D Saliba, Sepulveda GRECC, GLAS VAMC; LV Rubenstein, Center for the Study of Healthcare Provider Behavior, GLAS VAMC; B Simon, Center for the Study of Healthcare Provider Behavior, GLAS VAMC; E Hickey, CHQOER, Bedford VAMC; B Ferrell, UCLA Division of Geriatrics; E Czarnowski, CHQOER, Bedford VAMC; D Berlowitz, Center for Health Quality, Outcomes and Economic Research, Bedford VAMC

Objectives: Pressure ulcer prevention is a key element of nursing home care quality. Clinical practice guidelines for pressure ulcer prevention have been disseminated on a national level, but guidelines improve care only if they are implemented. This study aims to assess (1) overall nursing home (NH) implementation of clinical practice guidelines for pressure ulcer (PU) prevention and (2) variation in implementation rates among a geographically diverse sample of NHs within the VA system.

Methods: We translated the Agency for Health Care Policy Research (now AHRQ) practice guidelines for PU prevention into explicit utilization and review criteria for use in abstracting NH medical records. The review instrument provided specific parameters and time intervals for identifying resident characteristics, risk factors and indicated interventions outlined in the guideline recommendations. Study nurses used this instrument to abstract information from nursing, physician, consultant, physical therapy and dietician notes contained in each resident's NH medical record. Records to be abstracted were identified through the VA Patient Assessment File (PAF) using nested random sampling in a geographically diverse sample of 36 VA NH. 10% of records were re-abstracted to determine inter-rater reliability. Overall and facility specific rates of adherence were determined for 15 PU prevention guideline recommendations and for a subset of six key recommendations identified as having strongest evidence for PU prevention and being most feasible to measure with medical record review.

Results: The NH records of 854 residents were reviewed (mean number of residents per site = 24). Inter-rater agreement for abstraction was 95%. For the 15 guidelines, 6283 guideline indications were identified in the medical records (mean number of indications per resident = 7). Overall adherence to the 15 guideline recommendations when indicated was 41%. Overall adherence for the six recommendations judged by experts as most critical was 50%. For individual guideline recommendations, adherence rates ranged from 94% (skin inspection) to 1% (education of residents or families). While skin inspection achieved the highest rate of implementation, standardized risk assessment-a critical step for systematically targeting subsequent assessments and interventions-was performed in only 60%. NHs differed significantly in their implementation of guidelines. Facility adherence rates ranged from 29% to 51% for the 15 guideline recommendations (p<.001), from 24% to 75% for the 6 key recommendations and from 0% to 100% for standardized risk assessment.

Conclusions: Explicit NH utilization and review criteria reveal low rates of documented adherence to best care practices for PU prevention in NH. In a geographically diverse sample of NH, overall adherence with 15 national guideline recommendations and a subset of six key guidelines was low, and variance in adherence to quality standards among facilities was high

Impact: The high level of variation and low level of adherence we observed in VA nursing homes for this important and common clinical condition indicate a high need for quality improvement. Our findings also identify specific guideline areas in which general improvement across all nursing homes is needed, and others in which some nursing homes excel and could serve as models for improvement.