15947 Inaccuracy of self-reported family history of melanoma in high-risk patients

Abstract

Family history of melanoma is a known major risk factor for melanoma. However, self-reported family histories for numerous cancers, including melanoma, are commonly inaccurate. Three large melanoma studies have explored this with positive predictive values (PPVs), 28%-83%. Given these wide variations, we explored this question in individuals at high-risk of melanoma, and thus, more likely to be aware of their family history of melanoma. We queried the Utah Population Database (UPDB), a unique, extensive family history database linking >11,000,000 individuals with public records (driver’s license, marriage/divorce records), medical records and death records. This offers a more complete multi-generational dataset for a much larger cohort than would be feasible through manual chart review and is the largest cohort study to date. Charts from 1780 high-risk patients (having at least one melanoma risk factor) were reviewed to quantify self-reported family history of melanoma. These data were combined with UPDB and Utah Cancer Registry data serving as the “gold standard.” Self-reported family history of melanoma in first- and second-degree relatives had an overall sensitivity of 62%, specificity of 67%, PPV of 31%, and NPV of 89%, with increased accuracy for first-degree relatives. A personal history of melanoma was the only factor associated with increased accuracy in self-reported family history of melanoma. Age, sex, number of nevi, and number of prior personal melanomas were not significant. Dermatologists should educate patients on the differences between melanoma, nonmelanoma skin cancer and pre-cancers. Confirming self-reported family history of melanoma will limit over-screening of otherwise low-risk patients.