TL;DR
This study developed a mobile app to deliver personalized polygenic risk scores (PRS) for coronary artery disease (CAD) and observed its impact on health behaviors. They found that communicating a high CAD PRS was associated with increased and earlier initiation of lipid-lowering therapy, suggesting PRS can be a valuable tool for reinforcing preventive health guidelines.
What Problem Does This Paper Address?
Despite established guidelines for preventing coronary artery disease (CAD), a significant portion of individuals who would benefit from lipid-lowering therapy do not adhere to them. This gap exists even among those with high genetic predisposition. Previous attempts to influence behavior with genetic risk information have had limited success, often due to broad risk assessments or lack of actionable insights. This paper addresses the challenge of effectively communicating personalized genetic risk for CAD and its impact on real-world health decisions, specifically the initiation of lipid-lowering therapy, using a novel mobile application-based approach.
Methods
The researchers developed a smartphone application called "MyGeneRank" to conduct a prospective observational cohort study. This app facilitated the automated generation, communication, and electronic capture of participant responses to their personalized CAD polygenic risk score (PRS). Participants, who already had genetic profiling from 23andMe, self-referred to the study. The app provided two interfaces for risk communication: one integrating the CAD PRS with the 10-year ASCVD Pooled Cohort Equations (PCE) derived risk, and another showing a percentile rank relative to a simulated population. Participants completed baseline and follow-up surveys electronically to report their actions and intentions. Statistical analysis involved comparing responses across different CAD PRS tiers (low, average, high) using one-tailed two-proportion z-tests and Mann-Whitney U tests.
Datasets
The study utilized data from 3,800 participants who self-referred to the MyGeneRank study. These participants already possessed genetic profiling data from 23andMe. The study did not introduce a new benchmark dataset but rather leveraged existing genetic data and self-reported health information collected through the mobile application.
Key Insights and Ideas
- PRS as a "Risk Enhancer": The study reinforces the idea that a CAD PRS can act as a "risk enhancer," prompting individuals to consider therapeutic interventions beyond traditional risk factors.
- Digital Communication Efficacy: A fully digitized, app-based approach can effectively communicate complex genetic risk information and influence health behaviors.
- Actionability of PRS: Participants perceived the CAD PRS information as understandable and actionable, without inducing significant health anxiety.
- Reinforcement of Guidelines: The findings suggest that CAD PRS can be a valuable tool for reinforcing existing clinical guidelines, particularly for individuals who might otherwise not meet the threshold for intervention.
- "Worried Well" vs. General Population: The study acknowledges that early adopters of such technology might be more health-conscious, potentially influencing the observed signal.
Key Results
- Lipid-Lowering Therapy Initiation: 19% of participants provided complete responses. At follow-up, 20% of high CAD PRS individuals initiated lipid-lowering therapy compared to 7.9% of low CAD PRS individuals (p=0.002).
- Statin and Non-Statin Initiation: High CAD PRS individuals were more likely to initiate statin therapy (15.2% vs. 6.0%, p=0.018) and non-statin lipid-lowering therapy (6.8% vs. 1.6%, p=0.022).
- Earlier Initiation: High CAD PRS individuals initiated lipid-lowering therapy at an earlier average age (52 years) compared to low CAD PRS individuals (65 years, p=0.007).
- Overall Lipid-Lowering Use: 42.4% of high CAD PRS individuals used any lipid-lowering therapy at follow-up, compared to 28.5% of low CAD PRS individuals (p=0.009).
- Participant Intentions: Participants with high CAD PRS were more likely to intend to meet with a physician (60.3% vs. 33.3%) and make changes to their statin use (19.0% vs. 5%) after receiving their PRS.
- Risk Perception: High CAD PRS individuals were more likely to agree with statements like "My chances of developing Coronary Artery Disease are high" and "My genetics make it more likely that I will get Coronary Artery Disease."
Limitations and Future Work
- Loss to Follow-up: The primary limitation acknowledged is the loss to follow-up, with only 18.7% of participants completing the follow-up survey. While the study argues this rate is higher than some digital health studies and that respondents were balanced across PRS tiers, it remains a potential source of bias.
- Self-Reported Outcomes: The study relies on self-reported data for lipid-lowering therapy use, which may not always be accurate. Future work should incorporate objective measures like Electronic Health Records (EHRs).
- Convenience Sample: The study population consisted of self-referred "early adopters" of genetic testing, predominantly of European ancestry, which may limit generalizability to the broader population.
- "Worried Well" Effect: The enrichment of non-smokers and health-conscious individuals in the study suggests a potential "worried well" effect that might amplify the observed behavioral changes.
- Future Work: The authors suggest the need for alternative communication routes, longer-term studies with EHR-based outcomes to assess durability and adherence, and studies in more diverse populations. They also propose exploring how PRS can influence intensification of therapy, adherence, and ultimately hard cardiovascular outcomes.
Summary updated