Case
Abstract
Together with co-authors from institutions including Johns Hopkins, American Association of Retired Persons, the University of California–San Francisco, the Hearing Loss Association of America, and OptumLabs, we conducted a research study which focused on development of comorbidities, utilization of health care services, and health service expenditure associated with untreated hearing loss. This project resulted in two articles published in JAMA Otolaryngology—Head & Neck Surgery. The study used administrative claims and other “real-world data” to compare patients of age 50 years and older diagnosed with (but not treated for) hearing loss to matched controls without hearing loss over 2-, 5-, and 10-year follow-up periods. As is necessary in a clinically focused journal, the publications emphasized the outcomes of interest (incident comorbid conditions, health care costs, and health care service utilization) and devoted minimal space to the complex methodological strategies required to execute population health studies of this type. These challenges center on issues of measurement, selection bias, violations of normality assumptions, and the use of real-world data to address questions related to health outcomes. In this case study, we review the methodological challenges we faced and the strategies we employed to mitigate their impact.