Redefining risk assessment in congenital heart disease

As more patients with congenital heart disease (CHD) survive into adulthood (1), there is an increasing need to understand lifelong and dynamic risks associated with living with CHD to guide clinical care and improve patient outcomes. The JACC Scientific Statement on Clinical Risk Assessment and Prediction in Congenital Heart Disease Across the Lifespan presents a comprehensive overview of challenges in current risk prediction models in CHD, fundamental and emerging concepts in risk measurement and the current state of risk measurement in CHD, identifying opportunities for assessment and incorporating a longitudinal perspective for risk evaluation and management.

Opotowsky et al (2) focus on the necessity of development of comprehensive, broadly applicable, robust, and validated risk models in CHD, as current risk prediction models are restricted by retrospective data, wide variation in CHD anatomy and surgical approaches, sparse patient numbers and clinical outcomes, which limit generalizability to broader CHD populations. Clinical risk models in CHD focus on short to medium term outcomes in the context a specific event (ex: pregnancy, surgery, cardiac catheterization), and incorporate some anatomical and surgical diversity, but have little application to extend beyond follow up periods and don’t always incorporate patient reported outcomes. Therefore, there remain no robust models to guide long-term preventive efforts or models that focus on maintaining long-term health rather than avoiding adverse outcomes in the context of episodically increased risk. 

To adopt a broader perspective, risk prediction models in CHD should consider a lifelong perspective and incorporate risk factors with persistent long-term influences as risk modifiers and identify pivotal moments of intervention to alter each person’s disease trajectory. Use of emerging technologies in machine learning and artificial intelligence can potentially standardize data collection across institutions, developing increasingly multidimensional data sources and multi-institutional data sharing with complex data analysis that accounts for consensus-driven methodologies to provide insight into time-varying risk. Psychosocial factors, including mental health, social determinants of health, socioeconomic environment, racism, adverse childhood events, and social support, also influence the course of disease and must be integrated into models of risk. To improve generalizability, rigorous prospective validation studies and incorporating patient reported outcomes would be needed to capture a more complete spectrum of health-related quality of life in patients with CHD. 

Finally, another crucial point that this paper highlights in addition to improving patient outcomes, is to guide and counsel patients and families of the “unknown”. Effective communication with patients and caregivers of CHD-related risks requires establishing trust with patients, fostering empathy, and conveying group-based estimates along with personalized risk assessment to empower patients and caregivers in collaborative decision making.

This paper is novel in its forward-looking approach towards risk assessment in CHD by advocating for large representative cohorts, longitudinal and multidimensional data collection and use of artificial intelligence for data processing and the development of comprehensive risk models considering the complexities of CHD care. However, it is not clear how these risk models can be built. Another way to incorporate real world data is patient engagement in research allowing for real world data collection at a large scale to ensure the entire CHD population is represented to facilitate future research. (3) 

In conclusion, the single most important takeaway is that risk assessment in CHD requires a longitudinal, holistic, multidimensional and personalized lens, which incorporate not only the anatomical and surgical heterogeneity, but also risk factors with persistent long term influence and socio-economic factors that can impact patients’ access to healthcare and outcomes over time. Identifying high risk patients for preventable or modifiable cardiac risk and non-cardiac risk is crucial to identify opportunities for intervention and potentially change the trajectory of long term outcome. By incorporating emerging technologies and integrating psychosocial determinants, clinicians can better predict adverse outcomes and tailor interventions accordingly.