Abstract
BACKGROUND: Despite the growing utilization of data-driven methods to investigate multimorbidity patterns, there is currently no consensus or guidance on the conditions to include when identifying patterns. This scoping review aims to systematically examine the nature of conditions included in existing studies using data-driven techniques. METHODS: A comprehensive search of three electronic databases (MEDLINE, Web of Science and Scopus) was conducted to identify relevant publications from inception to 28 February 2022 using predefined search terms and inclusion/exclusion criteria. The reference lists and citations of relevant papers were also searched. RESULTS: Among 7326 search results, 5444 relevant articles were identified. After screening against the eligibility criteria, 60 articles were included in the review. Half of the reviewed studies reported selection criteria for conditions, with prevalence in the population of interest being the most common criterion (40%). Most studies included at least one neurological [59 (98.3%)], musculoskeletal [58 (96.7%)], respiratory [57 (95.0%)] or mental health [56 (93.3%)] condition. In contrast, only a small proportion of studies included skin [17 (28.3%)], infections [14 (23.3%)] or autoimmune conditions [10 (16.7%)]. Nine conditions (hypertension, diabetes, cancer, arthritis, COPD, asthma, depression, stroke and osteoporosis) were included by more than half of the studies. CONCLUSIONS: This review highlights the considerable heterogeneity among the conditions included in analyses of multimorbidity patterns. Researchers should provide a clear rationale for the selection of conditions to facilitate comparisons across studies and ensure reproducibility, as well as consider selecting a diverse range of conditions to capture the complexity of multimorbidity.