Abstract
Rapidly rising health care costs continue to be a significant concern in the United States. High cost-sharing strategies thus have been widely used to address rising health care costs. Since high cost-sharing policies can reduce needed care as well as unneeded care use, it raises the concern whether these policies for physician care are a good strategy for controlling costs among chronically ill patients, especially whether utilization and costs in inpatient care will increase in response. This study examined whether high cost sharing in physician care affects inpatient care utilization and costs differently between individuals with and without chronic conditions. Findings from this study will contribute to the insurance benefit design that can control care utilization and save costs of chronically ill individuals. Prior studies suffered from gaps that limit both internal validity and external validity of their findings. This study has its unique contributions by filling these gaps jointly. The study used data from the 2007 Medical Expenditure Panel Survey, a nationally representative sample, with a cross-sectional study design. Instrumental variable technique was used to address the endogeneity between health care utilization and cost-sharing levels. We used negative binomial regression to analyze the count data and generalized linear models for costs data. To account for national survey sampling design, weight and variance were adjusted. The study compared the effects of high cost-sharing policies on inpatient care utilization and costs between individuals with and without chronic conditions to answer the research question. The final study sample consisted of 4523 individuals; among them, 752 had hospitalizations. The multivariate analysis demonstrated consistent patterns. Compared with low cost-sharing policies, high cost-sharing policies for physician care were not associated with a greater increase in inpatient care utilization (P = .86 for chronically ill people and P = .67 for healthy people, respectively) and costs (P = .38 for chronically ill people and P = .68 for healthy people, respectively). The sensitivity analysis with a 10% cost-sharing level also generated consistent insignificant results for both chronically ill and healthy groups. Relative to nonchronically ill individuals, chronically ill individuals may increase their utilization and expenditures of inpatient care to a similar extent in response to increased physician care cost sharing. This may be due to cost pressure from inpatient care and short observation window. Although this study did not find evidence that high cost-sharing policies for physician care increase inpatient care differently for individuals with and without chronic conditions, interpretation of this finding should be cautious. It is possible that in the long run, these sick people would demonstrate substantial demands for medical care and there could be a total cost increase for health plans ultimately. Health plans need to be cautious of policies for chronically ill enrollees.