Patterns of polysubstance use disorder among human trafficking survivors: A latent class analysis

Author: Dell, Nathaniel; Carbone, Jason; Anasti, Theresa; Grimes, Lauren; Preble, Kathleen; Gezinski, Lindsay & Thibodeau, Hilary

Abstract: Substance use is commonly documented among human trafficking (HT) survivors in emergency department (ED) settings. Multiple substance use disorders (poly-SUD) are associated with poor health and psychosocial outcomes. This study identified latent classes and demographic covariates of HT-related ED visits by the types of SUDs documented in survivors’ medical records. We used cross-sectional data from the United States 2019–2021 Nationwide Emergency Department Sample, including visits of patients aged 12–64 years with an ICD-10-CM code documenting either sex or labor exploitation (N = 4,212). A bias-adjusted three-step latent class analysis was conducted, with SUDs documented via ICD-10-CM codes included as indicators in the model. The optimal three-class solution had superior fit based on pre-selected indicators, low classification error, and acceptable entropy. The largest class comprised 76.01 % of the sample and showed a lower predicted probability of the SUD classes considered. The second largest class (17.27 %) was characterized by high predicted probability of stimulant use disorder with moderately high predicted probability of opioid use disorder. The smallest class (6.72 %) was characterized by high predicted probability of each SUD considered. Class membership was differentially associated with disposition from the ED, nicotine use disorder, and income. Although most ED visits were classified as having relatively low probability of SUD, nearly one quarter of the sample had high risk of either stimulant use disorder or high poly-SUD. Poly-SUD in HT survivors is associated with increased risk of hospitalization. Findings provide direction for tailoring intervention programs to support SUD recovery among HT survivors.

Keywords: human trafficking, poly-substance use disorder, latent class analysis, emergency department