• CONTEXT
    • Acute otitis media (AOM) is one of the most common problems in pediatrics. An accurate diagnosis of AOM can guide proper treatment and follow-up.
  • OBJECTIVE
    • To systematically review the literature regarding precision and accuracy of history taking and physical examination in diagnosing AOM in children.
  • DATA SOURCES
    • We searched MEDLINE for English-language articles published from 1966 through May 2002. Bibliographies of retrieved articles and textbooks were also searched.
  • STUDY SELECTION
    • We located studies with original data on the precision or accuracy of history or physical examination for AOM in children. Of 397 references initially identified, 6 met inclusion criteria for analysis.
  • DATA EXTRACTION
    • Two authors independently reviewed and abstracted data to calculate likelihood ratios (LRs) for symptoms and signs.
  • DATA SYNTHESIS
    • Four studies of symptoms used clinical diagnosis as the criterion standard and were limited by incorporation bias. Ear pain is the most useful symptom (positive LRs, 3.0-7.3); fever, upper respiratory tract symptoms, and irritability are less useful. One study of clinical signs used tympanocentesis as the criterion standard, and we adjusted the results to correct for verification bias. A cloudy (adjusted LR, 34; 95% confidence interval [CI], 28-42), bulging (adjusted LR, 51; 95% CI, 36-73), or distinctly immobile (adjusted LR, 31; 95% CI, 26-37) tympanic membrane on pneumatic otoscopy are the most useful signs for detecting AOM. A distinctly red tympanic membrane is also helpful (adjusted LR, 8.4; 95% CI, 6.7-11) whereas a normal color makes AOM much less likely (adjusted LR, 0.2; 95% CI, 0.19-0.21).
  • CONCLUSIONS
    • Although many of the studies included in this analysis are limited by bias, a cloudy, bulging, or clearly immobile tympanic membrane is most helpful for detecting AOM. The degree of erythema may also be useful since a normal color makes otitis media unlikely whereas a distinctly red tympanic membrane increases the likelihood significantly.