Self-reported shift work, recall bias, and belief about disease causation in a case-control study of breast cancer

• Published in: Cancer epidemiology 2017-10, Vol.50 (Pt A), p.9-15
• Main Authors: Lizama, Natalia, Heyworth, Jane, Thomson, Allyson, Slevin, Terry, Fritschi, Lin
• Format: Article
• Language: English
• Subjects: Alcohol; Beliefs; Bias; Birth defects; Breast cancer; Breast neoplasms; Cancer; Case-control studies; Causality; Causation; Congenital diseases; Disease control; Employment; Epidemiology; Family medical history; Health risk assessment; Health risks; Interviews; Lifestyles; Melanoma; Pregnancy; Questionnaires; Recall; Shift work; Womens health; Working conditions
• ISSN: 1877-7821; 1877-783X
• DOI: 10.1016/j.canep.2017.07.007

•Cases who report working shifts are more likely to think shift work increases risk.•Belief about shift work does not affect controls’ reported shift work exposure.•Recalling shift work does not modify women’s beliefs about its effect on risk.•Recall bias does not appear to affect self-reported shift work exposure.•The combination of disease and exposure may modify beliefs about shift work. Recall bias is a potential source of misclassification in case-control studies. Studies have shown that the association between exposure and disease can differ according to participants’ beliefs or knowledge about the effect of that exposure on disease. We investigated the association between belief about breast cancer causation and self-reported shift work exposure in a case-control study. Women completed a questionnaire asking whether they believed that shift work caused cancer either before or after reporting their history of shift work. We measured: whether belief modified the association between reported shift work and disease; whether belief was associated with reported shift work exposure; and whether being prompted to recall shift work exposure was associated with an increased likelihood of believing that shift work increased breast cancer risk. There was a significant association between believing shift work increased breast cancer risk and reporting exposure to shift work. Being prompted to recall shift work was not associated with a belief that shift work increased risk. The association between pre-existing belief about breast cancer risk and reported shift work is likely to be due to exposed individuals believing that exposure increases risk, rather than resulting from recall bias.