Pure Effect of Macroeconomic Fluctuations on Medical Care Utilization

That macroeconomic conditions may affect an individual's decision to seek medical care has recently become an area of considerable research. Because of the complex, confounding relationships between health status, insurance, employment, and medical utilization, much of this research has centered on the evaluation of specific transmission channels, notably changes in employment and insurance status, through which macroeconomic forces may influence individual decisions. However, the pure macroeconomic effect itself is a valuable policy question.

For example, individuals who are both employed and insured may elect to seek less care during downturns if they are concerned about potential, future staff reductions at their current employer in an attempt to demonstrate their value to the firm – perhaps in hope that more intensive care seekers, and presumably more frequently absent, are those let go. Alternatively, such individuals may seek more care under these circumstances due to fears that they may soon be uninsured. Policy changes that seek to increase access to care among the unemployed or uninsured may create incentives that could potentially alter the utilization decisions among those currently employed and insured.

We evaluate the pure effect of the business cycle by examining the choices of Medical Expenditure Panel Survey (MEPS) respondents to seek a variety of medical care services while controlling for important factors such as employment, insurance and self-reported health status. We do so by exploiting the longitudinal nature of MEPS using a fixed-effects Poisson model. Employing the state identifiers associated with each respondent, we are able to use state-level economic measures allowing us to increase the power of this panel. Additionally, we modify the traditional robust asymptotic covariance matrix to account for the multi-level survey design and compare the effects of this adjustment to the unmodified standard errors.

State unemployment rates are statistically significantly (p<.01) associated with all of the medical care services considered in the study when using traditional robust standard errors. After adjusting the standard errors for the multi-level design, only 25% of the analyses obtained similarly significant results. Within the significant results, hospital outpatient visits and emergency room visits increased 20% and 10%, respectively, for respondents who are both employed and insured throughout the sample for each one percentage point increase in the state unemployment rate.

These results suggest that reforms designed to increase access may provide additional benefits to those with access by mitigating the effects of employment risk on their decisions to seek care.