Speech, admittance and tone tests in school screening. Reconciling economics with pathology and disability perspectives.

In theory a criterion can be set on any test score to 'pass' or 'fail' a desired percentage of any population; but in practice, particular measures used in screening can have both intrinsically appropriate criteria and zones of great uncertainty in relation to the condition being screened. For example, the criteria for possible abnormality on oto-admittance measures in single-test screening for middle ear disease are intrinsically rather too sensitive, as potentially pathological signs are displayed by a high proportion of the population. In contrast, criteria on pure tone measures are intrinsically insensitive. One solution to such dilemmas and to other deficiencies of screening lies in a two-test approach. In two investigations we have examined these accepted screening tests in relation to a proposal to introduce also a disability measure as a means of broadening the information available from school screening systems and reducing over-referral from oto-admittance screening. A minimal-pair monosyllable 2-alternative forced-choice speech-in-noise test with a picture-pointing identification response was developed and refined. The results obtained usefully complement existing screening methods and we examine ways of combining them with either pure tone threshold of oto-admittance measures. A sequential set of criteria for passing, referral or continuation of testing allows the test to be applied in short form for those children failing a pre-screen at a stringent level. In this way a reduction of approximately 10% of the total screened sample is obtained in the number of children needing referral or re-testing; the total testing time need not increase substantially; and more precise disability information is obtained on the marginal cases, where it is needed. Multiple regression enables a severity index (on the same scale as the widely appreciated pure-tone threshold scale) to be derived from speech and oto-admittance measures jointly. An adequate basis of referral for medical assessment and treatment of pathology would still require admittance methods at some stage, whatever the drawbacks may be in their sole use as a screen. The two-test sequential approach reconciles the pathology perspective and the disability perspective.