Project Background: Risk adjustment is recognized as critical for accurate assessment of quality, fair comparison of providers, and benchmarking across healthcare systems. As chart-based data -- the 'gold standard' for risk adjustment -- are costly and time-consuming to collect, administrative data have remained the basis of risk adjustment, despite inadequacies in capturing patient severity. Recent initiatives by the Agency of Healthcare Quality and Research (AHRQ) to enhance administrative data with automated data on laboratory tests and vital signs are an important step towards improving the accuracy of risk-adjustment models. Project Objectives: Taking advantage of readily available VA automated data, the primary goal of this project is to develop cost-effective and clinically sound enhanced risk-adjustment models to profile VA facilities on 30-day mortality following hospital admission. Enhanced risk-adjusted mortality rates -- and length of stay and readmission (secondary outcomes of interest) -- will be obtained for selected patients cohorts by combining administrative data with automated data on pharmacy claims, laboratory test measures, and vital signs. Our specific objectives are to: 1) examine performance of models for predicting 30-day mortality and length of stay using administrative data; 2) estimate the improvement in model performance from sequentially adding risk factors identified in pharmacy claims, laboratory tests and vital signs data; and 3) assess the impact of enhanced risk adjustment on facility rankings of risk-adjusted mortality. Project Methods: Separate enhanced risk-adjustment models will be estimated for nine medical conditions that account for a sizable proportion of all VA admissions (16 percent) and inpatient deaths (30 percent) - acute myocardial infarction, congestive heart failure, cirrhosis and alcoholic hepatitis, chronic obstructive pulmonary disease, gastrointestinal hemorrhage, hip fracture, pneumonia, acute renal failure and acute stroke. Based on all admissions to VA facilities during FY2003- 2008, disease-specific cohorts will be extracted. We will merge VA inpatient and outpatient administrative data with laboratory test and pharmacy claims data from Decision Support System (DSS) files and vital signs data from Corporate Data Warehouse (CDW). Starting with a standard risk-adjustment model based on administrative data, we will evaluate the improvement in predicting 30-day mortality and length of stay with each increment of additional data from other sources. Model development and hypothesis testing will be based on estimation of hierarchical, multivariate logistic and linear regression models and bootstrap sampling. Project Implications: Findings from this study can become the basis for developing formal quality monitoring and improvement mechanisms for medical inpatient admissions, similar to VA's National Surgical Quality Improvement Program (NSQIP) for surgical admissions.
Risk adjustment is recognized as critical for accurate assessment of quality, fair comparison of providers, and benchmarking across healthcare systems. Taking advantage of the rich VA automated clinical databases, this study will develop clinically sound enhanced risk adjustment models for profiling VA facilities on short-term mortality for disease conditions that rank high on the number of VA inpatient deaths. Enhanced risk-adjusted inpatient mortality rates -- and length of stay and readmission (secondary outcomes of interest) -- will be obtained for selected patients cohorts by combining administrative data with automated data on pharmacy claims, laboratory test measures, and vital signs.
