Health recommender systems (HRSs) are increasingly being proposed to support personalized and data-driven healthcare decisions, yet their translation into real-world practice remains limited. Existing reviews typically analyze HRSs through isolated dimensions such as algorithms, application domains, or evaluation methods, offering limited insight into why technically advanced systems rarely progress beyond prototypes. Following PRISMA 2020, we reviewed 136 peer-reviewed journal articles published between 2014 and 2025 from Scopus and PubMed. We adopted a lifecycle-oriented analytical framework, examining HRSs across six interrelated dimensions: clinical intent, data governance, recommendation logic, user interaction, evaluation strategy, and clinical integration and ethics. Our findings show that limited clinical adoption is not primarily driven by algorithmic immaturity but by persistent misalignment between intended use, evaluation design, and integration pathways. By reframing translational stagnation as a lifecycle coherence problem rather than a technical one, this review provides a unifying perspective for designing clinically credible and deployment-aware HRSs.
MIAYAR et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: