Assessing the accuracy of administrative data in health information systems.

摘要: Background: Administrative data play a central role in health care. Inaccuracies such are costly to systems, they obscure research, and affect the quality of patient Objectives: We sought prospectively determine accuracy primary secondary diagnoses recorded administrative sets. Research Design: Between March July 2002, standardized patients (SPs) completed unannounced visits at 3 sites. abstracted 348 medical records from these obtain written made by physicians. also examined files identify on encounter forms extracted computerized databases. Because correct diagnosis was defined SP visit, we could whether final set and, if not, it caused physician diagnostic error, missing forms, or incorrectly filled out forms. Subjects: General internal medicine outpatient clinics 2 Veterans Administration facilities large, private center participated this study. Measures: A total 45 trained SPs presented physicians with 4 common conditions. Results: The for 57% visits. Thirteen percent errors were 8% 22% entered data. Findings varied condition site but not level training. Accuracy (27%) even poorer. Conclusions: Although more research is needed evaluate cause inaccuracies relative contributions patient, provider, system effects, appears that significant common. Interventions aimed correcting appear feasible.