BACKGROUND: Few studies have evaluated the implementation and impact of real-world mental health programmes delivered at scale in low-resource settings. AIMS: To describe the cross-country research methods used to evaluate district-level mental healthcare plans (MHCPs) in Ethiopia, India, Nepal, South Africa and Uganda. METHOD: Multidisciplinary methods conducted at community, health facility and district levels, embedded within a theory of change. RESULTS: The following designs are employed to evaluate the MHCPs: (a) repeat community-based cross-sectional surveys to measure change in population-level contact coverage; (b) repeat facility-based surveys to assess change in detection of disorders; (c) disorder-specific cohorts to assess the effect on patient outcomes; and (d) multilevel case studies to evaluate the process of implementation. CONCLUSIONS: To evaluate whether and how a health-system-level intervention is effective, multidisciplinary research methods are required at different population levels. Although challenging, such methods may be replicated across diverse settings.