Patient Similarity Analysis and Decision Support Dashboards in Cancer Care: Patients’ and Other Stakeholders’ Perspectives.

Patient Similarity Analysis can be used to develop personalised aids for patients and clinicians making complex healthcare decisions, including those encountered in the oncology domain. However, little is known about patients’ views on how insights derived through Patient Similarity Analysis may be incorporated into decision making at the point of care. In this study, we developed a set of Patient Similarity-based dashboard mock-ups and consulted oncology stakeholders in Aotearoa New Zealand regarding acceptability. Aim: To explore oncology stakeholders’ (patients, clinicians, researchers and advocates) perspectives on the acceptability of using Patient Similarity-based decision dashboards to guide decision making. Methods: This is a qualitative descriptive study using non-random, purposive sampling, combined with advertisement and snowball recruitment. We interviewed patients, healthcare providers and other stakeholders. One-off, semi-structured interviews were conducted to elicit perspectives on the acceptability of using decision support dashboards, including: participants’ attitudes, the tools fit with the participants’ values and goals, and the potential benefits and burdens. Data was analysed using Directed Content Analysis. Results: Thirty-one participants were interviewed: 19 patients with breast or prostate cancer, seven clinicians, and five other stakeholders. Participants found the dashboard mock-ups generally acceptable, with the information presented considered relevant by most. Participants expressed enthusiasm about using Patient Similarity-based insights, as they made them feel confident in making decisions knowing what patients like them decided in similar circumstances. Participant-identified benefits of using Patient Similarity-based dashboards were: increased relatability to care recommendations based on other similar patients’ experiences; information consolidation, including about non-standard treatments; and shareability with whānau. Participants thought such dashboards would enable patients to actively participate in care decisions, enhancing equitable access to health information. A key challenge noted by both clinicians and patients related to clarity regarding the definition of ‘similar’, which would impact meaningfulness and reliability of recommendations. Conclusion: The proposed dashboard mock-ups were considered acceptable, with Patient Similarity insights reported as valuable and desired by participants. Several considerations and challenges are reported. These findings are relevant to researchers, software developers, health care providers and policy makers developing and/or implementing decision support tools in cancer care and other healthcare settings.