The diet of study participants was assessed by different instruments that had been developed and validated previously in the local context. Following the results of the methodological studies, three dietary assessment methods were adopted:
1. Quantitative dietary questionnaires, containing up to 260 food items and estimating individual average portions systematically, were used in the Netherlands, and Germany (both self-reported) and Greece (where dietary questionnaires were interviewer-administered). Questionnaires similar to the quantitative dietary questionnaires, although structured by meals, were used in Spain, northern Italy, France, and Ragusa (southern Italy). To increase compliance, the centres in Spain and Ragusa performed a face-to-face dietary interview using a computerized dietary program, whereas the dietary questionnaire was self-reported in France and northern Italy.
2. Semi-quantitative food-frequency questionnaires (with the same standard portion(s) assigned to all participants) were used in Denmark, Norway, Naples in Italy, and Umeå in Sweden.
3. Combined dietary methods were used in the United Kingdom and Malmö (Sweden). The two British centres used both a semi-quantitative food-frequency questionnaire and a 7-day record, whereas a method combining a short semi-quantitative food-frequency questionnaire with a 14-day record on hot meals (lunches and dinners) was developed in Malmö.
The EPIC study aims to increase the overall dietary variation in the investigation of diet–disease relationships by combining study populations characterized by different types of lifestyle and dietary habits and different cancer incidence rates. For this, dietary assessments in each of the subcohorts need to be comparable. To overcome the fact that the use of different dietary assessment methods across the 23 EPIC centres introduces comparability issues, additional dietary measurements were collected through a computer-assisted 24-hour dietary recall instrument (EPIC-SOFT) in representative subsamples of 8% of study participants in each subcohort. In total, 36 900 24-hour recalls were collected from EPIC participants. Baseline dietary assessments conducted on all EPIC participants, used locally to estimate long-term usual dietary intake, were used to rank participants within centres, while the 24-hour dietary recall is used as a reference method to correct for random and systematic misclassification in baseline dietary assessments.
A common food composition database for a number of nutrients, standardized across the European countries involved in EPIC, was developed.