Keywords: disparities, health care quality, variable selection
We investigate the relationship between health insurance premium costs and health care quality for members of the Medicare population enrolled in private “Medicare Advantage” plans. The policy concern is that beneficiaries who cannot pay high premiums may receive low quality care. Our quality metrics include patient experience and immunization receipt measures. We implemented mixed-effect regression models to predict quality-of-care measures from plan premiums using nationally representative 2014 Medicare Advantage CAHPS data (n=184,764). Models included random effects for location and fixed effects for beneficiary characteristics. Medicare beneficiaries in plans with moderate versus low premiums (e.g. $110 vs. $25 per month) reported significantly better experiences with quality of care (p < 0.01). We did not detect further improvements in plans with high premiums. Despite these average quality differences, due to large remaining plan-to-plan variation within premium levels, there are many low premium plans with as high or higher quality than plans with moderate and high premium levels.