Investigating the Impact of AI on Shared Decision-Making in Post-Kidney Transplant Care (PRIMA-AI): Protocol for a Randomized Controlled Trial.
Osmanodja, B., et al.JMIR Research Protocols 2024; 13: e54857.
Aims
The aim of this protocol for a randomised controlled trial is to determine the impact of using an artificial intelligence (AI)-based risk prediction tool for promoting conversations about post-kidney transplant treatment options following graft loss, as well as to evaluate how it influences the associated shared decision making (SDM) process.
Interventions
Participants will be randomised to receive either routine care alone or to AI-supported care in addition to routine care.
Participants
Kidney transplant recipients with low estimated glomerular filtration rate (eGFR).
Outcomes
The primary outcome is the frequency of conversations regarding treatment options following graft loss as perceived by the patient. The main secondary outcome is to assess SDM using two different tools, the CollaboRATE mean score and the Control Preference Scale.
Follow-up
24 months post-randomisation.
CET Conclusions
This interesting protocol describes a single-centre study investigating the impact of use of an AI-based graft loss prediction tool in facilitating conversations about post-transplant renal replacement therapy options. Patients in routine post-transplant follow-up with an eGFR < 30 ml/min will be randomised to use of the graft survival prediction decision-support tool or standard care. Primary endpoint is the frequency of conversations about treatment options following graft loss. Secondary endpoints include clinical and qualitative assessments. The main issue with the study design is that the primary endpoint is very subjective and open to manipulation or biases. There is no blinding (this would be challenging) and so clinicians and patients will know the allocation, and this may influence their decision to discuss treatment options in both arms of the study. The additional awareness simply through participating in the study and reading the patient information materials for consent may influence the frequency of discussions, creating a baseline different to standard practice outside of the clinical trial setting. Measuring the impact of clinical decision support tools in clinical practice is challenging, but one would think that a less bias-prone endpoint, such as rates of pre-emptive relisting for transplant, may be more appropriate.
Trial registration
ClinicalTrials.gov - NCT06056518