Predictors of self-referral into a cancer genetics registry

Abstract

Recruitment is a challenging part of developing and maintaining any population-based disease registry and different methods are used to increase enrollment. However, recruitment methods may attract different subgroups of individuals, so examining characteristics of samples recruited using different methods can help detect threats to the external validity of research results.

Objectives: There were two main objectives of this study: to examine differences between participants who were self-referred and those who were recruited into a cancer genetics registry and to identify predictors of self-referral status.

Methods: A cross-sectional survey was done with two groups of cancer genetics registry members (n = 268): (a) members recruited through self-referral and (b) members recruited through population-based sampling.

Results: There were no significant differences in demographic variables between the two samples except for education (higher in the self-referral group; P < 0.01). The self-referral group showed significantly higher levels of anxiety, depression, and cancer history and was more likely to report the strongest response to statements about cancer risk, screening intentions, and views on genetic testing. Logistic regression modeling indicated these predictors of self-referral status: previous cancer diagnosis, viewing self as a candidate for genetic testing, education higher than high school, and wanting assistance with personal future risk (R2 = 0.41).

Conclusions: Our results suggest that whereas groups recruited via different strategies may seem to be the same based on demographic variables, underlying psychosocial variables differ between those self-referring and those recruited via population-based screening. To accurately estimate the generalizability of population-based studies or studies conducted using a cancer genetics registry, method of recruitment should be examined when interpreting and analyzing results.