This Article Full Text (PDF) References Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Add to My Marked Citations Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Agardh, E. E. Articles by Östenson, C.-G. Search for Related Content PubMed PubMed Citation Articles by Agardh, E. E. Articles by Östenson, C.-G. Social Bookmarking                   What's this?

The magnitude of bias in a cross-sectional study on lifestyle factors in relation to Type 2 diabetes

Department of Molecular Medicine, Endocrine and Diabetes Unit, Karolinska Institutet, emilie.agardh{at}phs.ki.se

Anders Ahlbom

Department of Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Department of Epidemiology, Stockholm Center of Public Health, Stockholm, Sweden

Tomas Andersson

Department of Epidemiology, Stockholm Center of Public Health, Stockholm, Sweden

Claes-Göran Östenson

Department of Molecular Medicine, Endocrine and Diabetes Unit, Karolinska Institutet

Aim: In cross-sectional studies it may be difficult to ascertain the temporal order of exposure and disease, which may have consequences for causal inference. The authors aimed to illustrate the possible magnitude of this potential bias using data from a cross-sectional study on coffee consumption and work stress in relation to type 2 diabetes. Methods: By a series of computer simulations the authors examined to what extent the observed negative association between type 2 diabetes and high coffee consumption and positive association between type 2 diabetes and high work stress could be due to reverse causality, by assuming that cases changed their exposures in response to development of the disease. Results: If the negative association between coffee and type 2 diabetes was a consequence of reversed causality, 30—40% of the cases would have to decrease their coffee consumption from5 cups of coffee per day to 3—4 cups per day and from 3—4 cups per day to2 cups of coffee per day. Moreover, approximately 60% of the cases would have to increase their work stress from low to medium work stress and from medium to high work stress, in order to produce the positive association with diabetes that was observed. Conclusion: Even if the type 2 diabetic patients to some extent may have changed their exposure in response to disease development, it seems unlikely that the associations observed between type 2 diabetes, coffee consumption, and work stress are due to this bias.

Key Words: epidemiology • cross-sectional studies • causal inference • type 2 diabetes • lifestyle

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?