Comparing sampling strategies to recruit migrants for an epidemiological study. Results from a German feasibility study.

Background: In 2011, almost 20.0% of the population of Germany had a migration background. Studies on their health tend to have low participation rates. The aim of our study was to compare different sampling strategies and to test different approaches to recruit migrants for an epidemiological study.

Methods: Four recruitment centres of the German National Cohort recruited persons of Turkish origin and ethnic German immigrants from former Soviet Union countries. A register-based (random samples from residents' registration offices) and a community-orientated strategy were applied. Participants underwent a medical examination and self-completed a questionnaire.

Results: Used approaches: The community-orientated strategies comprised the acquisition of key persons from migrant networks to support the recruitment, invitation talks and distribution of study materials in migrant settings, etc. The identifying variables in the registry data were name, nationality or country of birth. All but one centres used bilingual study material and study staff.

Participation: When comparing the two strategies, the register-based participation rates ranged from 10.1 to 21.0% (n = 668 participants) and the community-oriented recruitment resulted in 722 participants.

Conclusion: Register-based recruitment should use a combination of name, nationality and country of birth in order not to be limited to identifying persons with a foreign nationality. However, according to the study staff, the community-oriented approach involving key persons of the same cultural background leads to a better acceptance by the participants. Also, it covers a more heterogeneous group. Yet, it is time-consuming and needs considerably more staff. Further research should establish the effectiveness of a combination of both strategies.