A data collection system to audit post-newborn hearing surveillance programme: problems and possibilities.

Background: Guidance documents on post-newborn hearing surveillance and screen (Sutton et al.2006; Bamford et al. 2007) indicated the need for a wider system to identify children with hearing loss after neonatal hearing screening. Recommendations were made for systems to be in place for recording screening activity and audit of the school entry hearing screen to provide information on coverage, referral and yield.

Method: This project has two phases: * development of the data collection system for audit; * assess local service performance.The focus of the work was on data entered into the child health system from children eligible for universal infant and school entry hearing screen. Linking information from a paediatric register of hearing impaired children allowed analysis of birth cohort data related to new diagnoses of sensorineural hearing loss. Available guidelines have not specified gold standards for coverage rates and locally endorsed benchmarks were set at 80% as minimum standards. Analysis of data was carried out on 2003, 2004, 2005, 1998, 1999 and 2000 birth cohorts. The child health system and the paediatric register were the main data sources for the audit exercises. Data extracted were computed for coverage, referral and yield.

Results: Factors and situations contributing to difficulties in establishing a robust system were identified and addressed. Usable information could be obtained to influence current practice.Coverage rates for 2003, 2004 and 2005 cohort were 64.7%, 78.1% and 73.1%. Their respective referral rates were 1.4%, 1.2% and 2.6%. Coverage rates for 1998, 1999 and 2000 cohort were 74.9%,75.6% and 71.4%. Their respective referral rates were 5.2%, 4.2% and 6.6%. The overall yield from universal screens was low.

Conclusion: Our study showed that it was achievable to collect and analyse data on childhood hearing loss in the context of routine surveillance. There were, however, limitations to analysis of data and findings have to be interpreted with this in mind.