Co-research in the development of AI and digital health tools for cancer management and care: a systematic review.

backgroundParticipatory approaches (including co-research, co-design, Patient and Public Involvement [PPI], and Participatory Action Research [PAR]) are increasingly being incorporated into digital health innovation. However, their application in cancer-related digital and Artificial Intelligence (AI) tool lifecycle remains insufficiently synthesised, limiting understanding of how participatory approaches influence ethical outcomes, trust, and the real-world implementation of these technologies. This review moves beyond documenting participatory approaches to critically examine how far participation extends across the AI lifecycle, and how this limits ethical and trust-related outcomes.

objectiveTo examine how participatory approaches are used in the lifecycle of digital and AI-based tools for cancer care, and to analyse their ethical and trust-related impacts.

methodsA systematic review was conducted following JBI methodology. Searches of PubMed, Embase, Scopus, and ScienceDirect identified studies published in English between 2010 and 2025. The inclusion criteria required empirical research reporting the participatory involvement of patients, survivors, caregivers, or public contributors in the design, development, implementation, or governance of digital or AI tool development for cancer care. Quality appraisal was performed using the CASP checklist. Data were synthesised using a meta-aggregation approach and organised into thematic domains.

resultsOf 2,742 records identified, 40 studies met the inclusion criteria. The largest proportion was published in 2024. Co-design emerged as the most frequently used participatory approach, particularly in early development stages (e.g., requirements identification, content co-creation, usability testing), while participatory involvement in later stages, such as implementation, auditing, or governance, was limited. Participation centred primarily on patients/survivors and clinicians, with caregivers and policy-level stakeholders comparatively under-represented. Most studies focused on general digital health technologies such as mobile apps and telehealth, with relatively few addressing AI-specific development components such as model training, validation, or explainability. Very few studies reported explicit measurements of ethical outcomes such as trust, fairness, or transparency.

conclusionsParticipatory approaches are widely referenced but remain predominantly concentrated in early design phases of digital tool development. Future research should extend co-research into later lifecycle stages, include broader stakeholder groups beyond patients and clinicians, and systematically capture ethical and trust-related outcomes. These shifts are necessary to move beyond symbolic participation toward demonstrable and equitable impact in the development of digital and AI-enabled technologies for cancer care.