Nursing facility (NF) residents, typically suffer from a constellation of chronic illnesses that leave them frail and vulnerable. For these individuals, hospital care may be necessary to cure serious acute illness, manage clinically complex chronic care, and restore lost function. However, it can also be disruptive and escalate functional decline. Data from NMES indicate that hospitalization rates among NF residents are notably higher than for the elderly as a whole. Further, a recent study found that nearly half of all hospitalization for NF residents was inappropriate. Both the costs and the clinical implications for this population, point to the need for a more comprehensive evaluation of factors contributing to variations in practice style and unnecessary hospitalization. This project will develop a comprehensive model of the determinants of NF hospitalization rates that includes economic and environmental incentives, NF and resident characteristics, and differentiates types of hospitalization. Because many studies point to the need for greater medical attention within Nfs, we also include a set of variables labeled physician-NF interaction that describe NF practices that might increase physician presence within NFs. We use an expert panel of practicing geriatricians to distinguish hospitalizations that are unavoidable from those that either could have been prevented or involved considerable discretion in the decision to hospitalize. The clinical capabilities available within a NF may influence these classifications. Data for the study come from two existing (Minimum Data Set (MDS)) databases on NF residents (from 1992-1995, and 2001-2002) which will be linked to Medicare hospital claims data. These will be supplemented by surveys of NF Medical Directors and Directors of Nursing to obtain information on policies and procedures surrounding the hospitalization of residents, the availability of clinical resources and physicians within the facility, and staff attitudes towards hospitalization. NF characteristics from the Medicare's facility certification files and local area resources (from HCFA's area resource file) will be appended to the two databases. We will also develop a (case-mix adjusted) NF profile of hospital use to monitor hospitalization rates and potentially quality of care. The validity of our profiling tool will be assessed through Comparison with other NT quality indicators (from the MDS) thought to be related to hospitalization and by an analysis of mortality rates.
