Back to 2016 Annual Meeting
Are Statistics Subjective? An Assessment of Risk Thresholds to Proceed with Prostate Biopsy
Kevin Koo, MD, MPH, MPhil, Elias S. Hyams, MD.
Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA.
BACKGROUND: “Shared decision-making” (SDM) tools, such as the Prostate Cancer Prevention Trial nomogram, provide quantitative risk information to guide patients’ decisions regarding prostate biopsy (PBx). Newer diagnostic tools (eg, 4Kscore Test) purport to further refine risk predictions. It is unknown if and how patients systematically consider these risk data in their decisions for PBx. We sought to delineate the range of “risk thresholds” for PBx in a urology clinic population to determine how statistical prediction may impact decision-making.
METHODS: Men 45-75 years old attending a multispecialty urology clinic were invited to complete a survey at registration. Collected data included demographics, educational background, employment status, personal and family history of prostate cancer (CaP), and PBx history. Respondents were presented with a summary of the procedural details, risks, and benefits of trans-rectal ultrasound-guided PBx, then asked to indicate on a matrix the risk threshold (%) for high-grade CaP (Gleason score ≥ 7) at which they would proceed with PBx. This was intended to mirror a clinic visit in which we share this information with patients in an SDM format. Data were compiled for descriptive analysis.
RESULTS: 103 men completed the survey (53% response rate). Mean age was 61 years (interquartile range [IQR] 56-66.5), and 97% were high-school graduates. Eighteen respondents (17%) had a personal history of CaP, and 31 (30%) had undergone PBx. The median risk threshold for high-grade CaP to proceed with biopsy was 25% (IQR 10-50%). Respondents’ thresholds did not vary significantly with race, educational background, or employment, but thresholds were higher for men with lower income (p=0.048). Personal history of CaP or PBx was significantly associated with lower mean thresholds (18.7% vs. 31.7% [p=0.02] and 23.2% vs. 32.8% [p=0.04], respectively). Family history of CaP, lethal CaP, or biopsy complications did not significantly affect thresholds. The sample was divided into quartiles based on thresholds for biopsy; there were significantly higher rates of history of CaP and prior PBx among lowest vs. highest quartile respondents (36% vs. 1% [p=0.01] and 46% vs. 17% [p=0.008], respectively).
CONCLUSIONS: We found a wide range of high-grade CaP “risk thresholds” to proceed with prostate biopsy, underscoring the importance of individualized discussions with men in an SDM setting. Men with a prior history of CaP or biopsy reported lower thresholds to proceed with biopsy, which may reflect their greater concern for and exposure to this disease. It is unclear to what extent refined risk prediction tools will improve SDM or influence decisions to proceed with PBx. Further study of risk thresholds in a more CaP-naïve primary-care population may help clarify patient perspectives.
Back to 2016 Annual Meeting