INTRODUCTION: Residential clients frequently report high psychological distress at intake, but little is known about changes in distress throughout treatment. This study aimed to identify in-treatment trajectories for psychological distress and factors associated with trajectory classes.

METHODS: A retrospective cohort of adults attending Australian non-government residential substance use treatment between 2012 and 2023 was identified from routinely collected data, NADAbase. Participants (N = 1492) completed ≥ 3 Kessler-10 Psychological Distress Scale (K10) assessments within 90 days of intake. Latent growth curve analyses identified classes of K10 trajectories. Multinomial regression identified demographic and clinical correlates (Severity of Dependence Scale [SDS] and EUROHIS Quality of Life scale [EQoL-8]) of class membership.

RESULTS: A five-class model describing K10 trajectories (1: moderate-low improved; 2: high-low improved; 3: very high-moderate improved; 4: very high-high improved; 5: very high unchanged) had the best model fit. Compared to high-low improved (34.5%; referent), moderate-low improved (45.4%) were less likely to identify as female, have higher SDS and lower EQoL-8 scores, or use cannabis; very high-moderate improved (13.1%) were more likely to have lower SDS scores, be aged under 25 and use opioids; very high-high improved (5.6%) were less likely to identify as male, be aged over 25, have higher EQoL and SDS scores; and very high unchanged (1.3%) were more likely to have lower EQoL scores and have left without completing treatment.

DISCUSSION AND CONCLUSIONS: Four K10 trajectory classes showed improvement after 90 days. Around 7% reported sustained high to very high psychological distress. Routine monitoring of psychological distress provides opportunities to identify non-improving clients and review treatment plans to improve outcomes.