Family Practice Vol. 16, No. 3, 262-268
© Oxford University Press 1999
Criteria, performance and diagnostic problems in diagnosing acute otitis media
Department and Research Unit of General Practice, University of Aarhus, Vennelyst Boulevard 6, DK-8000 Aarhus C, Denmark.
Dr J Lous.
Peter M Jensen and Jørgen Lous. Criteria, performance and diagnostic problems in diagnosing acute otitis media. Family Practice 1999; 16: 262–268.
Received 26 January 1998; Revised 25 September 1998; Accepted 28 January 1999.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
Objective.We aimed to assess criteria when diagnosing acute otitis media and related performance in general practice in Denmark. Furthermore, we aimed to identify the scale of and the reasons for diagnostic uncertainty.
Method. We conducted: (i) a survey among GPs assessing criteria; and (ii) prospective registration of acute otitis-media-related consultations performed by GPs assessing performance. The survey was sent to all 790 GPs in Funen, North Jutland and Ringkøbing counties, Denmark. A total of 568 (72%) of all GPs in the three counties responded. A total of 368 children with acute otitis media or previous acute otitis media visiting 151 GPs were studied. The main outcome measures were: (i) criteria for symptoms and findings suggesting the diagnosis acute otitis media, criteria for use of equipment and reasons for diagnostic uncertainty; and (ii) prevalence of symptoms and findings in diagnosed cases, equipment used and multivariate analysis of factors predicting diagnostic certainty.
Results. The symptoms of earache, fever, reduced hearing, findings of bulging eardrum, red eardrum and purulent otorrhea were important criteria used during both diagnosis of acute otitis media by the GPs and assessment of performance. In the prospective study, diagnostic certainty of acute otitis media was 67% (95% CI 58–76) in children under 2 years and 75% (95% CI 69–81) in older children. Diagnostic certainty was statistically related (P < 0.05) to a good view of the eardrum and the findings of purulent otorrhea or a bulging eardrum. Logistic regression revealed that the two most important factors predicting diagnostic certainty were a satisfactory view of the eardrum, with an odds ratio (OR) 11.0 (95% CI 4.1–29.5), and purulent otorrhea OR 10.1 (95% CI 3.1–32.9). Main reasons for diagnostic uncertainty given by GPs were differential diagnostic doubts, insufficient view of the eardrum and lack of knowledge.
Conclusion.Danish GPs' criteria for the diagnosis of acute otitis media were stricter than criteria used internationally. The discrepancy between diagnostic criteria and performance was small. Diagnostic accuracy and certainty could be substantially improved by cleaning the ear canal when needed and by widespread use of pneumatic otoscopy.
Keywords. Denmark, diagnostic criteria, general practice, otitis media, performance.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
Examination for acute otitis media (AOM) is an everyday occurrence for the GP. Consensus has not been reached between countries with regard to diagnostic criteria. This divergence is demonstrated in different inclusion criteria used in clinical trials addressing treatment of AOM. Efforts have been made to reach consensus, resulting in several published reports over the years.1–4 Diagnostic criteria accepted by these consensus conferences2,3 are very broad and cannot serve as a basis for scientific studies of treatment, especially because distinguishing AOM from secretory otitis media is not sufficiently addressed. GPs' knowledge of diagnostic criteria is limited.
An international study5 has shown GPs to be certain of the diagnosis of AOM in 58% of children under 1 year, increasing to 73% in children older than two and a half years. Reasons for diagnostic uncertainty could not be extracted from the study. Diagnostic certainty is, however, influenced by the diagnostic criteria used, diagnostic equipment used, the GP's knowledge of the topic and practical examination problems such as cerumen or a crying child. Knowledge of the importance of factors compromising diagnostic certainty is essential when trying to improve performance.
Prevalence of antibiotic resistant pathogens in the upper respiratory tract has increased in the out-patient population6–8 with potential hazards associated with the future treatment of bacterial infections. The use of exact diagnostic criteria when diagnosing AOM could counteract this development, as the AOM diagnosis often implies prescription of antibiotics.
