A father brought in his 2 year old boy because of a “cold” and also concern about his hearing and speech development. He thought the boy sometimes heard well and other days appeared deaf. Clinical examination showed dull tympanic membranes with retraction on the left. I wondered how accurate the clinics microtympanometer was for the diagnosis of otitis media with effusion and eustrachian tube dysfunction and formulated the question, what is the accuracy of microtympanometry for the diagnosis of hearing loss from a middle ear effusion in young children?
Searching terms and evidence source: (tympanomet* OR impedance) AND (otitis media OR middle ear) AND audiomet* (MEDLINE)
Read the article and decide:
 Is the evidence from this article valid?
 If valid, is this evidence important?
 If valid and important, can you apply this evidence in caring for your patient?
Completed Diagnosis Worksheet for EvidenceBased General Practice
Citation
Holty I, Forster DP. Evaluation of pure tone audiometry and impedance screening in infant schoolchildren. J Epidemiol Community Health 1992; 46: 2125.
Are the results of this diagnostic study valid?
 Was there an independent, blind comparison with a reference (“gold”) standard of diagnosis?
 Yes – the audiometry and tympanometry were done by different examiners 5 days apart and without knowledge of the previous result.
 Was the diagnostic test evaluated in an appropriate spectrum of patients (like those in whom it would be used in practice)?
 Schoolaged children were screened.
 Was the reference standard applied regardless of the diagnostic test result?
 Yes – all children were supposed to undergo both tests, and 94.1% did.
Are the valid results of this diagnostic study important?
Your calculations:
Target Disorder (abnormal audiometry)  Totals  

Present  Absent  
Diagnostic Test Result (tympanometry) 
Positive (type B or C)  99 a 
92 b 
191 a + b 
Negative (type A)  73 c 
310 d 
383 c + d 

Totals  a + c 172 
b + d 402 
574 
begin{align} mathit{Sensitivity} &= a/(a+c) \
&= 99/172 \
&= 58% end{align}
begin{align} mathit{Specificity} &= d/(b+d) \
&= 310/402 \
&= 77% end{align}
begin{align} text{Likelihood Ratio for a positive test result ($LR+$)} &= mathit{sens}/(1mathit{spec}) \
&= 58%/23% \
&= 2.5 end{align}
begin{align} text{Likelihood Ratio for a negative test result ($LR$)} &= (1mathit{sens})/mathit{spec} \
&= 42%/77% \
&= 0.54 end{align}
begin{align} text{Positive Predictive Value} &= a/(a+b) \
&= 99/191 \
&= 52% end{align}
begin{align} text{Negative Predictive Value} &= d/(c+d) \
&= 310/383 \
&= 77% end{align}
begin{align} text{Pretest Probability ($prevalence$)} &= (a+c)/(a+b+c+d) \
&= 172/574 \
&= 81% end{align}
begin{align} mathit{Pretestodds} &= mathit{prevalence}/(1mathit{prevalence}) \
&= 30%/70% \
&= 0.43 end{align}
$$ text{Posttest odds} = text{Pretest odds} times text{Likelihood Ratio} $$
$$ text{Posttest Probability} = text{Posttest odds}/(text{Posttest odds} + 1) $$
Can you apply this valid, important evidence about a diagnostic test in caring for your patient?
 Is the diagnostic test available, affordable, accurate, and precise in your setting?
 Yes. Many practices, including ours, have one of these simple cheap instruments.
 Can you generate a clinically sensible estimate of your patient’s pretest probability (from practice data, from personal experience, from the report itself, or from clinical speculation)
 Parental concern is a poor predictor of hearing problems (Rosenfeld, Arch Otolaryngol Head Neck Surg 1998 Sep;124(9):98992). I would adjust the prevalence slightly to a pretest value of 40%.
 Will the resulting posttest probabilities affect your management and help your patient? (Could it move you across a testtreatment threshold?; Would your patient be a willing partner in carrying it out?)
 A positive test would predict about a 63% chance of an abnormal audiogram (and warrant an audiogram); a negative test a 27% chance (and warrant a repeat test in a several weeks).
 Would the consequences of the test help your patient?
 Yes – a recent trial (Maw, Lancet 1999 353: 9603) of delayed versus immediate surgery for OME showed a benefit in language development but that the delayed group also later caught up.
Additional Notes
While the “disease” of interest is otitis media with effusion, I have taken audiometry as the gold standard since it is really the hearing impairment that is important to the child, not the presence of some middle ear fluid.
Tympanometry is a moderate predictor of audiometric hearing loss in school children
Clinical Bottom Line
Tympanometry is moderately helpful in the assessment of possible childhood hearing problems, but cannot rule out problems (sensitivity 58%).
Citation
Holty I, Forster DP. Evaluation of pure tone audiometry and impedance screening in infant schoolchildren. J Epidemiol Community Health 1992; 46: 2125.
Clinical Question
What is the accuracy of microtympanometry for the diagnosis of hearing loss from a middle ear effusion in young children?
Search Terms
(tympanomet* OR impedance) AND (otitis media OR middle ear) AND audiomet* (MEDLINE)
The Study
The study attempted to screen 610 school aged children with both audiometry and tympanometry done 5 days apart (half in each sequence) by different examiners unaware of the previous result. 94.1% had both tests.
The Evidence
Target Disorder (abnormal audiometry)  Totals  

Present  Absent  
Diagnostic Test Result (tympanometry) 
Positive (type B or C) 
99 a 
92 b 
191 a + b 
Negative (type A) 
73 c 
310 d 
383 c + d 

Totals  a + c 172 
b + d 402 
574 
begin{align} mathit{Sensitivity} &= a/(a+c) \
&= 99/172 \
&= 58% end{align}
begin{align} mathit{Specificity} &= d/(b+d) \
&= 310/402 \
&= 77% end{align}
begin{align} text{Likelihood Ratio for a positive test result ($LR+$)} &= mathit{sens}/(1mathit{spec}) \
&= 58%/23% \
&= 2.5 end{align}
begin{align} text{Likelihood Ratio for a negative test result ($LR$)} &= (1mathit{sens})/mathit{spec} \
&= 42%/77% \
&= 0.54 end{align}
Comparison of the 94.1% of children who completed both tympanometry and audiometry showed a sensitivity (for hearing loss) of 58% and a specificity of 77%, suggesting only a modest accuracy for tympanometry. However, the same study showed that audiometry had a testretest repeatability of 67%. This inaccuracy in the reference standard will lead to an underestimate of the accuracy of tympanometry.
Comments
The results for Type B and Type C tympanometries have been combined here as “abnormal”, but may have differential accuracy.
Appraised By
Paul Glasziou