Development of clinical and economic prognoses from Medicare claims data.

Using a 5% nationally random sample of Medicare beneficiaries, we calculated the probability of dying, the probability of being readmitted, and the mean level of inpatient hospital expenditures within various time periods following discharge for those beneficiaries who were discharged alive from an acute-care hospital during 1983. We then examined the 674 most common principal discharge diagnoses and found significant variations by diagnosis code for all three outcomes. We believe that analyses of claims data by diagnosis code can provide useful information to clinicians and their patients regarding the clinical and economic prognosis of specific diseases, help managed-care programs identify patients likely to incur substantial costs over a several-year period, and inform insurers regarding the expected level of resources that will be used following discharge for patients with specific diseases.