The aim of this study was to assess Danish GPs' implicit criteria and performance with regard to diagnosis of AOM. Furthermore, we aimed to identify reasons for diagnostic uncertainty.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
Definitions:
Explicit criteria. . These are specific criteria that have been established and agreed upon with regard to a specific issue (e.g. the AOM diagnosis), often published in official guidelines.9
Implicit criteria. . These are the criteria based on individual experience, knowledge, economic considerations, attitudes, etc.,10 i.e. the individual GP's consideration of good clinical practice.
Performance. . The current clinical practice, i.e. what the GP actually performs.
Material
The data were obtained from two sources: a survey among GPs and a prospective registration of AOM-related consultations.
The survey. .
All GPs (n = 790) in three Danish counties (Ringkøbing, Funen and North Jutland), which constitute 25% of all Danish GPs, received a mailed questionnaire in September 1994. They were asked to evaluate implicit criteria for diagnosis, treatment of AOM and follow-up examination after AOM. Implicit criteria were assessed by direct questions and case vignettes. Examples of direct questions are: ‘Which equipment do you use when diagnosing AOM?’ (a list of equipment used is shown in Table 1
, item 6) and ‘which symptoms and findings do you demand present to diagnose AOM?’ (a list of symptoms/findings is shown in Table 1, items 2 and 3; ranked reply choices were: ‘Speaking against the diagnosis’, ‘no importance’, ‘important’ and ‘very important’). An example of a case vignette is: ‘Consider a child older than two years, in good general condition, temperature 38.5, continuous earache during three hours and moderately red eardrum’ followed by subsequent questions about treatment etc.
|
Reminders were sent out in October 1994. A total of 602 (76%) questionnaires were returned on 16 November 1994, and 34 of these were not duly completed, resulting in a response rate of 72%.
The prospective registration. .
All 568 GPs participating in the survey were invited to participate in a prospective registration of AOM-related consultations. During a period of 2 weeks in November 1994, the GPs were asked to include all consultations with patients under thirteen consecutively, when the consultation resulted in an AOM diagnosis or the consultation was a follow-up examination after AOM. After each consultation, the GP completed a questionnaire containing topics shown in Table 1
. With regard to symptoms and findings, reply choices included ‘Present’, ‘not present’ and ‘not examined’.
Diagnostic certainty was assessed by a direct question: ‘Are you certain of the diagnosis?’, including the following reply choices: ‘very uncertain’, ‘somewhat uncertain’, ‘certain’ and ‘positive’.
A total of 151 GPs (27%) participated and 368 registrations of contacts were performed. A total of 277 were new cases of AOM.
Data management and statistics. . All questionnaires were reviewed before inclusion. SPSS/PC+ 4.0.1 was used. SOLO was used in the comparison of the participating GPs with the background population. STATISTIX 4.1 was used for estimating odds ratios in the multiple logistic regression models. Five per cent was used as the level of significance.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
The study populations
The response rate was 72% (568 of 790 GPs). Participation was not related to the GP's age, sex or number of patients enrolled in the practice. One hundred and fifty-one GPs participated in the prospective study. As the 151 GPs were drawn from the 568 GPs participating in the survey, detailed analysis of selection bias was possible. This analysis, which was based on 81 questions asked in the survey, revealed no differences between the groups regarding the following: use of equipment for ear-examination, implicit criteria for importance of AOM symptoms, antibiotic prescription pattern (timing of treatment, choice of antibiotic, duration of treatment), frequency, duration, content and need of follow-up examinations after AOM, the pattern of ENT-referral, self-perceived diagnostic difficulties, and, finally, age, sex, otological experience and location of the practice. However, participating GPs showed a tendency to have slightly more equipment in their practices and more knowledge of the differences between AOM and otitis media with effusion (secretory otitis media) than the average.
