Challenges to Achieving Adequate Sample Sizes for Facility Level Evaluations of Nursing Home Resident Satisfaction

Abstract

Residents with cognitive limitations by necessity are excluded from nursing home satisfaction initiatives. We used national data from the Minimum Data Set (MDS) to examine the distribution of resident cognitive status across nursing homes to gain further insight into how residents with cognitive limitations could impact the sample size of respondents. We estimate the associations of cognitive status (using the MDS Cognitive Performance Scale, [CPS]) with length-of-stay, bed size, and hospital affiliation. The analyses were performed using approximately 3.5 million MDS records collected from nursing homes during 2001. The results show adequately precise information for long-stay residents will be available from 4 to 46% of hospital-based facilities and 67-90% of freestanding facilities, whereas information from short-stay residents will be available from 34 to 95% of hospitalbased facilities and 14-50% of freestanding facilities. Several initiatives are underway to measure and report the satisfaction of nursing home residents. In all of these large satisfaction initiatives the sampling strategy of respondents needs to be sufficient to allow for nursing home comparisons, and second, to allow interfacility benchmarking comparisons. We highlight the difficulty in obtaining information for these initiatives.