Background: There is little guidance on how to select the best approach to broad review questions. Challenged with a broad question about the effectiveness and harms of electroconvulsive therapy (ECT), both from the clinical and the patient perspective, our review team drew on published examples and guidelines in developing a novel methodological approach. Only a few worked examples of integrated reviews with qualitative and quantitative evidence exist, while no Cochrane reviews have embedded qualitative evidence in an effectiveness review. The Cochrane Handbook identifies four ways in which qualitative evidence can contribute, signalling an openness to embed qualitative evidence in Cochrane reviews. Objective: To present a novel integrative evidence approach and our experience on the input side. Methods: Two underlying principles guided our development of the integrative review approach: i) adherence to an explicit, systematic method as described by the Cochrane Handbook; and ii) recognition that the review question must drive the review method. Results: An integrative stepwise evidence approach (Figure 1) allowing for synthesis of findings from non-intervention studies within the framework of a standard effectiveness review was developed. Quantitative and qualitative data are advanced through four phases, and the results connected between the data analyses phases and again during interpretation of findings. Quantitative data have methodological dominance, but qualitative data inform the effectiveness review by refining the PICO elements and extend it by integrating first-hand patient experiences of benefits and harms. The approach resulted in a comprehensive search strategy where 4787 references were reviewed. The integration of the varied ECT evidence addressed the trade-off between breadth versus depth and generalisability versus specificity of a contested treatment. The appropriateness of the integrative review method for complex and controversial topics is highlighted. Conclusions: Methods for integrative evidence synthesis are still evolving and we present one promising approach.