User-Centered Methodological Framework for Development and Evaluation of Digital Technologies in Healthcare: A Case Study of a Machine Learning-Based Lower-Limb Exercise Training System for Older Adults with Knee Pain

The integration of digital technologies into healthcare promises great advances in improving the delivery of healthcare services. However, the use and adoption of digital health technologies are often hindered by poor usability, low user acceptance, or failure to deliver expected health outcomes. To address these challenges, this paper presents a user-centered methodological framework for the development and evaluation of digital health technologies, with a focus on usability, user acceptance, and health effects. The framework consists of five phases: Phase 1 − Pilot testing of technical feasibility, Phase 2 − Development of initial prototype and needs analysis of end-users, Phase 3 − Development of technology-integrated prototype and usability evaluation, Phase 4 − Longitudinal implementation and evaluation, and Phase 5 − Refinement and validation. We conducted a case study to demonstrate how this framework was applied to develop and evaluate a machine learning-based lower-limb exercise training system for older adults with knee pain. With the use of this framework, the advanced machine learning techniques are effectively deployed into healthcare applications; they are found to be usable and acceptable, and show health benefits among end-users. This methodological framework offers practical guidance for researchers and developers to create, evaluate, and refine digital health technologies that address adoption barriers while delivering clinical value.