Return toAll Abstracts
HSR&D 2004 National Meeting Abstracts

2039. VA Administrative Diagnoses of Schizophrenia Versus Chart Diagnoses
John Grabowski, MD

Objectives: Using administrative databases have advantages for research and quality improvement (QI) efforts, including large numbers, consistent data and low cost. However, administrative data may not be consistent with data in clinical records. We examine the concordance of VA administrative and chart note diagnoses of schizophrenia/schizoaffective disorder.

Methods: Using administrative data from two VA medical centers, we identified patients whose most frequent diagnosis in the last 12 months had been schizophrenia or schizoaffective disorder (N=684). Trained MSW RAs collected diagnostic data from the progress notes of patientsí primary psychiatrists.

Results: 72% (n=494) of patients with administrative diagnoses of schizophrenia/schizoaffective disorder had a current clinical diagnosis of schizophrenia or schizoaffective disorder on chart review. An additional 8% (n=53) had clinical diagnoses of another psychotic disorder, 2% had dementia, and 7% had a current diagnosis of bipolar disorder. 12% had a variety of non-psychotic diagnoses such PTSD, major depression, and substance abuse.

Conclusions: The majority of patients flagged by VA administrative data as having schizophrenia or schizoaffective disorder had this diagnosis confirmed on chart review. Most discrepancies involved a chart diagnosis of another psychotic disorder or a disorder often accompanied by psychosis, such as dementia or bipolar disorder.

Impact: Administrative data is a useful first screen for identifying patients with a current clinical diagnosis of schizophrenia or schizoaffective disorder. However, depending upon their purpose, researchers and QI personnel may need to follow this screen with a review of the clinical record because of because of inconsistencies between administrative and chart data.