Does the perceived accuracy of urine drug testing impact clinical decision-making?

Background: Urine drug testing techniques have different rates of false-positive and false-negative test results. However, clinicians may have highly varying perceptions of test accuracy and may compensate for perceived inaccuracy by incorporating other factors into their interpretation of observed test results. Thus, there is the potential for adverse consequences from decisions based on inaccurate test results or interpretation. Methods: We surveyed 466 members of the American Society of Addiction Medicine to examine clinicians' perceptions of the accuracy of 2 types of urine drug tests, immunoassay (IA) and liquid chromatography-tandem mass spectrometry (LC-MS/MS), and the extent to which behavioral and demographic factors influence the interpretation of test results. Participants read 4 brief vignettes describing positive and negative test results in hypothetical patients who differed along several dimensions (gender, age, race/ethnicity, comorbid mental disorder, court-ordered versus voluntary status, treatment compliance). Outcome variables include likelihood of renewed drug use, likelihood of test error, whether to request additional testing, and whether to report the violation to a probation officer. Results: The strongest predictor of study outcomes was treatment compliance (consistent versus inconsistent attendance), as this was the only independent variable to generate effect sizes of medium strength. Significant effect sizes were also found for type of test used (IA versus LC-MS/MS), legal status (court-mandated versus voluntary), presence of a comorbid mental disorder, treatment history, and race, although effect sizes for these variables were small and less consistently observed. Conclusions: These results highlight the potential for error in clinician judgments about urine drug testing. Not only were participants likely to underestimate the accuracy of "confirmatory" LC-MS/MS testing, but vignettes suggested that a number of historical and demographic factors may influence interpretation of test results.