Spending and readmission for multiple chronic conditions in vulnerable population: Results from county level panel data analysis

Multiple chronic conditions (MCC) have become a major public health issue because it significantly increases the risk of poor outcomes including unnecessary hospitalization, and financial burden. A significant gap exists in our understanding of the health care delivery to MCC afflicted individuals, especially in regions that has higher minority population and are deprived on socioeconomic indicators. In this study, I evaluate the variation in healthcare spending and readmission rates with a specific focus on socioeconomic indicators of counties across the United States.

The study sample included annual estimates on 3074 counties in US for the period 2008–2012. The data on prevalence rates, related standardized healthcare spending, readmission rates, and ED visits for different MCC levels [0–1; 2–3; 4–5; and 6plus conditions] were obtained from the CMS website. CMS reported this data based on 100% Medicare claims, and MCC categories were counted from a set of 17 chronic conditions. The socioeconomic indicators were obtained from US Census Bureau, National Center for Health Statistics, and RTI spatial impact factor database. These indicators include urbanity index, percentage Hispanic and non–Hispanic Black population, poverty rates, percentage rural population, unemployment rate, percent population with less than high school, and college/higher degrees, and percentage population with poor/no English speaking capability.

The data was analyzed in two steps. First, temporal and cross–sectional pattern for spending, readmission and ED visits were examined. Next, using multilevel mixed effects model the temporal variation for each MCC group was analyzed adjusting for state fixed effects. We also adjusted for nonlinear growth over time by including year and year^2 as covariates, and accounted for autocorrelation effects.

My results showed average standardized spending, readmission rates, and ED visits rose exponentially with increase in MCC across years. Within each MCC groups these measures were lower in counties with higher prevalence of focal MCC compared to low prevalence counties. Furthermore, counties with higher Hispanic population  (above 20%) had lower spending and ED visits compared to counties with lesser Hispanic population. However, counties with higher non–Hispanic Black population had higher spending, readmission, and ED visits compared to counties with lesser non–Hispanic Black population. Poverty, urbanity, rural population, and lower education were associated with higher levels of spending, readmission, and ED visits.

These results elucidate the disparities in care delivery to MCC population, especially in counties that has higher minority population, and are deprived on several socioeconomic indicators.