The Fluharty Preschool Speech and Language Screening Test: a population-based validation study using sample-independent decision rules.

Two cohorts of 4- and 5-year-old children (N = 700) were screened with the Fluharty Preschool Speech and Language Screening Test. Two stratified samples (n = 51 cohort 1; n = 147 cohort 2), based on speech/language screening results, were administered criterion tests for articulation (AAPS-R or Templin-Darley) and language (TOLD or TALC-R). Clinical validation indices for combined speech or language outcome in the two cohorts were as follows: sensitivity, .43 and .31 respectively; specificity, .82 and .93; predictive value, .43 and .54; overreferral, 14% and 5%; underreferral, 14% in both cohorts, and percent agreement, 72% and 80%. The measure of sensitivities for language outcomes was lower than the above, whereas sensitivities for articulation was higher. These results suggest that the Fluharty is too insensitive to be relied on for screening programs aimed at identifying preschool children with language disorders, although it appears to have promise for the identification of children with articulation impairments.