Abstract
BACKGROUND: Despite increasing evidence to support risk-based breast cancer screening, individuals' understanding of personal risk is not well understood. This study compares women's perceptions of risk to their estimated risk, and examines factors associated with perceived risk, including breast density notification, within a population-based screening program. METHODS: A survey of 5784 women measured their perceived risk via three questions: a number from 0 to 100 (numeric), a category from very low to very high (verbal), a comparative category relative to an average woman (comparative). Descriptive analyses assessed correlations between perceived risk variables and estimated risk (using the Gail Model), and modelled relationships using K-fold cross-validation. A Graded Response Model was used to obtain an index of unobserved (latent) overall perceived risk from the three questions. Multivariable modelling was used to investigate factors associated with overall perceived risk. RESULTS: Most participants perceived themselves as being at neither high nor low risk, although perceived risk was higher than estimated risk, on average. All three perceived risk measures were positively correlated with each other and with estimated risk. Overall perceived risk was weakly associated with estimated risk (adjusted R(2) = 0.12). Women who received multiple breast density notifications, were younger, or had a family history, perceived their risk as higher relative to respective reference groups. Those who identified as Asian perceived their risk as lower than those who identified as European/Caucasian. CONCLUSION: Individuals' understanding of breast cancer risk is poor. New strategies are needed to improve education and awareness of personal risk.