:781-783. 12. 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, 1:239-246. 13. McIsaac WJ, White D, Tannenbaum D, Low DE: A clinical score to reduce unnecessary antibiotic use in patients with sore throat. Can Med Association J 1998, 158:75-83. 14. Walsh BT, Bookheim WW, Johnson RC, Tompkins RK: Recognition of streptococcal pharyngitis in adults. Arch Intern Med 1975, 135:1493-1497. 15. Gerber AM, Shulman TS: Rapid diagnosis of pharyngitis caused by group A streptococci. Clin Microbiology Review 2004, 17:571-580. 16. Bartlett JG: Approach to acute pharyngitis in adults. 2006 [http://www. uptodate.com]. 17. Worrall G, Hutchinson J, Sherman G, Griffiths J: Diagnosing streptococcal sore throat in adults. Can Fam Phys 2007, 53:666-671. doi:10.1186/1447-056X-9-9 Cite this article as: Bakare and Schattner: The usefulness of a clinical ‘scorecard’ in managing patients with sore throat in general practice. Asia Pacific Family Medicine 2010 9:9. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
Asia Pacific Family Medicine – Springer Journals
Published: Jul 29, 2010
Access the full text.
Sign up today, get DeepDyve free for 14 days.