RESPONSE RATES WITH DIFFERENT DISTRIBUTION METHODS AND REWARD, AND REPRODUCIBILITY OF A QUANTITATIVE FOOD FREQUENCY QUESTIONNAIRE

Objectives: To evaluate the use of a self-administered quantitative fo od frequency questionnaire (QFFQ) ina national dietary survey concerni ng (a) response rates with different distribution methods and reward; (b) degree of underreporting of energy intake; (c) reproducibility of the QFFQ; and (d) seasonal variation on reported intake. Design and su bjects: A pilot study was performed in 1992 to test response rates to the QFFQ with three different distribution methods, with and without r eward, in a random sample of 1200 adults aged 16-79 y. Zn another stud y, the QFFQ was distributed to a nation-wide, representative random sa mple of 5008 adults aged 16-79 y duping June, September, November 1993 and March 1994. Reproducibility was evaluated among 90 responders to the survey who answered another QFFQ six weeks later. Results: The dis tribution method combining postal distribution and collecting the QFFQ by interviewer as well as an offer to participate in a lottery, gave the highest response rate (72%). The possibility to get a reward incre ased the response rate by 9, 14 and 57%, respectively, depending on th e distribution method used. The mean daily energy intake and the perce ntage of subjects claiming to have unlikely low energy intake did not differ significantly between the different ways of distribution In the main survey the mean ratio between energy intake and estimated basal metabolic rate was 1.58 among men and 1.47 among women, and 37% of men and 45% of women had a ratio below 1.35. Spearman rank correlations b etween the two QFFQ ranged from 0.48 (edible fats) to 0.91 (coffee) wi th a median coefficient of 0.70. For nutrients correlations ranged fro m 0.55 (carbohydrate E%) to 0.81 (alcohol), with a median coefficient of 0.72. The season of questionnaire administration was of minor impor tance for the reported intake of the main foods and nutrients. Conclus ions: The QFFQ-method is suitable for use in a Norwegian nutritional s urveillance system.