Development and Validation of an Incident Hypertension Risk Prediction Model for Young Adults

Area of Science:

• Cardiovascular disease research
• Public health
• Predictive modeling in medicine

Background:

• Hypertension risk identification in young adults is crucial for timely intervention.
• Current blood pressure screening guidelines may not optimally identify high-risk individuals.
• Early detection of hypertension risk can inform targeted screening strategies.

Purpose of the Study:

• To develop and validate predictive models for incident hypertension in young adults (18-39 years).
• To compare the performance of novel prediction models against existing US Preventive Services Task Force (USPSTF) recommendations.
• To assess the efficiency of a new model in identifying high-risk young adults for early intervention.

Main Methods:

• Developed sex-specific hypertension prediction models using Cox elastic net and random survival forests.
• Utilized diverse contemporary cohorts including Kaiser Permanente Southern California and three external cohort studies (CARDIA, FHS, HCHS/SOL).
• Included predictors such as age, blood pressure, BMI, smoking, social determinants of health, comorbidities, lipids, and pregnancy-related hypertensive disorders.

Main Results:

• The developed model demonstrated good discrimination and calibration in external validation cohorts.
• The new model showed similar sensitivity but improved specificity compared to USPSTF recommendations for identifying high-risk individuals.
• The model estimated 28.5 million US young adults at high risk, versus 45.3 million by USPSTF, suggesting more efficient identification.

Conclusions:

• The novel hypertension risk prediction model offers a more efficient approach to identifying high-risk young adults.
• This tool can potentially enhance early intervention strategies for hypertension prevention.
• The findings support the refinement of blood pressure screening recommendations for young adults.