Time trends analysis of hearing loss: an alternative approach to evaluating hearing loss prevention programs

Two general approaches for evaluating the performance of hearing loss prevention programs (HLPPs) are described in the literature: (1) a comparison of the rate of hearing loss in an HLPP with a reference population and (2) a comparison of audiometric variability or annual incidence of hearing loss with an established set of criteria that rank HLPP performance on a graded scale. This article discusses a third method, time trends analysis, which assesses patterns in hearing loss over time. Patterns may reflect program improvement (decreased hearing loss incidence over time), deterioration (a pattern of increased incidence), or stasis (unchanged incidence). To demonstrate this method, a time trends analysis was conducted on a population of 44,547 industrial workers. Subjects were divided into 11 subgroups based on year of enrollment in the HLPP (1980-1990) and followed retrospectively for 3 years to determine the incidence of hearing loss. Hazard ratios (HRs) were estimated for each subgroup by gender using the Cox Proportional Hazards model and adjusting for age, race, and hearing threshold at enrollment in the HLPP. For women, plots of adjusted HRs against enrollment years produced a statistically significant (p < 0.05) quadratic trend of an initial increase in hearing loss, followed by decreasing incidence over time. For men, there was a statistically significant linear trend of decreasing hearing loss over time. The downward trend, particularly during the late 1980s, indicates improved HLPP performance during the latter portion of the decade. Time trends analysis can be a valuable tool for assessing HLPP performance for those with access to follow-up data and ability to work with statistical models.