Evaluation of two food frequency methods of measuring dietary calcium intake

To assess the ability of food frequency methods to measure current dietary calcium intake in elderly women, the authors administered two types of food frequency instruments to 37 randomly selected women who attended two senior citizens centers in San Francisco, and they compared those responses to seven-day food records. A 34-item food frequency instrument (with portion sizes rated as small, medium, or large) correlated well (r = 0.76) with the estimated calcium intake from seven-day records. Limiting the instrument to the top 15, 10, or five foods that contribute to dietary calcium intake had little effect on this correlation (r = 076, 0.75, and 0.67, respectively). Rating all portions as "medium" reduced these correlations somewhat. An 18-item instrument, which asks respondents to rate portion sizes in ounces or cups, did not correlate as well with the seven-day records (r = 0.49). The authors conclude that brief food frequency instruments which rate portion sizes on a simple qualitative scale may be suitable for many clinical uses and adequate for some types of epidemiologic studies of calcium intake in elderly women.