For observational studies derived from Internet based cohorts, the representativeness of demographic, behavioral and health characteristics of subjects is not well-established in European countries.
OBJECTIVES: To compare distributions of self-reported characteristics of subjects recruited via Internet in Germany & UK to a national representative sample in each country.
METHODS: PROCEED is a multinational observational cohort of overweight & obese subjects (body mass index >25 kg/m2) recruited through an existing Internet panel in Germany & the UK in 2005. Eligibility criteria were: age 35–75; not pregnant; willing to lose weight in the next year and weight < 180 kg. Recruitment was stratified to balance gender and overweight and obese categories. Baseline demographics and selected health and behavior characteristics of the PROCEED cohort were compared with estimates from a relevant subset (same age, BMI and not pregnant) of two National Surveys (1998 GNHIS and 2003 HSE). PROCEED data were standardized for gender and BMI category in each of the national surveys.
RESULTS: PROCEED subjects in Germany (n = 203) and the UK (n = 216) presented similar characteristics to each national survey population in terms of level of alcohol consumption, and prevalence of hypertension, high cholesterol and diabetes. More PROCEED subjects reported having college or higher education (22% versus 12% in Germany; 31% versus 15% in the UK). PROCEED subjects were also more likely to be single. The German PROCEED cohort had a higher proportion of current smokers compared to GNHIS data (48% versus 24%) while the UK PROCEED cohort was very similar to HSE data.
CONCLUSIONS: Despite few differences in education level, marital and smoking status, most demographic and health characteristics were similar between the Internet cohort of overweight/obese subjects and the German and UK national surveys.The internet seems to be an appropriate tool for recruiting subjects in observational studies.
