Inter-hospital transfers of patients with systemic lupus erythematosus: characteristics, predictors, and outcomes.

Abstract

OBJECTIVE: To describe the reasons for inter-hospital transfers of patients with systemic lupus erythematosus (SLE), to identify predictors of transfers, and to compare the risk of in-hospital mortality between patients who were transferred and those not transferred. METHODS: Data on acute care hospitalizations of patients with SLE in New York and Pennsylvania in 2000-2002 were obtained from state health planning agencies. We identified inter-hospital transfers from discharge and admission codes, and categorized the major reason for transfer (rehabilitation, procedure, or continued medical care). Patient and hospital characteristics were examined as predictors of transfers. We used a matched cohort design with propensity adjustment to compare in-hospital mortality between patients transferred for continued medical care and those who were not transferred. RESULTS: We identified 533 inter-hospital transfers in 490 patients, 524 of which involved one transfer per hospitalization episode. Of these 524 transfers, 122 (23.3%) were for rehabilitation, 158 (30.1%) were for procedures, and 244 (46.6%) were for continued medical care. Patient characteristics and transfer destinations varied among these groups. Transfers for continued medical care were more common among younger patients, those who were more severely ill, had an emergency or urgent admission, or were hospitalized in a smaller, rural or non-teaching hospital, or in Pennsylvania, and were less common among those at proprietary hospitals. In the matched cohort analysis, the risk of in-hospital mortality was 2.25 times higher (95% confidence interval 1.31, 3.85; p = 0.004) among those transferred compared with those who were not transferred. This risk differed with the experience of the attending physician at the receiving hospital: among patients of physicians who treated 3 or fewer patients with SLE per year, this risk was 2.5 times higher (95% CI 1.42, 4.36; p = 0.002), while among patients of physicians who treated more than 3 patients with SLE per year, this risk was 0.56 times (95% CI 0.06, 5.12; p = 0.62) that of matched controls. CONCLUSION: Patient and transferring hospital characteristics vary with the reason for transfer. Transfers for continued medical care are associated with higher risks of in-hospital mortality, but these risks may differ with the SLE-related experience of the attending physician at the receiving hospital.