Reference no: EM133329371

Question: A group of investigators is discussing the design of a case-control study of pancreatic cancer in relation to past alcohol use. The only feasible approach to case ascertainment is through a cancer registry in that region; however, delays in registration, ranging from 12-18 months after diagnosis, would lead to a notable loss of cases because several patients with pancreatic cancer die within one year or so after their diagnosis. Therefore, the study team agrees that it would not be feasible to interview cases about their past alcohol use, and that they would need to rely on someone else, such as a partner, to ascertain information about cases' past alcohol use. The investigators also know from the prior literature that individuals often have incomplete knowledge of their partner's history and tend to under-report their partner's behaviors, such as alcohol use.

This under-reporting has been found, however, to operate to a similar extent in partners of cases and partners of controls in prior studies. There is disagreement within the study team, however, about involving partners for controls. Colleague A suggests that controls should be interviewed directly about their own past alcohol use (without involving their partners), "so that the best available information