Background
Urban-rural comparisons are of limited relevance in examining the effects of urban migration in developing countries where urbanisation is due to growth of existing urban populations, expansion of urban boundaries, and rural in-migration. Cultural, genetic and life-style backgrounds of migrants and host populations further limit the value of rural-urban comparisons. Therefore we evaluated a sib-comparison design intended to overcome the limitations of urban-rural comparisons.

Methods
Using the framework of a current cardiovascular risk factor screening study conducted in Indian factories, we recruited the non-migrant rural sibs of migrant urban factory workers and the urban sibs of non-migrant factory workers. The response rate, completed interviews and examinations conducted were assessed. Adequacy of generic food frequency questionnaires and WHO quality of life questionnaire were assessed.

Results
All the urban factory workers and spouses approached agreed to be interviewed. Of the 697 participants interviewed, 293 (42%) had at least one rural dwelling sibling. Twenty (22%) siblings lived further than 100 km from the study site. An additional 21 urban siblings of non-migrant factory workers were also investigated to test the logistics of this element of the study. Obesity (BMI >25 kg/m2) was more common in rural sibs than urban factory workers (age adjusted prevalence: 21.1% (17.1 to 25.0) vs. 16.1% (11.9, 20.3). Diabetes prevalence (fasting plasma glucose greater than 126 mg/dl) was higher than expected (age-adjusted prevalence: 12.5% (22 out of 93) in urban migrants and 4.5% (8 out of 90) in rural non-migrant sibs.

Conclusion
The sib-comparison design is robust and has been adopted in the main study. It is possible that simple urban-rural study designs under-estimate the true differences in diabetes risk between migrants and non-migrants.