A total of 368 patients were recruited, and 277 of those were new cases of AOM. The remainder, which will not be discussed further, were either follow-up examinations after AOM or repeated consultations due to persistence of already diagnosed AOM.
Diagnostic criteria and performance
The GPs' implicit criteria as regards importance of symptoms and findings when diagnosing AOM are shown in the left part of Table 2. Earache and fever were the most important symptoms, and a bulging or red eardrum or purulent otorrhea were the most important findings. The current appearance of symptoms and findings in 273 new cases of AOM, together with the frequency of the individual symptoms and examination findings, are shown in the right part of Table 2. The equipment available in the practices for examination of the patients' ears, the GPs' criteria for using diagnostic equipment, and the current equipment used in the 277 new cases of AOM are shown in Table 3. Instruments for removal of cerumen were used in only 3% of new AOM cases.
|
|
Diagnostic certainty
The GPs' opinion about the frequency of difficulties when diagnosing AOM and reasons for diagnostic difficulties are shown in Table 4
. Cerumen was the most frequently encountered problem. In the prospective study, the GPs were certain that the diagnosis was correct in 67% (95% CI 58–76) of new AOM cases as regards children under 2 years. With regard to children aged 2 years or over diagnostic certainty increased to 75% (95%CI 69–81). Of the 32 different variables listed in Table 1 under items 1–4, diagnostic certainty was statistically related to obtaining a good view of the eardrum, and the findings of either purulent otohrrea or a bulging eardrum. Based on 32 variables, a logistic regression model was built to predict the GPs' diagnostic certainty. The model shows that a satisfactory view of the eardrum, purulent otorrhea and a crying child increase the GPs' self-evaluated diagnostic certainty (Table 5).
|
|
A sufficient view of the eardrum was not obtained in 46 cases (12%). In 44 cases, this was caused by cerumen or otorrhoea. Equipment for cleaning the ear canal was used in only four cases (8%, 95% CI 1–18).
In cases in which the GP was uncertain of the diagnosis, he was asked for the reason. A reason was given in 52 of the 72 cases. Grouping of answers revealed that the reason was an insufficient view of the eardrum in 11 cases (21%, 95% CI 10–32). In only one of these cases (9%, 95% CI 0–26), was equipment for cleaning the ear canal used. Thirteen (25%, 95% CI 13–37) cases were associated with diagnostic uncertainty attributed to lack of knowledge, e.g. the GP was uncertain of the interpretation of the findings of a red eardrum with normal mobility and bullae (myringitis bullosa). In the remaining 28 cases (54%, 95% CI 40–67), the reason given was differential diagnostic doubts, especially in relation to an alternative diagnosis of otitis media with effusion (secretory otitis media). Pneumatic otoscopy was used in only one such case (4%, 95% CI 0–10).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
There were no differences in the descriptions of GPs' participating/not participating in the survey. Furthermore, the differences between the 151 GPs participating in the prospective study and the 417 GPs that did not, were negligible. Therefore, selection bias was assumably small.
Diagnostic criteria
In the prospective study, the inclusion of AOM cases was based on the GPs' opinion of whether AOM was present or not; that is, no inclusion criteria was given. This was imperative to the objective of assessing GPs' diagnostic criteria. Furthermore, there is no consensus on explicit criteria for the AOM diagnosis when myringotomy followed by bacteriological examination is not clinically indicated.
In the survey, the GPs were asked their opinion of the importance of different symptoms and findings. In the clinical situation, importance of a symptom correlates to the predictive value of either a positive or negative test. Thus, it is a result of diagnostic sensitivity, specificity and prevalence of the symptom in children with and without AOM. Therefore, when comparing diagnostic criteria with performance, the GP's opinion of importance should be compared with whether or not patients were examined for the symptom, and not with the current prevalence of symptoms in diagnosed cases.
