This study will evaluate the structure of error in self-reporting instruments on diet, in following up the Observing Protein and Energy Nutrition Study (OPEN) conducted in 1999 and 2000. The OPEN study, the largest of its kind, resulted in a wealth of nutritional biomarker data-measurable indicators of changes in organisms at the level of molecules or cells. A biomarker could help in understanding how environment and disease are related and about disease risks. Furthermore, its findings showed that the relative estimates of disease risk may be weakened by a food frequency questionnaire and a 24-hour questionnaire. Scientists have long recognized that self-reported information from questionnaires and interviews contains errors, but there has been uncertainty about the structure of such errors. For the proposed study, a food record and a food checklist-adjusted food frequency questionnaire will be compared with the previously used questionnaires, to determine whether the newer study instruments present less chance of measurement error. The results may provide better tools for a study on nutritional epidemiology, that is, the incidence and other aspects of disease and nutrition. OPEN study participants will be invited for follow-up.

Patients who completed the 1999-2000 study who are not currently pregnant or a liquid weight loss diet may be eligible for this study. The original study recruited men and women 40 to 69 years of age. A letter with summary results from the original OPEN study will be sent to each of the 482 participants.

Both the Food Record and the food checklist, called the Daily Food List, are instruments to be given. The Food Record asks the participant to record all foods and beverages consumed. The Daily Food List asks respondents to write in to the number of times a limited number of food categories are consumed. In addition, respondents will be asked to complete the National Cancer Institute Diet History Questionnaire (DHQ), identical to the one that they completed in the original OPEN study, regarding the previous 12 months. Participants will randomly be assigned to one of two groups. All will receive four study mailings over a 3-month period: the DHQ, First 4-Day Food Record (4-day FR), Second 4-day FR, and Booklet of 7-Day Food Lists (7-day FL), the latter listing selected foods that the participants will mark if consumed on the reporting day. Study Group A will receive the DHQ, first 4-day FR, second 4-day FR, and 7-day FL. Study Group B will receive the DHQ, 7-day FL, first 4-day FR, and second 4-day FR. Within 10 days after receiving the DHQ, participants will be called to schedule an appointment for a 30- to 45-minute visit. Participants will undergo the following procedures:

• Complete the Physical Activity Questionnaire (by telephone for those who have moved out of the local area or are unable to attend the visit)
• Have weight measured
• Be asked about whether he or she has recently developed certain medical conditions, experienced dramatic changes in diet, or gained special education in nutrition

The completion time for each instrument in dietary assessment is approximately as follows:

• DHQ-60 minutes.
• Each 4-day FR-80 minutes (20 minutes a day).
• 7-day FL-35 to 49 minutes (5 to 7 minutes a day).

NCI carries out extensive research and surveillance activities that rely on self-reported dietary information from questionnaires and interviews. Although scientists have long recognized that this information contains error, considerable uncertainty remains about its structure. The interpretation of findings from NCI's nutritional epidemiology and surveillance efforts is critically dependent on the assessment of dietary measurement error. We evaluated such measurement error in the previously conducted Observing Protein and Energy Nutrition Study (OPEN) conducted in 1999 and 2000. This study, the largest study of its kind, used unbiased biomarkers of dietary intake (doubly labeled water for energy and 24-hour urinary nitrogen for protein) to assess measurement error in dietary assessment methods commonly used in epidemiology and surveillance research around the world. The landmark findings indicate that relative risk estimates in nutritional epidemiology may be severely attenuated by the use of food frequency questionnaires and that the conventional use of 24-hour dietary recalls as reference measurements may lead to severe underestimates of the extent of this attenuation. Key questions remain regarding whether these findings hold in diverse populations, for other dietary-report or physical activity instruments, across varying nutrients and food groups, and how measurement error affects relative risks in existing prospective cohorts. In light of the OPEN study findings, we would like to re-contact participants, for whom we already have a wealth of nutritional biomarker data, in order to assess the structure of measurement error with respect to two other self-reported dietary assessment instruments: the food record and a food checklist-adjusted food frequency questionnaire. Our intent is to evaluate whether these instruments might have more favorable measurement error structure than the food frequency questionnaire, thereby providing better tools for nutritional epidemiology.

• INCLUSION CRITERIA:

As this is a follow-up to a previous study with NCI SSIRB approval, there will be no recruitment of new participants. In the original OPEN study, we recruited a sample of men and women between 40-69 years of age. This age group represents those most likely to be recruited for large-scale epidemiological studies of diet and cancer. It is in this same group that we seek to assess measurement error problems for other dietary assessment instruments.

Volunteers.