From Evidence to Data Framework: Decision Factors and Structured Data for AI-Driven Clinical Decision Support Systems in Offloading Footwear

BackgroundDiabetes-related foot ulcers (DFUs) are a serious complication of diabetes, often resulting in infection, amputation, or even mortality. Offloading footwear is a key intervention that promotes ulcer prevention and healing by reducing pressure on affected areas of the foot. However, patient adherence to prescribed footwear remains low. Current clinical guidelines, such as those from the International Working Group on the Diabetic Foot (IWGDF), offer general recommendations but lack a standardized and structured framework. Furthermore, they do not support personalized footwear prescription or integration into clinical decision support systems (CDSSs) and artificial intelligence (AI) applications. ObjectiveThe primary objective of this study was to identify and systematically categorise the key factors influencing footwear prescription and design for patients with DFUs. A secondary objective was to integrate these factors into a structured framework aligned with HL7 Fast Healthcare Interoperability Resources (FHIR) standards, with the goal of informing the development of AI-powered CDSSs. MethodsA systematic scoping review was conducted in accordance with the PRISMA-ScR guidelines. Seventeen academic databases and Google Scholar were searched for relevant studies published between 2014 and 2025. Eligible studies included original research involving adults with diabetes or diabetic foot ulcers (DFUs) and focused on offloading footwear or orthotic interventions. Data were extracted and categorized using the WHO Dimensions of Adherence to Long-Term Therapies framework and mapped to FHIR resources using standardized terminologies such as SNOMED CT, and LOINC. Where existing standards were insufficient, custom FHIR extensions were proposed. ResultsA total of 81 studies met the inclusion criteria, encompassing data from 5,001 participants. Key outcome measures across the studies included plantar pressure reduction (n = 55), adherence (n = 27), gait and balance (n = 20), and ulcer recurrence (n = 13). The review identified 90 unique decision factors influencing footwear prescription, classified according to the five WHO dimensions of adherence to long-term therapies: patient-related, condition-related, therapy-related, socioeconomic, and healthcare system- related factors. Of the 90 factors identified, 48 were categorized as input variables and 42 as output variables. The review found that only 16/90 of the decision factors could be mapped directly to existing FHIR standards, while the remaining 74/90 required custom extensions--particularly those related to detailed footwear attributes and socioeconomic patient data. This is because of the unique nature of the domain as the result of lack of standardized definition related to the domain. ConclusionsThis study introduces the first structured and interoperable framework for prescribing offloading footwear in the management of diabetic foot ulcers (DFUs). By aligning key decision factors with HL7 FHIR standards and the WHO Dimensions of Adherence, the framework enables seamless integration into electronic health records and supports the development of AI-driven clinical decision support systems (CDSSs). Future work should prioritize expert validation of the identified factors, implementation studies, and real-world testing to evaluate the frameworks usability, clinical relevance, and impact on patient outcomes.