Follow-up procedures in EPIC-Germany--data quality aspects. European Prospective Investigation into Cancer and Nutrition

Abstract

With 475,000 participants throughout Europe, EPIC is one of the largest cohort studies investigating the association between diet and cancer and other chronic diseases. The German part of EPIC comprises about 53,000 participants in Potsdam (n = 27,616) and in Heidelberg (n = 25,546). In the German study centers, follow-up started in 1998 and will be continued in 2-year intervals over the next 10-15 years. To ensure high follow-up data quality at an European level, an international working group developed guidelines for endpoint data collection in every country. A follow-up phase in Germany comprises mailing of a questionnaire, tracing of individuals to whom mail could not be delivered, obtaining information on deceased participants including cause of death, and verifying self- reported diagnoses. Furthermore, activities aimed at motivating study participants are part of the follow-up. The first round of follow-up of those who entered the study in 1994 and 1995 included 8, 706 participants in Potsdam and 6,289 in Heidelberg. Due to a comprehensive and intensive reminder and tracing system, vital status of the study subjects is known from almost 100% in Potsdam and 99% in Heidelberg. Two years after baseline examination, and with twice as many addresses in Potsdam as in Heidelberg, addresses had to be traced or checked via population registry (13 versus 6%). Tracing, the application of different mailing strategies, and intensive reminder activities resulted in a 95% return of the questionnaire in Potsdam and 90% in Heidelberg. The system of follow-up data entry and control, including completion of missing information via telephone, verification of self-reports and causes of death, has been set up for EPIC-Germany and works efficiently and successfully. The aim of this paper is to describe the follow-up procedures in EPIC-Germany with a focus on the generation of valid and complete outcome data.