OBJECTIVE Population ascertainment of children having bilateral moderate to profound hearing impairment was undertaken to find out how many had passed (false negatives) and how many had failed (true negatives) a neonatal screening test based on transient-evoked otoacoustic emissions (TEOAE). METHODS Neonatal screening using a purpose-built TEOAE instrument was undertaken in neonates, at eight hospitals in districts distributed around the United Kingdom starting in 1988. Screening was targeted on neonates at risk of hearing impairment. A total of 7500 babies had been tested by the end of 1995. Searching of audiological records in the districts completed at the end of 1995 ascertained 218 children born between January 1988 and December 1993 who had hearing threshold levels in both ears of 50 dB or more, averaged over the speech frequencies 0.5, 1, 2, and 4 kHz. Of those, 47 had completed the neonatal TEOAE screening test. Retrospective examination of their TEOAE records indicated whether they had passed or failed the screening test. RESULTS Eleven of the 47 had passed the screening test, although two of those had documented acquired hearing impairment occurring after screening. Disregarding those two cases leaves nine false negatives out of 45, giving a sensitivity estimate of 80% (36 divided by 45). In two of the nine cases, there was documented evidence of progression, one of whom had a family history of progressive hearing loss. CONCLUSIONS Targeted neonatal hearing screening programs based on TEOAE can expect to identify hearing impairment in approximately 80% of babies screened. The cause of false negatives is a matter for conjecture. There may be several reasons: the test may give an incorrect result, there may be a later acquired hearing impairment, or there may be a progressive hearing loss of unknown origin. Passing a neonatal screening test is not a valid reason to forego further surveillance, or to disregard parental suspicion of hearing impairment.