Authors: Adam Mohamed Saber Mohamed , Amira Gaber Ibrahim Mostafa , Moutaz Ali Kasasbeh , Mohammed Abdullah Alhejji , Abobakr Elmotesim Abdulla Elamin , Hassan Ahmad Khubrani , Sameera Gamal Moubark Moahmed

Abstract

Background: Substance use disorders (SUDs) are chronic relapsing conditions with high relapse rates and substantial burden. Inpatient rehabilitation length of stay (LOS) has been identified internationally as a determinant of treatment outcome, but evidence from Saudi Arabia is limited. This study examined the association between inpatient LOS and relapse, and identified demographic and clinical factors associated with relapse.

Methods: A retrospective cross-sectional correlational study using inpatient medical records and follow-up data at ReLib Behavioral Rehabilitation Center, Riyadh. Records of 125 adult patients (aged 18 years or above) admitted between January and December 2025 were reviewed. Patients were observed from discharge to a common census date of 30 June 2026 (median follow-up 306 days, IQR 164–424). Relapse was the primary outcome. Analyses comprised chi-square and Fisher’s exact tests, the Mann–Whitney U test, and multivariable binary logistic regression with collinearity diagnostics. Sensitivity analyses restricted the cohort to patients with guaranteed minimum follow-up windows. Significance was set at p < 0.05.

Results: Mean age was 33.8 ± 10.8 years; 80.8% were male. The relapse rate was 36.8% (n = 46). Mean LOS was 56.6 ± 53.3 days, 79.3 days in non-relapsed versus 17.5 days in relapsed patients (Mann–Whitney p < 0.001). Relapse by LOS category was 97.1% (1–14 days), 29.0% (15–30 days), 6.9% (31–90 days) and 6.5% (>90 days); χ2= 77.31, p < 0.001. This gradient persisted in sensitivity analyses restricted to patients with at least 90 and at least 180 days of follow-up. In multivariable analysis (model χ2= 81.64, p < 0.001; Nagelkerke R2 = 0.655; AUC = 0.935) LOS (OR = 0.957 per day, 95% CI 0.933–0.980), readmission (OR = 6.87, 1.25–37.90) and discharge against medical advice (DAMA; OR = 6.07, 1.70–21.71) were independently associated with relapse. Variance inflation factors were 1.19–1.81, indicating no problematic collinearity.

Conclusion: Shorter inpatient LOS, readmission and DAMA discharge were strongly associated with relapse, with an apparent inflection near one month. Because the design is observational, these associations should not be interpreted as establishing that extending LOS prevents relapse; early discharge may equally be a marker of unmeasured severity, engagement or motivation. The findings identify early-discharge patients as a group warranting intensified aftercare and generate a threshold hypothesis requiring prospective validation.

Keywords: Substance use disorders, length of stay, relapse, inpatient rehabilitation, Saudi Arabia, logistic regression, predictors of relapse, discharge against medical advice.