The goal of a health care system should be to get and keep everyone in optimal health. If people with certain socio-demographic or socio-economic factors such as race, income-level, gender, or religion appear to be disproportionately not meeting this goal, it is worth understanding why that is and what can be done to address this disparity. And there is increasing controversy over the failure to include adjustments for such factors in pay-for-performance and other quality-oriented programs. At its root the issue is the extent to which a health care provider or health plan can realistically impact those factors and the behaviors that may be related to them. An article in the Journal of the American Medical Association examines the association of certain of these factors with outcomes for heart attack, heart failure and pneumonia hospitalizations. (JAMA Article) The authors were trying to ascertain if race or income level was tied to outcomes within specific hospitals, of whether, any variation, if it existed, was more systemic. The primary outcomes were mortality and readmissions within the Medicare population. A large number of hospitals could not be included in the analysis because they lacked sufficient racial and income diversity in their patient populations, but these tended to be smaller, rural hospitals, and a large percent of hospital discharges for the relevant diagnoses was included.
For all three conditions, African-American patients had slightly lower mortality than did white patients treated in the same hospital, but had slightly higher readmission rates. In regard to neighborhood income levels there were very small, statistically insignificant differences between patients from lower-income neighborhoods compared to higher-income ones, when treated at the same hospital, although lower-income patients generally had those very slightly worse outcomes in regard to mortality and readmissions. These results suggest that there is no meaningful difference based on race or income in how patients fare when hospitalized for these conditions. In addition, performance on these measures did not appear to vary in any significant manner between hospitals depending on the proportion of black or poorer patients that were treated. Readmission rates were again very slightly, but not statistically significantly, worse at hospitals with a higher proportion of African-American or poor patients, but mortality rates were indistinguishable. While there are differential outcomes by race or income nationally, these are not due to hospitals treating these patients differently or have different outcomes for them. They instead appear to be systemic and I would respectively suggest are clearly due to differences in responsible health and health care seeking behavior. Helping people to improve those behaviors, which most often occur in all aspects of their lives, is the key to eliminating disparities in outcomes.