Return toAll Abstracts
HSR&D 2004 National Meeting Abstracts

1026. Does Better Performance Lead to Better Outcomes in the VA System?
John E Schneider, PhD, Iowa City VAMC and University of Iowa College of Public Health, LE Kazis, CHQOER, Bedford VAMC and Boston University School of Public Health, YH Huang, CHQOER, Bedford VAMC, J Yankey, Iowa City VAMC, T Vaughn, Iowa City VAMC and University of Iowa College of Public Health, BN Doebbeling, Roudebush VA and Indiana University School of Medicine

Objectives: We compare the effects of two different VA Medical Center (VAMC) summary performance measures on the SF-36 summary scores of patients treated at VAMCs, controlling for differences in patient and facility characteristics.

Methods: Diagnosis data were linked to the Large Veterans Health Survey (n = 824,263) and merged with facility-level benchmarking data based on the External Peer Review Program (EPRP) and the Veterans Health Survey (VHS). EPRP is based on chart review of chronic disease quality measures and VHS assesses preventive measures for a random sample of primary care patients; both indicators measure adherence to evidence-based practices (higher is better). Descriptive and multivariable models were estimated to identify variables associated with SF-36 summary scores.

Results: In 1998, patients treated in VAMCs characterized by higher EPRP scores had SF-36 physical component scores 3.4 points lower, whereas those treated in VAMCs characterized by higher VHS scores had physical summary scores 0.80 points higher (p = 0.01). In 1999, both the EPRP and VHS effects were negative but smaller in magnitude (1.70 and 0.44 points, respectively; p = 0.01).

Conclusions: (1) There is a significant relationship between facility performance scores and patient outcomes; and (2) the direction of the relationship differs according to the underlying methodology and content of the performance measure.

Impact: Improving organizational efficiency and quality turns on the ability to link specific organizational actions with the quality of the final “product.” However, our results suggest that similar measures may capture different elements of performance.