Back to 2016 Annual Meeting
Impact of Numeracy on Understanding of Prostate Cancer Risk Reduction in PSA Screening
Kevin Koo, MD, MPH, MPhil, Charles D. Brackett, MD, MPH, Ellen H. Eisenberg, MD, Kelly A. Kieffer, MD, Elias S. Hyams, MD.
Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA.
BACKGROUND: Prostate-specific antigen (PSA) screening in men of average risk remains controversial. Results from PSA clinical trials are widely cited in patient education materials, but patients’ ability to incorporate probability and risk data into their decision-making may depend on their numeracy, or facility with quantitative concepts. This study assessed patients’ numeracy and its impact on their understanding of the risk reduction benefits of PSA screening.
METHODS: We used a randomized, cross-sectional survey design. Men 40-75 years old attending a general medicine clinic were randomized to complete one of four surveys. All surveys included demographics, personal PSA and prostate cancer (CaP) history, and a validated three-item numeracy test. Numeracy was scored as the number of items correctly answered (range 0-3). Surveys differed in their presentation of PSA testing risk reduction data derived from the European Randomized Study of Screening for Prostate Cancer (absolute (ARR) vs relative risk reduction (RRR), with or without baseline risk (BR)) (Figure). Respondents were asked to adjust their perceived risk of CaP mortality using the risk reduction data presented. Accuracy of risk reduction was evaluated relative to how risk information was presented. Descriptive analysis was performed.
RESULTS: 200 men completed the survey (60% response rate). Mean age was 60 years, and 91% were high-school graduates. 51% had received a PSA test, and 5% reported a CaP diagnosis. Demographic characteristics were not significantly different among the four survey groups. A majority of respondents incorrectly answered one or more of the three numeracy items; half could not convert “1 in 1000” to a percentage, and one-quarter were unable to calculate 1% of 1000. Overall accuracy of adjustment in perceived risk was 20% among all groups. Accuracy varied with data presentation format: when presented with RRR, respondents were 13% accurate without BR and 31% accurate with BR; when presented with ARR, they were 0% accurate without BR and 35% accurate with BR. Including BR data significantly improved accuracy for both RRR (p=0.03) and ARR groups (p=0.0001). Accuracy was significantly related to numeracy: 6% of respondents were accurate with a numeracy score of 0, 5% accurate with a score of 1, 9% accurate with a score of 2, and 36% accurate with a score of 3 (p=0.006). Neither PSA testing history nor CaP history was associated with accuracy.
CONCLUSIONS: Patients’ numeracy was significantly associated with the accuracy of interpreting quantitative benefits of PSA screening. Although accuracy improved when the presentation of risk reduction data was framed by baseline risk, numeracy in this screening population of men was poor overall. The findings suggest that alternative methods of communicating concepts of risk to patients may facilitate shared decision-making.
Back to 2016 Annual Meeting