Leonard T Nguyen1, Maggie Guo1, Brenda Hemmelgarn2, Hude Quan2, Fiona Clement2, Tolulope Sajobi2, Roger Thomas3, Tanvir C Turin3, Christopher Naugler4. 1. Department of Pathology and Laboratory Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. 2. Department Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. 3. Department of Family Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. 4. Department of Pathology and Laboratory Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada; Department Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada; Department of Family Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. Electronic address: christopher.naugler@albertahealthservices.ca.
Abstract
BACKGROUND: There is widespread variation in testing practice among practitioners, however there has been no objective way to pinpoint target tests for utilization management. We propose to take advantage of inter-physician variance in clinical practice as a quantitative measure to generate lists of potentially misutilized tests. METHODS: Testing frequencies from a database of clinical testing volumes for outpatients in Calgary, Canada, were obtained for the study period of 2016. For each chemistry, microbiology or hematology test, an arithmetic coefficient of variation (CV) was calculated from family physicians' ordering frequencies. RESULTS: The mean CV for all 358 tests considered was 219% (95% CI 206-231%) with a range of 52-729%. The highest variance was observed for human T-lymphotropic virus antibody testing and several tests for heavy metal levels (mercury, copper, zinc and chromium). Among the 100 most commonly run tests, high variance was found for several endocrinology tests including cortisol. CONCLUSIONS: The utility of ranking clinical tests by ordering variance presents a practical approach to evaluate relative variation in physician practice strategy and to identify potential areas of misutilization.
BACKGROUND: There is widespread variation in testing practice among practitioners, however there has been no objective way to pinpoint target tests for utilization management. We propose to take advantage of inter-physician variance in clinical practice as a quantitative measure to generate lists of potentially misutilized tests. METHODS: Testing frequencies from a database of clinical testing volumes for outpatients in Calgary, Canada, were obtained for the study period of 2016. For each chemistry, microbiology or hematology test, an arithmetic coefficient of variation (CV) was calculated from family physicians' ordering frequencies. RESULTS: The mean CV for all 358 tests considered was 219% (95% CI 206-231%) with a range of 52-729%. The highest variance was observed for human T-lymphotropic virus antibody testing and several tests for heavy metal levels (mercury, copper, zinc and chromium). Among the 100 most commonly run tests, high variance was found for several endocrinology tests including cortisol. CONCLUSIONS: The utility of ranking clinical tests by ordering variance presents a practical approach to evaluate relative variation in physician practice strategy and to identify potential areas of misutilization.
Authors: Leonard T Nguyen; Maggie Guo; Brenda Hemmelgarn; Hude Quan; Fiona Clement; Tolulope Sajobi; Roger Thomas; Tanvir C Turin; Christopher Naugler Journal: Data Brief Date: 2019-08-12