BACKGROUND/RATIONALE:
While the universal access to care provided by the VA to poor veterans and other system design features are reasons to believe that racial disparities are smaller within the VA than in non-VA settings, this study will provide explicit evidence about the degree to which that is true.
OBJECTIVE(S):
To compare racial disparities in 30-day mortality among patients treated in VA hospitals for six inpatient medical conditions (acute myocardial infarction, stroke, hip fracture, gastrointestinal hemorrhage, pneumonia, congestive heart failure) with disparities among similar patients treated in non-VA hospitals to determine how well the VA system does in reducing disparities.
METHODS:
This observational study of outcomes uses secondary data to estimate risk-adjusted mortality for white and black patients admitted between October 1995 and September 2002 for the six study conditions. Data sources include the VA inpatient discharge dataset (Patient Treatment File [PTF]); Medicare (MEDPAR) discharge dataset; state hospital discharge data from Pennsylvania and California, and the Healthcare Coast and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS). Comparisons will be made nationally for all male patients 65 and older, while comparisons will be limited to Pennsylvania, California, and HCUP for male patients under age 65. Identify hospitals with high and low racial disparities and examine the hospital level factors associated with different degrees of disparities.
Estimates of racial disparities will be based on multi-level logistic regression models of 30-day mortality using covariate adjustment of comorbidity, age, year of discharge, hospital where admitted, and SES. Analysis of hospital and VISN-level factors will produce race-specific odds ratios for each hospital based on multi-level logistic models and will identify hospital-level predictors that account for substantial hospital- and/or VISN-level variation in the race-specific odds ratios.
FINDINGS/RESULTS:
Analysis of the VA data indicated different patterns among the under and over 65 populations. In models adjusted for age, year, comorbidities, means test eligibility and area-level SES, and hospital random effects, blacks had consistently significantly lower odds of mortality than whites in the over 65 population. However, among the under 65 population this pattern was not observed; CHF was the only condition for which blacks under 65 had significantly lower odds of mortality than whites.
Comparative analysis utilizing the OSHPD and PHC4 data demonstrates that factors associated with better 30-day mortality for blacks in the over-65 population may not be unique to VA. For the under-65 population the odds of mortality are similar for black and white patients in both the VA and non-VA settings.
Analyses for the Medicare population indicate within the Medicare population, 30-day mortality rates are lower for Blacks than for Whites for hospital admissions for common high-mortality conditions. However, as time from admission increases, mortality for Blacks relative to Whites increases for 5 of our 6 study conditions. Consequently, 2-year mortality rates are higher for Blacks for these same hospital admissions.
Analysis of the characteristics and performance of VA hospitals that care for blacks versus others indicate that hospital care for veterans is concentrated. Just 9/160 hospitals care for 25% of black veterans, and 42/160 care for 75% of black veterans. VA hospitals that disproportionately care for blacks are: larger, in the South, urban, are teaching hospitals and have greater technological capabilities. Despite this concentration of care, analysis found a lack of significant variability in black-white differences in mortality across sites.
IMPACT:
Our findings demonstrate that blacks have had better outcomes than whites for a number of high mortality conditions in both VA and non-VA hospitals and that relative to white patients, black patients within VA have better long term outcomes. Within the VHA, care for blacks is concentrated and thus QI efforts targeted on those hospitals that disproportionately care for blacks can have a substantial impact on care for minorities.
PUBLICATIONS:
Journal Articles
- Polsky D, Lave J, Klusaritz H, Jha A, Pauly MV, Cen L, Xie H, Stone R, Chen Z, Volpp K. Is lower 30-day mortality posthospital admission among blacks unique to the Veterans Affairs health care system? Medical Care. 2007; 45(11): 1083-9.
- Volpp KG, Stone R, Lave JR, Jha AK, Pauly M, Klusaritz H, Chen H, Cen L, Brucker N, Polsky D. Is thirty-day hospital mortality really lower for black veterans compared with white veterans? Health Services Research. 2007; 42(4): 1613-31.
- Gurmankin AD, Polsky D, Volpp KG. Accounting for apparent "reverse" racial disparities in Department of Veterans Affairs (VA)-based medical care: influence of out-of-VA care. American Journal of Public Health. 2004; 94(12): 2076-8.
