BABBI: Biological Age Prediction based on Blood Biomarkers - A Comparative Study of Linear and Non-linear Predictive Models

The BABBI project focuses on improving how biological age (BA) is estimated using routinely collected blood biomarkers, moving beyond chronological age as a proxy for health and aging. Because individuals of the same chronological age can have very different health trajectories and mortality risks, BA has emerged as a more informative indicator of physiological aging. BABBI builds on established blood-based aging measures such as Klemera–Doubal Age and PhenoAge, while addressing their limitations by systematically comparing traditional linear models with more flexible non-linear and machine-learning approaches trained on large public datasets and applied to real-world data from Care Preventive AG (CARE), a Swiss preventive healthcare platform.

Methodologically, the project leverages longitudinal blood biomarker data, demographics, comorbidities, lifestyle factors, and health-coaching information from several hundred consenting CARE customers in Switzerland. Models are pre-trained on publicly available datasets (e.g., NHANES) and then evaluated on CARE data using standard performance metrics such as mean absolute error, root mean square error, and explained variance. Beyond pure prediction accuracy, BABBI includes extensive data exploration to understand how demographic characteristics, existing health conditions, and lifestyle behaviors influence BA estimates and their trajectories over time.

A key innovation of BABBI is its focus on intervention impact. By analyzing BA changes before and after CARE-recommended health coaching, the project assesses whether personalized lifestyle and behavioral interventions are associated with decelerated biological aging. This longitudinal perspective addresses a major gap in prior research, which has largely relied on cross-sectional data. Overall, BABBI aims to identify the most accurate and robust modeling strategies for BA estimation, inform personalized healthy-aging interventions, and provide actionable evidence for integrating biological age metrics into real-world preventive healthcare platforms.