Variations in length of stay and outcomes for six medical and surgical conditions in Massachusetts and California

Objectives. —To determine the extent to which interinstitutional variations in length of stay are explained by differences in patient characteristics and to determine whether patients in hospitals with shorter lengths of stay had worse outcomes. Design. —We reviewed patients' medical records and surveyed patients between 3 and 12 months after hospital discharge using a questionnaire. Setting. —Six teaching hospitals in California and Massachusetts. Patients. —A cohort of 2484 selected patients who had been hospitalized for acute myocardial infarction or to rule out acute myocardial infarction, coronary artery bypass graft surgery, total hip replacement, cholecystectomy, or transurethral prostatectomy. Between 73% and 84% of the patients with each condition completed a follow-up questionnaire. Outcome Measures. —In-hospital complications, deaths, length of stay, functional status after hospital discharge, readmission, and patient satisfaction with hospital care were analyzed. Results. —Significant interinstitutional differences in length of stay were noted for all conditions except rule-out acute myocardial infarction. Statistical adjustment for case-mix differences accounted for most of the interinstitutional differences in length of stay for total hip replacement but explained little of the differences in the other conditions. When we controlled statistically for other predictors, length of stay did not have a significant impact on deaths, functional status after hospital discharge, the probability of readmission, or patient satisfaction with hospital care. Conclusion. —More research is needed to determine the medical practices that are related to variations in lengths of stay. Routinely available outcome data may help preserve quality in the face of efforts to decrease costs by effecting more standardized practices of care. (JAMA. 1991;266:73-79)