The symptoms earache, fever and reduced hearing, and the findings bulging eardrum, red eardrum and purulent otorrhea were important for the GPs as criteria for the diagnosis AOM (Table 2). Except for the symptom reduced hearing, they were also important for the GPs when considering performance. These symptoms and findings were examined in more than 80% of AOM cases in children aged 2 years or over (Table 2). The discrepancy between criteria and performance as regards reduced hearing is understandable, as a valid examination of hearing is difficult in a crying younger-aged child.
The same factors were important for the GPs when diagnosing AOM in children under 2 years, only with the adjustment that earache, which cannot be reported by infants, in the current performance was assessed by indirect measures such as crying. Compared with the diagnostic criteria reported by the consensus conferences on otitis media,1,2,4 the Danish GPs, in this study, gave more weight to symptoms associated with earache, resulting in stricter diagnostic criteria, which might indicate lower use of antibiotics.
|
Diagnostic certainty |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
As discussed, consensus on explicit criteria for the AOM diagnosis does not exist. Therefore, the GPs' diagnostic certainty was based on the GPs' own opinion of the degree of certainty; thus, interpretation of certainty was with regard to the implicit criteria.
The level of diagnostic uncertainty was substantial, but in agreement with the level found by others.5 When diagnosing AOM, understanding of the natural history of the disease is important. AOM is an inflammation of the mucous membrane in the middle ear with exudation causing the middle ear cavity to be filled with purulent exudate. Therefore, the mobility of the eardrum will be reduced, which is an important sign when excluding both myringitis and external otitis from the differential diagnosis. In secretory otitis media (otitis media with effusion), the middle ear cavity is also filled with fluid; although the eardrum has reduced mobility, it will often be observed as retracted, less often in a neutral and never in a bulging position. Therefore, both the position and the mobility of the eardrum are important in the differential diagnosis of conditions like myringitis, secretory otitis media and otitis externa, which all can cause earache, and conditions like reddening of the eardrum, caused by crying or cleaning of the ear canal from cerumen.2,11 Using an ordinary otoscope when diagnosing AOM is therefore often insufficient.
Only 55% of the GPs considered the finding of an immobile eardrum important, and the mobility of the eardrum was only examined in few of the new AOM cases. More extensive use of techniques evaluating whether middle ear fluid is present or not and the position of the eardrum has the potential to improve diagnostic accuracy considerably. Furthermore, in more than half of the cases when the GP reported diagnostic uncertainty, the reason was differential diagnostic doubts, especially in relation to otitis media with effusion. This uncertainty could also in part be removed by applying techniques evaluating middle ear fluid and the position of the eardrum. In Table 3 it is seen that such equipment was used in only 14% of AOM cases, namely pneumatic otoscope (11%) and tympanometry (3%).
Lildholdt et al.12 showed that introduction of tympanometry in general practice significantly changed the diagnostic pattern of middle ear diseases. The use of tympanometry in general practice requires, however, substantial expenses for the GP and theoretical knowledge on interpretation of the tympanometry curves. The use of pneumatic otoscopy in 11% of the cases should be compared to the fact that 36% of the GPs had access to a pneumatic otoscope (Table 3). The pneumatic otoscope is relatively cheap, and the need for theoretical knowledge is limited, although practice is needed to be a skilled user.11 Furthermore, the pneumatic otoscope identifies reduced mobility of the eardrum with a sensitivity of 85–90% and a specificity of 70–79% in the hands of an experienced ENT specialist.13 These figures are to be compared to those of the tympanometer, i.e. a sensitivity and specificity over 90%.14
As seen in Table 5, the most important factor for diagnostic certainty was, not surprisingly, a satisfactory view of the eardrum. Although insufficient view of the eardrum was almost exclusively caused by cerumen or otorrhea, the ear canal was cleaned in only 8% of the cases.
