Objective: This study examined the accuracy of a multiple-pass, 24-hour dietary recall method for estimating energy intakes of men and women by comparing it with energy intake required for weight maintenance.
Design: Three-day, multiple-pass, 24-hour recalls were obtained on randomly selected days during a self-selected diet period when subjects were preparing their own meals and during a controlled diet period when all meals were provided by the study. During the dietary intervention, weight was maintained; body weight and dietary intake were monitored closely, thereby allowing estimation of the energy intake required for weight maintenance.
Subjects/setting: Seventy-eight men and women (22 to 67 years old) from the Dietary Effects on Lipoprotein and Thrombogenic Activity (DELTA) study participated in this study. All 24-hour recalls were collected using a computer-assisted, interactive, multiple-pass telephone interview technique. Energy requirements for each individual were determined by the energy content of the DELTA study foods provided to maintain weight.
Statistical analysis: Paired and independent t tests were conducted to examine differences among study variables. Agreement between recalled energy intake and weight maintenance energy intake was analyzed using the Bland-Altman technique.
Results: Compared with weight maintenance energy intake, during the self-selected diet period men and women underestimated energy intake by 11% and 13%, respectively. During the controlled diet period, men underestimated energy intake by 13%, whereas women overestimated energy by 1.3%.
Applications/conclusions: Men had a tendency to under-estimate energy intake irrespective of the recording period. The accuracy of the recalled energy intake of women may be influenced by recording circumstances. Researchers should examine the factors influencing underreporting and overreporting by individuals and their impact on macronutrient and micronutrient intakes. Also, strategies need to be developed to minimize underreporting and overreporting.