Validation of a self-administered, computerized tool for collecting and displaying the family history of cancer

Abstract

Purpose: A detailed family history is important for cancer risk assessment, but obtaining it is time consuming and infrequently accomplished in practice. The Genetic Risk Easy Assessment Tool (GREAT) conducts a computer-administered family history interview and immediately generates a pedigree diagram in digital form. The purpose of this study was to validate family cancer histories produced by patients using the computer tool in comparison with pedigrees made by genetic counselors.

Methods: Patients scheduled for genetics consultation recorded their family histories using the GREAT, separate from their genetic counseling session. The presence of each relative; presence, type, and age at diagnosis of cancers; and cancer geneticist's risk assessment were compared for 120 pairs of pedigrees produced by counselors versus computer tool.

Results: The automated telephone interview took a mean of 33.5 minutes and was highly acceptable to respondents. Ninety-four percent of first-degree relatives, 67% of second-degree relatives, and 38% of third-degree relatives were identical on paired pedigrees; computer-generated pedigrees included additional relatives. Sixty-three percent of all cancers were identified by both family histories, with 90% agreement on the type of cancer. There was very good agreement (kappa = 0.70; correlation = 0.77) between the geneticist's breast cancer risk assessments based on computer versus counselors' pedigrees. In a subsample of 61 users, test-retest reliability for the computer-administered questionnaire was high (phi = 0.94 for cancers in first-degree and phi = 0.91 in second-degree relatives).

Conclusion: The GREAT computer-administered questionnaire provides an acceptable, reliable, and valid way of collecting an unverified but extensive family history of cancer and displaying it as a pedigree, in an entirely automated process.