Background and patient relevance Lay or community health workers (LHWs), i.e., people trained to perform healthcare functions but with no formal professional certification, are one strategy for addressing shortages of health workers to manage chronic conditions in primary and community care. Because LHW interventions are complex and involve a range of tasks, it is challenging to synthesise available trial results for decision-making. We are updating a systematic review on the effects of LHWs for chronic conditions and wanted to maximise the relevance of the findings to stakeholders, including patients and policy users. Objectives To describe two methods innovations to improve the relevance to stakeholders of complex health systems review findings. Methods Innovation 1 - intervention taxonomy: Using the intervention descriptions from 66 randomised trials of LHWs for diabetes treatment, we iteratively developed a draft intervention taxonomy organised by purpose of care. We then tested and refined this taxonomy on a second set of trials of LHWs for hypertension treatment. Innovation 2 – prioritising outcomes: To identify outcomes that patients and other stakeholders view as important, we convened online stakeholder panels, including patients, clinicians, researchers and service managers, to examine the outcomes assessed in included trials; review related clinical core outcome sets; and provide guidance on which outcomes and assessment timepoints should be prioritized. These judgements then guided data extraction and analysis. Results Our intervention taxonomy comprises six categories focused on the purpose of the intervention for patients: health education; social and emotional support; self-management; case management and patient navigation; clinical management; and multi-faceted interventions. We were able to successfully use taxonomy to organise the included studies for synthesis. Included LHWs trials assessed a wide range of outcomes, and we found no core outcome sets for LHWs trials. Stakeholders were key to identifying critical and important outcomes for each condition that we could then use to guide data extraction and analysis. Conclusions Multiple methods innovations are needed within reviews of complex health systems interventions to improve their relevance to stakeholders. Two innovations are presented here as exemplars to be potentially applied to other reviews addressing complex questions.