Purpose Longer length of stay (LOS) in residential drug and alcohol treatment has been associated with more favourable outcomes, but the optimal duration has yet to be determined for reliable change indices. Optimal durations are likely a function of participant and problem characteristics. The current study aims to determine whether LOS in a residential therapeutic drug and alcohol treatment community independently predicts reliable change across a range of psychological recovery and well being measures.. Design/methodology/approach Three hundred and eighty clients from Australian Salvation Army residential drug and alcohol treatment facilities were assessed at intake and three months post-discharge using the Addiction Severity Index 5th ed., the Depression, Anxiety and Stress Scale, the Recovery Assessment Scale, the Mental Health Continuum- Short Form and the Life Engagement Test. Findings The findings confirm LOS as an independent predictor of reliable change on measures of well being and client perceived assessment of recovery. The mean LOS that differentiated reliable change from no improvement was 37.37 days. Originality/value The finding of LOS as a predictor of reliable change and the identification of an estimated time requirement may be useful for residential drug treatment providers in modifying treatment durations.