Developing a scoring method for evaluating dietary methodology in reviews of epidemiologic studies

We examined the quality of dietary assessment used by studies of prostate cancer and dietary fat in an attempt to explain the heterogeneity of their relative risk (RR) estimates. We reviewed the dietary assessment of 39 studies published in English that reported RRs for the association between prostate cancer and dietary fat intake derived from food frequency questionnaires (FFQs). We scored studies based on several objective measures of quality dietary assessment. Studies received no points for characteristics with unclear information. Studies scored 2 points for interviewer-completed FFQs, along with 2 points for quantitative assessments. They were scored 4 points for FFQs with more than 150 items, with an additional point for pretesting and 2 points for validated FFQs. Studies were given 1 point for describing each of the following characteristics: specifying the nutrient database used to convert foods to grams of fat, specifying quality control, attempting to measure dietary intake prior to diagnosis (recalled dietary period), and reporting the time needed to complete the FFQ. We then ranked studies based on their overall score: "high" for a score of 7 or greater out of 15 and "low" for lower scores. Two of the 39 studies that used quantitative methods other than a FFQ were excluded. Of the remaining 37 studies reviewed that used FFQs, only 16 were judged to have a high quality assessment of dietary fat. This review highlights the inconsistency of FFQ used in epidemiologic studies of dietary fat. Such variations in dietary measurement may be reflected in the variation in the magnitude of RRs reported for prostate cancer and dietary fat. The problems identified here include insufficient reporting of the details of dietary assessment, in addition to use of questionnaires with only a few food items to estimate a subject's dietary fat intake. It is imperative that journals include experts in the field of nutrition as reviewers of epidemiologic papers describing diet.