All Medicare and/or Medicaid certified nursing homes (NH) undergo an annual survey to ensure that acceptable quality of care is provided to residents. If substandard care is found during a survey, a deficiency citation can be issued to the facility, identifying the type and severity of the infraction. Although the Centers for Medicare and Medicaid services (CMS) provides training and information recording materials to surveyors, past research has uncovered wide variation in the issuance of deficiency citations across states that cannot be explained fully by differences in NH quality of care. Despite the critical role annual surveys play in monitoring NH quality of care, little is known about the extent to which variations in frequency, types, and severity of deficiency citations reflect actual differences in quality of care versus broader systemic differences in the annual survey process itself. In response, we propose to use spatial analysis techniques to examine regional variation in the issuance of NH deficiency citations across the United States. To this end, our specific aims include, to: (1) describe regional patterns of NH deficiency citations, (2) examine alternative definitions of regional areas (e.g., zip code location, county location, algorithmically defined clusters) and identify spatially meaningful patterns and clusters, and (3) examine whether certain facility characteristics (e.g., for-profit status, high Medicaid concentration) are regionally associated with higher concentrations of certain types/quantities of deficiency citations. A secondary analysis of four years of data from the On-Line Survey Certification and Reporting (OSCAR) system linked with data from the Area Resource File (ARF) and the 2000 U. S. Population Census is proposed. Data will be spatially assigned to map locations and standardized to adjust rates of deficiencies for differences in the density of NH beds across regions. Spatial analysis techniques will be used to visually and statistically examine the frequency, type, and severity of deficiency citations across geographic areas. Moran's I statistic will be used to test for spatial autocorrelation across regions and to identify 'hot spots' and clusters. Multilevel models will be estimated to explore the relative contribution of facility-level factors versus broader systemic influences to regional variations in the issuance of deficiency citations. This study will be the first to use geospatial methods to explore small area regional differences in the issuance of NH deficiency citations and will lay the foundation for future studies exploring regional variations in NH quality of care. Findings will provide greater understanding of regional variations in the issuance of deficiency citations and will assist ongoing efforts to improve the oversight, regulation and reporting of NH quality.of care as well as identify potential strategies for targeting quality of care improvement initiatives to those areas most likely to benefit from such programs. ? ? ?
