PATIENTS OVERESTIMATE THEIR RISK, BUT THE MAGNITUDE DECREASES WHEN INFORMATION COMES FROM THEIR DOCTOR WITHOUT PATIENT'S OWN ESTIMATIONS

Purpose: While providers spend substantial time discussing the risks and benefits of different treatment options with patients, extant data suggest that cancer patients generally overestimate their risks of recurrence/benefits of treatment. This study aims to examine whether patients' risk estimations differ from published estimates of efficacy in the literature, and whether patient estimates vary according to their source of information.

Method: We surveyed women aged 65 and older who received breast conserving surgery for stage I/II breast cancer from 5 radiation oncology clinics from May 2015 to May 2016. Participants were asked 1) whether their physician provided them with an estimate of their risk of recurrence (yes/no/unsure), 2) to indicate their 10-year risk of recurrence with or without receiving radiotherapy (visual analogue scales), and 3) to identify the source(s) of their risk estimate (my doctor told me/my own estimation/my estimation plus my doctor told me/other). Using multivariate linear regression, we evaluated the associations between the source(s) of their risk estimate and 10-year risk of recurrence without receiving radiotherapy, controlling for age, race, and education.

Results: The response rate was 69.8% (74/106). Sixty-five participants (87.8%) indicated their physician had provided them with risk estimates. Sixty-seven (90.5%) provided estimations of their 10-year risk of recurrence. Among them, 31.3% indicated their estimations came from physicians, 25.4% from their own estimations, 37.3% from both (their own estimation and their doctor), and 6.0% were missing or from other sources. Patients' risk perceptions differed significantly in relation to the attributed information source of their estimations (Figure). In the multivariate analyses, compared with participants whose source came from their own estimation, those whose source came from physicians had significantly lower risk estimates (-20.1%, p=.025). No difference was found for those whose source came from both (-8.0%, p=.331). Undergraduate or higher education (in comparison with high school education or below) was also significantly associated with a lower risk estimate (-14.3%, p=.0497). There was no significant interaction between sources and education on perceived risk of recurrence. 

Conclusions: Participants whose risk perception was based on their own estimation tended to overestimate absolute risk reduction attributed to RT, even among those whose doctors provided this information. Efforts to ensure patient-perceived risks come from doctors could help patients gain an accurate understanding of the intervention benefits.