Differences between referred and nonreferred patients in cancer research.

BACKGROUND: In Canada, provincial cancer registries have been established to provide rigorous population-based data for patients with colorectal cancer. Databases maintained by regional cancer agencies contain a broader scope of information and have been used as a surrogate source of information for colorectal cancer research. It is unclear whether these data can be reliably extrapolated to all patients affected by colorectal cancer. We sought to determine whether patients included in a referral-based database are systematically different from patients who are not included.
METHODS: We conducted a retrospective cohort study to compare patients referred to the British Columbia Cancer Agency with those who were not referred. Comparison was based on age, sex and geographic location. We used univariate and logistic regression analysis to identify significant differences between the cohorts.
RESULTS: Univariate analysis demonstrated that the referral and nonreferral cohorts differed in sex, age and geographic location. For patients with rectal cancer, the referral and nonreferral cohorts varied in age and geographic location. Multivariate analysis demonstrated significant differences in age and geographic location but not sex for patients with colon and rectal cancer.
CONCLUSION: Patients included in the referral database differed in age and geographic location from those included only in the provincial database. Studies using large data sets from referral centres must be interpreted with caution and may not be representative of the entire patient population.