No current consensus exists on the causal effect of gaining or losing health insurance on health care utilization and health outcomes.
To systemically search and review available evidence of estimated causal effects of health insurance on health care utilization and/or health outcomes among nonelderly adults in the United States.
A systematic search of 3 electronic databases (PubMed, JSTOR, EconLit) was performed. To be included in the review, studies had to have a publication date after 1991; a population of nonelderly adults; analyses comparing an uninsured group to an appropriate control group; and 1 of 3 study designs that account for potential reverse causality and provide estimates of causal effects (longitudinal cohort, instrumental variable analysis, or quasi-experimental design).
A total of 9701 studies, including duplicates, were primarily screened. Fourteen studies fulfilled the criteria to be included in this review—4 longitudinal cohort studies using standard regression or fixed effects analysis, 5 longitudinal cohort studies using instrumental variable regression analysis, and 5 quasi-experimental studies.
Results of our review of empirical studies that estimate causal relationships between health insurance and health care utilization and/or health outcomes consistently show that health insurance increases utilization and improves health. Specifically, health insurance had substantial effects on the use of physician services, preventive services, self-reported health status, and mortality conditional on injury and disease. These results both confirm and contradict comparable results from the RAND Health Insurance Experiment, the gold standard on relationships between health insurance, utilization, and health.
From the *University of Washington School of Medicine, †University of Washington School of Pharmacy, and ‡University of Washington School of Public Health and Community Medicine, University of Washington Health Services, Seattle, Washington.
Supported by Agency for Healthcare Research and Quality grant # T32-HS013853-01 (to J.F.B.).
Reprints: Joseph D. Freeman, University of Washington School of Medicine, 4625 92nd Ave NE, Bellevue, WA 98004. E-mail: email@example.com.