Relative validation of a pre-coded food diary in a group of Norwegian adults – Comparison of underreporters and acceptable reporters

Estimating dietary intake is important for both epidemiological and clinical studies. In large studies, a balance has to be achieved between methods with high accuracy and methods that are easy to use. The aim of the present study was to compare results from a pre-coded scanable food diary (PFD) with results from a weighed record (WR) in a group of Norwegian adults. We also explored differences in day-to-day energy intake and the distribution of energy intake across the day in acceptable reporters (ARs) and underreporters (URs). Participants (n = 114, mean age 35 years, 68% women) recorded dietary intake with the PFD for 7 consecutive days. One week after completing the PFD, participants completed a 7 days WR. No difference in mean energy intake was seen between methods. Few differences were seen for the macronutrients, the most noticeable difference being the percentage of energy (E%) from carbohydrates which was significantly lower with the PFD (47 E%) than with the WR (49 E%). For the micronutrients, intakes of calcium and vitamin A were both significantly higher with the PFD than with the WR. Pearson’s correlation coefficient ranged from 0.47 (tocopherol) to 0.76 (E% carbohydrates) for all nutrients. Bread intake was significantly lower with the PFD while the intakes of edible fats, cheese and beverages were higher. Twenty-eight percent of the participants were found to be URs with the PFD. No clear pattern of underreporting at certain recording days or times of the day was seen. In conclusion, the results showed similar energy intakes and few differences in food and nutrient intakes between the PDF and the WR at the group level. Somewhat larger differences between the methods were seen at the individual level. Because of the reduced work load on both participants and researchers, the PFD seems a suitable alternative to the WR.