Review of analysis techniques in mental health research with consumer instruments – a guide for researchers

Purpose: This integrative review aimed to identify and synthesis literature on analysis techniques and methodological approaches used to analyse consumer measures in mental health research. Design/methodology/approach: The review included papers published up to January 2020 across seven databases: CINAHL, Web of Science, Medline, PsycINFO, EMBASE, Scopus and Google Scholar. Data search and extraction was conducted according to the recommended guidelines for conducting review by Cochrane and Joanna Briggs Institute. Mixed method synthesis was used to integrate both qualitative and quantitative data into a single synthesis. Findings: The initial search yielded a total of 2,282 papers. A total of 32 papers were included in the synthesis. Most of the included papers (25/32; 78.12%) focused on psychometric properties, whereas 14% (5/32) targeted analysis techniques, and 6.3% (2/32) addressed methodological justification. The measurement models (e.g. psychometric properties) were analysed through validity and reliability testing as part of instrument development and adaptation. The structural models were analysed using techniques such as structural equation modelling, multivariable regression models, intraclass correlation coefficient and partial least squares–structural equation modelling. Practical implications: Although consumer-reported instruments are analysed using techniques involving linear, hierarchical and longitudinal effects, no attempt has been given to procedures that applied complex data mining or machine learning. Consumer researchers, clinicians and quality management are encouraged to apply rigorous analysis techniques to critically evaluate consumer outcome measures. Originality/value: This review provides evidence on the analysis techniques in mental health research to inform the training of mental health professionals, students and quality assessment practitioners.