A study of the reproducibility of long-term recall in the elderly.

We assessed reproducibility of long-term recall (over periods of up to 82 years) by repeat interviews 1-3 months apart in 63 people age 65-95 years. We assessed weight, height, physical activity, smoking, reproductive factors, and consumption of alcohol, coffee, tea, milk, cheese, and fruit. Reproducibility was acceptable for most variables. Spearman rank correlation coefficients ranged from 0.95 to 0.36; kappas ranged from 0.97 to 0.35. Past physical activity was the least reproducible variable. Older age, cognitive impairment, and male gender were associated with reduced reproducibility.