Objective: The primary aim was to assess the association between response behavior and health status at baseline, and survival in a 5-year follow-up period. A secondary aim was to assess whether reasons for nonresponse were associated with health status at baseline.
Study design and setting: Data came from a prospective study cohort consisting of 31,349 men and women aged 45-70 years. Objective retrospective and prospective health information derived from general practitioner registries was available for both respondents and nonrespondents.
Results: Results show that among respondents coronary heart disease was more prevalent. Compared with respondents, noncontacts had a higher mortality risk during follow-up. Refusals had hypercholesterolemia more often than did noncontacts, and coronary heart disease or diabetes mellitus less often.
Conclusion: The paradoxical results that respondents are less healthy at baseline but prospectively have a lower mortality risk may point to a selection effect indicating that the 'worried ill' are more inclined to participate. This effect could imply that observed relationships between risk factors or behaviors and outcomes in cohort studies may be attenuated.