In conclusion, the diagnostic accuracy could be improved simply by cleaning the ear canal when needed. Furthermore, the diagnostic accuracy could be substantially improved and the level of diagnostic uncertainty reduced by more extensive use of pneumatic otoscopy. Expenses for the GP would be small, and this would compensate for the extra time needed for ear examination.
|
Acknowledgments |
|---|
We would like to thank The Danish National Research Foundation for Primary Care (j. 2-01-143), The Danish Medical Research Council (j. 12-2252-1), and the Magda and Svend Aage Friederics Memorial Endowment. We thank Senior Lecturer Ian Williamson, University of Southampton for valuable comments on the manuscript.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionDiagnostic certaintyReferences |
|---|
1 Prellner K, Bagger-Sjöbeck D, Bohlin G et al. Barn med Öroninflammationer: Konsensuskonferens [Children with middle ear infection]. 23–25 April 1991. Stockholm: SPIR tryck 206, 1991.
2 Karma P, Palva T, Kouvalainen K, et al. Finnish approach to the treatment of acute otitis media: report of the Finnish consensus conference. Ann Otol Rhinol Laryngol 1987; 96 (Suppl. 129): 1–19.[Web of Science][Medline]
3 Lim DJ, Bluestone CD, DeMaria TF, Bakaletz LO. Recent advances in otitis media. Report of the fifth research conference. Ann Otol Rhinol Laryngol 1994; 103 (Suppl. 164): 3–80.
4 Pedersen CB, Olsen J. Danish approach to the treatment of secretory otitis media: report of the Danish consensus conference, Copenhagen, 1987. Ann Otol Rhinol Laryngol 1990; 99 (Suppl. 146): 3–28.
5 Froom J, Culpepper L, Grob P et al. Diagnosis and antibiotic treatment of acute otitis media: report from International Primary Care Network. Br Med J 1990; 300: 582–586.
6 Sjolund A. Daycare Institutions and Children's Development. Lexington, Mass.: Saxon House/Lexington Books, 1973.
7 McCracken GHJ. Emergence of resistant Streptococcus pneumoniae: a problem in pediatrics. Pediatr Infect Dis J 1995; 14: 424–428.[Web of Science][Medline]
8 El Sayed Y, Zakzouk S. Point prevalence of type B tympanogram in Riyadh. Int J Pediatr Otorhinolaryngol 1995; 31: 53–61.[Medline]
9
The Expert Panel. Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Arch Intern Med 1988; 148: 36–69.
10 Donabedian A. The criteria and Standards of Quality. Explorations in Quality Assessment and monitoring, Vol 2. Ann Arbor, Michigan: Health Administration Press, 1982.
11 Schwartz DM, Schwartz RH. Acoustic immittance findings in acute otitis media. Ann Otol Rhinol Laryngol 1980; 89 (Suppl. 68): 211–213.[Web of Science]
12 Lildholdt T, Felding JU, Eriksen EW, Pedersen LV. Diagnostik og behandling af øresygdomme i almen praksis. En kontrolleret undersøgelse af virkningen af indførelse af mellemøretrykmåling (tympanometri). [Diagnosis and treatment of ear dis-eases in general practice. A controlled trial of the effect of the introduction of middle ear measurement (tympanometry)]. Ugeskr Laeger 1991; 153: 3004–3007.[Medline]
13 Stool SE, Berg AO, Berman S et al. Otitis Media with Effusion in Young Children. Clinical practice guidelines, No. 12. AHCPR Publication No. 94-0622. Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, US Department of Health and Human Services, 1994.
14 Fiellau-Nikolajsen M. Tympanometry and secretory otitis media. Observations on diagnosis, epidemiology, treatment, and prevention in prospective cohort studies of three-year-old children (Thesis). Acta Otolaryngol (Stockh) 1983; 394 (Suppl): 1–73.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Blomgren and A. Pitkaranta Is it possible to diagnose acute otitis media accurately in primary health care? Fam. Pract., October 1, 2003; 20(5): 524 - 527. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||