Hospitalizations of nursing home residents are a frequent event. A recent review by Castle and Mor (1996) suggests that the hospitalization rate of nursing home residents is between 30 and 45% per year. Although the optimal hospital use rate is not known, it is suspected that current rates in most nursing homes are excessive and are associated with higher health care costs and considerable iatrogenesis. Recently Mor, Intrator et al. (1997) reported a reduced odds of hospitalization within 6 months of 0.70 in 1993 in comparison to 1990 but considerable inter-state variation was observed was observed. Facility and market factors explained as much variation on hospitalization rates as did resident characteristics in a model for hospitalization controlling for death. In this proposal we integrate together two organizational theories, contingency theory and resource dependency theory, in order to test a model of the effects of the effects of facility organization and market structure on the decision to hospitalize a nursing home resident. We propose specific hypotheses that arise from this theoretical model. We will concentrate on availability and utility of specialized technologies in the home, as well as the use of specific protocols, such as those developed in conjunction with the MDS, that my be instrumental in limiting hospital transfers. Throughout we will use state-of-the-art statistical methods to assess the affect of variables in the model on a multi-categorical outcome with multi-level modeling. We propose to use extensive resident level data on nursing home residents from the Case Mix Demonstration Project states (KS, ME, MS, NY, SD), and from a sample of 30 nursing homes for which we have collected detailed data on organizational structure. We have linked these data to HCFA Medicare claims files to build longitudinal histories of treatment setting. Merging information on facilities from the OSCAR (On-line Survey of Certification Automated Records) data, and the ARF (Area Resource File) we will provide the facility and market context (where the market is approximated by the county). Using the models developed as estimated for this large and diverse population we will be able to estimate much more accurately the proportion of explained variation that is attributable to facility and market factors, relative to that attributable to resident characteristics. We will apply the model to two specific conditions for which hospitalization if potentially avoidable. Hospitalization with primary admission diagnosis of dehydration, and hospitalization of diabetic residents with a primary admission diagnosis of hyperglycemia or hypoglycemia. Dehydration may be avoided with adequate nursing care, while hypo-and hyper-glycemia can be easily diagnosed if symptoms are accurately observed and by using a simple test procedure that can be administered at bedside. These conditions can be easily treated.
| Intrator, Orna; Mor, Vincent (2004) Effect of state Medicaid reimbursement rates on hospitalizations from nursing homes. J Am Geriatr Soc 52:393-8 |
| Intrator, Orna; Zinn, Jacqueline; Mor, Vincent (2004) Nursing home characteristics and potentially preventable hospitalizations of long-stay residents. J Am Geriatr Soc 52:1730-6 |
