Recruitment of participants to a familial cancer study: Balancing privacy protection and research quality

Inclusion of relatives of individuals at risk for or diagnosed with disease is critical for many public health studies. Debate exists about how to recruit relatives in a way that both protects their privacy and allows for reasonable participation rates. To inform the debate, empirical data is needed.

This research evaluates an investigator-driven recruitment strategy enabling index subjects to first discuss the study with relatives. We assessed whether index subjects' willingness to permit investigators to contact relatives was associated with their own and their relatives' demographic characteristics.

In 2001, we conducted a random digit dial phone survey in Connecticut to determine the accuracy of family cancer history reported by 1,380 index subjects (response rate=70%). After obtaining a family roster of 1st and 2nd degree relatives, a random sample was invited to participate.

The majority of index subjects (65.1%) contacted at least one relative to discuss the study, although non-whites were less likely to do so than whites (OR=0.52 [0.32-0.84]). Furthermore, 67.4% gave investigators permission to contact some or all sampled relatives for an interview. After adjustment for educational level, index subjects were more likely to give investigators permission to contact relatives with a history of cancer (OR=1.19 [1.08-3.37]), and those with whom they had a supportive relationship (OR=5.13 [2.66-9.92]). Overall, 90% of the 717 relatives whom the investigators were given permission to contact agreed to participate.

The recruitment strategy we examined appears effective for population based research, as high proportions of index subjects were willing to give investigators contact permission.