Ecological Momentary Genomic Phenotyping: A Dynamic Stage-Matched Risk Communication Framework Personalized Heart Failure Education

We implemented a multi-ancestry polygenic risk score, perhaps augmented with actionable pharmacogenomic information specific to a patient’s medications, such as the bleeding risk associated with warfarin based on the CYP2C9 genotype of the patient. Rather than simply disclosing an individual’s genetic risk once, we plan to provide real-time education (e.g., information about what genetic risk means for client health and options) matched to the patient’s current stage of change (as defined in the TTM). This latter approach allows us to use objective information (genomic risk) as an ecological momentary assessment input in the patient’s portal, which the client will complete six times per day over 6 months. The patient’s worry about genetic risk, behavioral engagement, and momentary self-efficacy can easily be assessed in this manner. Additionally, the responses can be scored in real time across the five stages of change (precontemplation, contemplation, preparation, action, and maintenance) defined in the TTM using ordinal logistic regression. Moreover, because the stochastic Markov model allows movement from state to state over time, in this case, transitions from precontemplation to contemplation, we can include the contribution of genomic risk to the transition probabilities to provide yet another dimension of individualized “real-time” educational feedback. The proposed intervention consists of an interaction of predictions from the closed-loop system regarding a patient’s readiness stage from a stage estimate described according to the Prochaska and DiClemente paradigm and enrichment of the estimates with EMA reflecting the patient’s genomic risk information in terms of polygenic risk scores (PRS). Plain-language educational messages associated with each stage can be varied according to the patient’s current stage and behavioral change-related self-efficacy. If a patient’s self-efficacy is judged to be low, for example, presenting educational material that is easier to read. Simulation results show that the educational message can be tailored to the fluctuation in the patient’s self-efficacy so that it remains relevant in the continuum of care. Thus, the closed-loop design is based on a pedagogical approach to genomic risk for the avoidance of information and then to the facilitation of self-management for HF.