51 We Test Everyone… Really? Bias in Drug Screening in an Academic Burn Center

Abstract

\n \n \n Previous studies have found that burn-injured patients who use illicit substances, such as methamphetamines, have worse outcomes and longer hospital lengths of stay when compared to those who do not use illicit drugs. It is our units’ stated practice to perform a urine drug screen on all patients admitted with a burn injury. We hypothesize that, while we intend to test all patients, we fall short of this goal. The purpose of this study is to examine our urine drug screening practices.\n \n \n \n Following IRB approval, a retrospective chart review was conducted using electronic medical records of all adult patients admitted to the burn center from 2016–2018. Data collected included information on the burn injury, drug screening, and demographics. Due to the fact that many patients receive opioids and benzodiazepines for pain and anxiety related to their burn injury, these were not considered positive if present on drug screen without gas chromatography to confirm use. Data analysis was conducted using chi-square, t-test, and logistic regression models.\n \n \n \n A total of 1134 patients (mean age 45.9 ± 17.3 years, 855 males (75.4%), mean burn size (TBSA) 12.1±15.5%) were analyzed. Of the 1134 patients admitted, 65% had a urine drug screen performed. Of those who were not screened, 12.1% had a test ordered but not performed while 87.8% had no test ordered. Globally, amphetamines/methamphetamines were the most commonly detected substances in 262 patients (23.1%) followed by cocaine with 14 patients (0.2%) Those with larger burns were more likely to be tested with the mean burn size of those tested being 15.0% while the mean size of those not tested was 6.6%, p< 0.0001. When looking at the effect of age on drug screening, those at the extremes of age (< 20 years (37.5%), 61–70 years (38.9%), 71–80 years (51.5%) and >80 years (76.9%)) were more likely to not be tested (p< 0.0001). All other age groups had a rate of not testing between 30.3% and 33.3%. Additionally, overall men were more likely to be tested than women (68.3% vs. 54.8%, p< 0.0001). On multivariate logistic regression, age (OR 0.99 (0.98–0.99), p< 0.0001), burn size (OR 1.07 (1.05–1.08), p< 0.0001), and female gender (OR 0.61 (0.45–0.81), p=0.0008) were independently associated with a patient receiving a urine drug screen.\n \n \n \n Although our intention is to perform a urine drug screen on all patients admitted with burn injury, we fall short of this goal. A significant number of patients do not get tested and these tend to be older patients, those with smaller burns, and women. Given that drugs of abuse can alter patient outcomes we need to be more rigorous in our efforts to obtain tests on all patients.\n