Adequacy of self-reported family history in electronic health record for genetic risk assessment for Lynch syndrome.

Abstract

1515 Background: Self-reported family cancer history (FCH) is one of the key indicators of hereditary cancer risk. Studies have shown that accurate FCH documentation by healthcare providers is suboptimal, but data regarding patient-provided FCH are limited. We evaluated the quality of FCH as entered by the patient into the electronic health record (EHR) to determine its adequacy for Lynch syndrome (LS) risk assessment. Methods: At our tertiary referral cancer center, FCH is self-reported via an online questionnaire sent prior to appointment, which is reviewed/updated by clinic nurse during initial visit and then imported into EHR review of systems. Records of all new patients from September 2016 to August 2017 were retrospectively reviewed and analyzed. FCH quality was estimated by calculating rates of reporting of 3 FCH variables required for PREMM5, a risk-prediction model for LS. Parameters required for the model were sex, age, personal history of cancer, and for FCH, degree of kinship (first/second degree), cancer site/type, and age at diagnosis. Results: Of 47,647 unique patients, 47.5% reported FCH for 1 or more relative (46.1% were first degree, 64.8% second degree, 3.0% other, and 2.4% missing). A cancer type/site was specified for 88.8% reporting FCH. Age at diagnosis was listed for 21.7% of the relatives’ cancers. Overall, only 20.9% provided all 3 FCH data elements required for running PREMM5 (9.9% of the total sample, n=4738). Fewer men (9.5%) than women (28.1%) provided all 3 FCH elements. Furthermore, 46.7% of breast cancer patients, 21.9% of gastrointestinal cancer patients, 47.2% and 23.1% of patients seen for cancer prevention screening and endoscopy respectively, reported 3 FCH elements. Lower rates were observed for other cancers. Conclusions: Patient self-reported FCH is suboptimal for estimation of LS risk and genetic counseling referral. Future steps to optimize online patient-facing FCH collection to enable routine automated risk-assessment in an essentially provider-free setting may include, patient education regarding importance of FCH, EHR prompts for FCH completion, and implementation of algorithms in EHRs using FCH to identify patients at risk for hereditary cancer predisposition syndromes.