Vaillant · The American journal of psychiatry 2001 · Prospective longitudinal cohort study · n=569

Successful aging.

Cited 339 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective longitudinal cohort study with 60-year follow-up.

PubMed 11384887 · doi:10.1176/appi.ajp.158.6.839 · record verified 2026-08-29

What was done

Two cohorts of adolescent males—237 college students and 332 core-city youth (total n = 569)—were prospectively followed for 60 years or until death. Participants underwent physical examinations every 5 years and psychosocial evaluations every 2 years. Thirteen predictor variables assessed before age 50 were analyzed: six classified as uncontrollable (parental social class, family cohesion, major depression, ancestral longevity, childhood temperament, physical health at age 50) and seven with some personal control (alcohol abuse, smoking, marital stability, exercise, body mass index, coping mechanisms, education). Outcomes at age 70–80 evaluated successful aging across four objective metrics (physical health, mortality/disability before age 80, social supports, mental health) and two subjective metrics (instrumental activities of daily living, life enjoyment).

What was found

The abstract reports no numerical values, effect sizes, or confidence intervals. Multivariate analysis indicated that health and functioning at age 70–80 were predictable from variables measured before age 50. When the seven modifiable variables were controlled, major depression was the only uncontrollable midlife/childhood variable that continued to predict the quality of subjective and objective aging.

Why it matters

This longitudinal investigation suggests that late-life biopsychosocial health is heavily shaped by modifiable behaviors and psychosocial habits established before age 50 rather than fixed ancestral or childhood social determinants.

Limits

The abstract reports no numerical data, hazard ratios, or precision estimates. The sample is exclusively male and derived from specific historical cohorts (Harvard undergraduates and Boston inner-city youth), limiting generalizability to women and broader populations. Residual confounding and survival bias across a 60-year observation period are inherent design constraints.

Cited by