Individual vitality changes over 20 years among aging adults, using predefined and data-driven patterns: the Doetinchem Cohort Study.

'Vitality' refers to healthy aging beyond medical measures, but long-term data on vitality is scarce. Hence, we examined individual vitality patterns over 20 years, and their lifestyle and health characteristics, using predefined and data-driven methods. We used data of 3,013 adults aged 26-70 at their first vitality measurement from the Doetinchem Cohort Study examined up to 5 times from 1995 to 2019 at 5-year intervals. Vitality was based on the SF-36 vitality scale, with scores > 50 indicating good vitality. Long-term vitality patterns were defined by two approaches, predefined and data-driven, and compared using cross-tabulation and Cramér's V. The predefined patterns were: persistent good, persistent poor, worsening, improving, and varying vitality. The lifestyle and health characteristics related to these patterns were identified by multivariate multinomial logistic regression, using persistent good vitality as the reference at both baseline (t1) and 20-year follow-up (t5). Data-driven analysis revealed patterns similar to the predefined method with varying vitality split in two variants. There was fair agreement between the approaches (Cramér's V: 0.49). Most participants exhibited persistent good vitality: 61% in the predefined, and 78% in the data-driven approach. Compared to the persistent good vitality group, others were characterized by younger age and poor health, including pain, chronic conditions, and particularly high prevalence of poor mental health. Sex, education, household and work status, along with lifestyle factors, played a role in only some patterns. Most adults showed persistent good vitality over 20 years, with good mental health as a key characteristic of long-term vitality.