Validation of a brief diet survey instrument among medical students.

Our objective was to assess the reproducibility and accuracy of fat and of fruit and vegetable items on a 43-item food frequency questionnaire (FFQ) previously self-administered by students at 16 US medical schools. Five in-person, 24-hour recalls were administered between two FFQ administrations to 88 medical students. Reported fat intake decreased from the first (34.7%) to the second (33.1%) FFQ administration ( P <.001); the reproducibility correlation was r =0.63. Fat intake from recalls (28.4%) was lower than that from the FFQ (33.8%, P <.001). The Pearson correlation was r =0.36. Fruit and vegetable servings per day were 3.9 and 3.7 from the first and second FFQ, respectively ( P =.5); the reproducibility correlation was r =0.77. Fruit and vegetable servings were marginally higher from recalls (4.3) than from the FFQ (3.8, P =.06). The Pearson correlation for fruit and vegetable servings was r =0.50. This brief FFQ provides acceptably reproducible and valid estimates of fruit and vegetable servings per day among most groups of medical students, but overestimates fat as a percentage of energy intake.