ISQUA16-3153THE IMPLICATIONS OF VARIATION IN RISK ADJUSTED COST OF CARE OF HOSPITALIZED PATIENTS ON COMPLICATIONS, COST, LENGTH OF STAY, AND READMISSION WITHIN 30 DAYS

Abstract

Objectives: 1. Present a risk adjustment strategy for addressing variation of cost of care 2. Describe variation in risk adjusted cost of hospital care at US academic medical centers 3. Illustrate implications of risk adjusted cost variation relative to outcomes of inpatient care: cost, length of stay, complications, in-hospital death, readmission. Method(s): This is a retrospective cohort study incorporating risk adjustment models to predict cost of inpatient care. The study cohort included adults hospitalized in US academic medical centers between October 1, 2014, and September 30, 2015. Risk adjustment for 330 base Medicare Severity Diagnosis Related Groups (categories of MS-DRGs related by having the same principal reason for care), was the basis for the cost risk adjustment models. Vizient's clinical database/resource manager (CDB/RM) was the source of the data; it is populated by nearly 300 hospitals (115 academic medical centers and 185 of their hospital affiliates). Variation was calculated as the standard deviation of risk adjusted costs by hospital and MS-DRG, and z scores were generated and averaged by hospital for hospital comparisons. Result(s): Cost variability was found to be a predictor of variation in outcomes, including length of stay, the rate of complications of care, in-hospital death, and readmissions. Greater cost variability was associated with longer stays, higher complication rates (nearly 100% higher in the highest quartile of hospital z scores), higher in-hospital mortality (more than 7 times higher), and more frequent readmissions. These observations were consistent across several dimensions: MS-DRG, Major Diagnostic Category (MDC), hospital, and time. Risk adjusted cost for patients with complex pneumonia, coagulation, and bone disorders as well as infected knee procedures and skin trauma had the most variation in cost while transplant patients (liver, kidney, pancreas, and pancreas/kidney) were among the case types with the least cost variation. Conclusion(s): Strategies for reducing cost variation should include initiative related to better managing characteristics of patient's stays, for example, length of stay (for example, timely discharge), complications (for example, investing in patient safety initiatives), failure to rescue (for example, building rapid response teams), or appropriate discharge to avoid readmission.