Unraveling aging and health trajectories through Big Data

The global increase in life expectancy is leading to a growing population of old adults worldwide. In consequence, disability and multimorbidity rates are also increasing, which constitutes a challenge for the healthcare system. However, the aging process presents high heterogeneity, and understanding the underlying relationship between healthspan and lifespan is essential for identifying distinct patterns in human aging. Electronic Health Records (EHRs) offer a unique opportunity to study aging dynamically, but their longitudinal complexity remains underexploited. In this work, we used machine learning techniques applied to EHR data from the Basque Country old individuals to explore aging trajectories. We identified sex-specific disease trajectories and their relationship with longevity, highlighting the heterogeneous and individual nature of aging. We also examined centenarians as a model of successful aging and confirmed that they fit the compression of morbidity theory, showing resilience against major age-associated conditions such as cancer, neurodegenerative and cardiovascular diseases, and COVID-19. Using a multivariate survival approach, we derived a Resilience Index that quantifies individual aging trajectories over time and reveals distinct aging patterns in men and women. Overall, our work provides an integrative, data-driven framework to study human aging and disease.