Design and implementation of Electronic Health Record Computerised Decision Support System (CDSS) trials: A Systematic Review

ABSTRACT Objectives Conduct a systematic review of the existing evidence base pertaining to the conduct of randomised controlled trials of clinical decision support systems embedded within electronic health record systems. Further, to describe whether key features of trial design and implementation were consistently reported. Materials and Methods A systematic search of MEDLINE was conducted in April 2022. Three independent reviewers screened the search results. A 27-item checklist was used to extract data from the screened studies. A subgroup analysis was conducted to classify trials of clinical decision support systems based on whether they encouraged guideline adherence or represented new knowledge generating mechanisms. Results 5,213 records were retrieved. Following screening, 106 studies were included in the review. The majority of studies evaluated active alerts seeking to improve adherence to clinical guidelines rather than generate new knowledge. Few studies quantified the existing ecosystem of decision support at the study site, or explored phenomena like alert fatigue. Discussion This systematic review provides a detailed analysis of the characteristics of trials evaluating clinical decision support systems. It highlights significant under-reporting of key factors which may affect the reproducibility and generalisability of trial results - particularly with respect to measurement of alert fatigue, description of the underlying digital ecosystem and additional co-interventions used within trials. Conclusion As clinical workflows undergo digital transformation, randomised controlled trials of clinical decision support systems require greater standardisation, in both conduct and reporting. This represents an area of expanding interest given the increasing use of artificial intelligence-enabled decision support. STRENGTHS & LIMITATIONS - This study presents the results of an updated systematic review of studies evaluating the effectiveness clinical decision support systems. - It used a comprehensive checklist to extract detail pertaining to five information domains on trial quality and description. - Studies were evaluated to determine whether the clinical decision support system was knowledge generating or designed to improve guideline adherence. - The review was limited to randomised trials and excluded quasi-experimental and observational studies of clinical decision support systems.