Previously, it has been estimated that the uninsured adults have a 25 percent higher overall risk of mortality compared to those with private insurance, resulting in 18,000 additional deaths annually. The evidence on the relationship between the presence of a usual source of care and health outcomes has been mixed, associated with short-term gains in health status, but not lower mortality.

This study analyzed data from the Health and Ways of Living Survey in the Human Population Laboratory (HPL), a longitudinal data set, collected among the non-institutionalized adult population in Alameda County, California between 1965 and 1999. A Cox proportional hazards model was estimated, allowing values of independent variables to change over time. Individual growth models and difference in differences models were estimated using physical and mental Health Related Quality of Life (HRQOL) scores as outcome measures. Treatment effects and bivariate probit models also were estimated.

The relationship between health insurance, usual source of care and health outcomes varied depending on outcome measure and analytic methods used. In the Cox models that included other determinants that changed over time, health insurance was not associated with reduced mortality hazard for this population. Health insurance was associated with 34 percent lower mortality hazard among women overall and 28 percent lower mortality hazard for women who reported a chronic condition at baseline. No differences were observed for men or among overall population who reported chronic conditions or hypertension at baseline. The presence of a usual source of care was not associated with differences in mortality. Being uninsured was associated with greater decline in physical health status over the study period. Not reporting a usual source of care was associated higher physical health scores and a greater increase in physical HRQOL scores compared to those who reported one. After correcting for endogeneity, usual source of care was found to be associated with better physical HRQOL scores at each wave (0.92, z < .001) and greater improvement between the first and last waves (1.11, z < .001). In instrumental variable analyses, health insurance was not associated with higher physical HRQOL scores or improvement between baseline and 1999.