Validity between self-report and biochemical testing of cannabis and drugs among patients with traumatic injury: brief report

Background The relationship between drug use and traumatic injury is well documented, yet only a small proportion of patients are biochemically tested for cannabis and other substances. The study objective was to determine whether patient self-report can be used as a proxy for biochemical drug testing following traumatic injury. Methods This study was a secondary analysis that included 320 patients admitted to four level I trauma centers in Colorado and Texas, primarily involved in motor vehicle crash (89%). If performed, biochemical testing was collected via urine toxicology screen (“tox screen”) for cannabis, amphetamines, barbiturates, cocaine, opiates, PCP, and benzodiazepines. All patients were screened for self-reported current drug use, which was evaluated for any drug and specifically for cannabis use. Analyses used to compare results of self-reported drug use and tox screen included sensitivity, specificity, positive, and negative predictive values, and percent agreement. Results Among 320 patients, 23% ( n = 75) self-reported drug use; cannabis was the most frequently reported drug ( n = 63). A tox screen was performed in 59% of patients ( n = 190); the proportion of patients who had a tox screen was similar for those self-reporting drug use (60.0%) to those who denied using drugs (59.2%), p = 0.90. Among patients who had a tox screen performed, 18% ( n = 35) tested positive for any drug, 12% ( n = 22) tested positive for THC, and 7% ( n = 13) tested positive for opiates. The percent agreement was 80% for any drug and 81% for cannabis. The specificity was 84–85%, indicating a high likelihood that a patient will not have a positive tox screen if they do not report using drugs. Negative predictive values were 90–95%, indicating a negative self-report correctly identified nearly all patients testing negative on tox screen. Sensitivity was only 60% and positive predictive values were 30–47% for cannabis and drugs, respectively. Conclusion These findings may negate the need for biochemical drug testing in this population, particularly as a “rule out” based on self-reporting. Future studies are needed to confirm these findings and should address risk of selection bias.