<p>Cross-sector interventions are increasingly suggested in care for trauma-affected refugees, but knowledge about how they influence mental health over time remains sparse. Using a qualitative longitudinal design, we explored patterns of mental health change and aspects contributing to change among refugees participating in a cross-sector intervention addressing post-migration stressors alongside treatment for post-traumatic stress disorder. Interviews were conducted with participants at four time points during the intervention and the year after. A pattern-oriented analysis helped identify three patterns of mental health change, in which participants experienced either consistent improvements, decline without the support of the intervention, or persistently poor mental health despite changes to post-migration stressors. The patterns differed in how refugees (a) perceived benefiting from the intervention, (b) desired or benefited from social interactions and support, (c) encountered challenging life events, and (d) explained their expectations and agency. The intervention supported improved mental health in two patterns; however, lasting improvement beyond the intervention period was identified in only one pattern. We discuss adaptations and alternative interventions. Findings support a personalized and cross-sectoral approach to mental health treatment to better support the individual needs of refugee patients.</p>