Background: Over the last few years, the USA has seen a sharp increase in the incidence and prevalence of substance use disorder (SUD) and related acute illnesses. The clinical course of many hospitalized patients is impacted by their diagnosis of SUD. At our institution, we put in place an inpatient addiction service where a physician specialized in addiction services would meet the patients in the hospital and initiate treatment as indicated. The primary objective of our study is to assess the impact of the inpatient addiction service on outcomes in patients with substance use disorder.
Methods: We included patients aged 18 or older, who were hospitalized for 24 hours or greater between March 2018 and January 2022 and who had a diagnosis of SUD. We performed a retrospective chart review of a randomly selected convenience subset sample of patients. We recorded demographic and clinical variables and analyzed the association between the occurrence of addiction medicine consultation (AMC) and the outcomes of left against medical advice (AMA) and 30-day readmission (READ). We performed univariate and multivariable analyses.
Results: There were 1,020 encounters included in our study: 949 (93%) without AMC and 71 (7%) AMC. AMC patients were younger (53.1 (+/- 7.9) vs. 56.93 (+/- 10.2), P < 0.001), and were less likely female (23.9% vs. 42.8%, P=0.002) and insured (73.2% vs. 85.5%, p=0.007). There was a similar distributions in race, Years of use (30 years), and hospital length of stay (4 days) between AMC and non-AMC groups. The AMC group had a higher rate of intravenous drugs (26.8% vs 9.9%, < p< 0.001) and opiates (49.3% vs. 19.5%, P< 0.001) but similar rates of cocaine usage (35.2% vs. 36.6%, P=0.819). There was no statistical difference in the rate of consultation with an addiction specialist (inpatient or outpatient) previous to current admission 38% vs. 30%, P=0.158. A higher proportion of AMA was in the AMC group (18.3% vs. 8.6% P=0.007), but less readmission (5.6% vs. 18.2%, P=0.007). We then performed a multivariate regression analysis and after adjusting for demographic variables, there was evidence that AMC was not associated with the outcome AMA (P=.183) but was shown to have a much lower likelihood of 30-day readmission (0.3[0.1-0.8], P=0.016).
Conclusions: We found that the practice of AMC was associated with similar rates of AMA but fewer 30-day readmission rates in patients with SUD. The similar rate of AMA could be due to the higher propensity of clinicians to consult AMC if patients are threatening to leave the hospital. The findings of our study should be confirmed in larger trials but hold significant promise.