In recent years, escalating health care costs have prompted legislators, third-party payers, and the public at large to demand health care reform and focus much of their attention on the largest component of health care expenditures, hospital care. One strategy used to stem inefficiencies in hospitals and to ensure their financial viability is to redesign/restructure the work or care delivery system. Nursing case management is one popular approach. The specific aims of this study were to: (1) describe a nursing case management care delivery system for medical-surgical patients, and (2) empirically test its effect on hospital output and costs. A multivariate matched sampling methodology, nearest available matching on the estimated propensity score, was used to secure 182 matched pairs from the case-managed and non-case-managed groups. An exact match was required on DRG. The additional matching variables were: age, sex, race, socioeconomic status, marital status, route of admission, day of admission, payer group, physician, nursing unit. For each patient in the matched pair, the following additional data were collected: presence of caregiver at home, ICU stay, ICU case management, discharge destination, severity of illness and nursing intensity. A stepwise multiple regression analysis of matched pair differences was used to control for non-matched covariates and to examine the differences in LOS and resource consumption (total cost, room and board costs, and ancillary costs) between the case-managed and non-case-managed groups. The results indicate nursing case management significantly increased length of stay (p-value =.001). Differences in total costs were not significant (p-value =.08). Room and board costs, a good proxy for length of stay, were significantly higher for the case-managed group (p-value =.002). No significant differences were found between the two groups in regard to ancillary charges (p-value =.68). Investigations of subgroups of the sample yielded similar results or showed that nursing case management had no effect on length of stay and resource consumption.
