CONTEXT

Each year approximately 5000 infants are born in the United States with moderate to profound, bilateral sensorineural hearing loss (SNHL). Universal newborn hearing screening (UNHS) has been proposed as a means to speed diagnosis and treatment, and thereby improve language outcomes in these children.

OBJECTIVE

To identify strengths, weaknesses, and gaps in the evidence supporting UNHS and to compare the additional benefits and harms of UNHS with those of selective screening of high-risk newborns.

DATA SOURCES

A keyword search of MEDLINE, CINAHL, and PsycINFO databases for relevant papers published from 1994 to August 2001, using terms for hearing disorders, infant or newborn, screening, and relevant treatments. We contacted experts and reviewed reference lists to identify additional articles, including those published before 1994.

STUDY SELECTION

We included controlled and observational studies of (1) the accuracy, yield, or harms of screening using otoacoustic emissions (OAEs), auditory brainstem response (ABR), or both in the general newborn population or (2) the effects of screening or of early identification and treatment on language outcomes. Nineteen articles, including 1 controlled trial, met these inclusion criteria.

DATA EXTRACTION

Data on population, test performance, outcomes, and methodological quality were extracted using prespecified criteria developed by the US Preventive Services Task Force. We queried authors when information needed to assess study quality was missing.

DATA SYNTHESIS

Good quality studies show from 2041 to 2794 low-risk, and 86 to 208 high-risk, newborns were screened to find 1 case of moderate to profound SNHL. The best estimate of positive predictive value is 6.7%. Six percent to 15% of infants who fail the screening tests are subsequently diagnosed with bilateral SNHL. In a trial of UNHS versus clinical screening at 8 months of age, UNHS increased the proportion of infants with moderate to severe hearing loss diagnosed by 10 months of age (57% vs 14%), but did not reduce the rate of diagnosis after 18 months of age. No good-quality controlled study has compared UNHS to selective screening of high-risk newborns. In fair- to poor-quality cohort studies, intervention before 6 months of age was associated with improved language and communication skills by 2 to 5 years of age. These studies had unclear criteria for selecting subjects, and none compared an inception cohort of low-risk newborns identified by screening to those identified in usual care, making it impossible to exclude selection bias as an explanation for the results. In a mathematical model based on the literature review, we estimated that extending screening to low-risk infants would detect 1 additional case before 10 months for every 1441 low-risk infants screened, and result in treatment before 10 months of 1 additional case for every 2401 low-risk infants screened. With UNHS, 254 newborns would be referred for audiological evaluation because of false-positive second-stage screening test results, versus 48 for selective screening.

CONCLUSIONS

Modern screening tests for hearing impairment can improve identification of newborns with SNHL, but the efficacy of UNHS to improve long-term language outcomes remains uncertain.