CLINICAL APPLICATIONS MEDICAL DECISION MAKING/MAY–JUNE 2007 233

Objective. The objective of this retrospective study was to assess the evidence for racial differences in discharge dispo-sition among patients hospitalized for stroke. Data. Hospital discharge data from the Maryland Health Services Cost Review Commission were used in the analysis. The data cov-ered the period from January 2000 to September 2003. Study design. Discharge-disposition categories were ordered such that higher numbers corresponded to less desirable out-comes: 1 = discharge to home; 2 = discharge to any medical care facility; 3 = death. We analyzed the influence of black race on the discharge disposition by estimating a partial pro-portional odds logit regression model that included demo-graphic and clinical covariates. Data extraction. The study inclusion criteria were 1) stroke (ICD9 431–434; 436–438) as a primary admission diagnosis and 2) patient race identified as black or white. Patients discharged against medical advice were excluded. The sample contained 51,564 stroke hospitalizations. Principal findings. Based on the relative odds ratios (OR; 95 % confidence interval [CI]), black males were more likely to be discharged to higher ranked (i.e., less desirable) discharge categories (OR = 1.66; CI 1.55–1.77) compared to white males. Black females were more likely to die (OR = 1.14; CI 1.02–1.28) and more likely either to die or to be discharged to medical care (OR = 1.38; CI 1.24–1.54) compared to white males. Conclusions. Blacks are at greater mortality risk following stroke hospitalizations and face less desirable discharge dispositions if they survive. These results are consistent with prior reports of lower sur-vival rates among blacks and are robust to adjustments for various confounding factors. Key words: patient discharge; African Americans; cerebrovascular stroke; in-hospital mor