Statistical requirements in medical diagnostic development across the UK, US and EU markets: a review of regulation, guidelines and standards

Background This review of legislation, official policy guidance and standards across the UK, the European Union (EU) and USA aimed to identify regulatory requirements relating to statistical methodology for diagnostics development outlining barriers and opportunities for novel study designs. Methods The search was conducted between February 2024 and April 2024 for legislation, official policy guidance and standards across the UK, EU and USA. Search terms included: Medical Device, In Vitro Diagnostic, Medical Diagnostic, Diagnostic and IVD . Documents were screened by title, with a subset of documents being double screened. Included documents underwent a search for 25 statistical key terms and statistical requirements were extracted. Eligibility criteria: documents were included if they related to medical diagnostic device development. Documents which related to risk management, postmarket surveillance and quality assurance, while important, were beyond the scope of this review and were excluded. Data sources: EUR-Lex, WestLaw UK, US Food and Drug Administration, Lexis+, Policy Commons, Medical Device Co-ordination Group and the British Standards Online Library . Results 404 documents (total=2479) met the inclusion criteria for data extraction. 128 (32%) had statistical requirements for medical diagnostic development. Of the 1146 requirements identified, 275 were mandatory for specific high-risk tests or conditions (UK only=4, EU only=170, EU/UK=74, USA=27). Recommendations for best practice covered: evidence requirements, statistical validity, study design and management . Discussion Mandatory statistical requirements exist for high-risk areas, limiting novel study designs; however, there remains opportunity to improve efficiency in the diagnostic development pathway for low risk diagnostics.