Variation in quality of physician communication in prostate cancer consultations

Clear communication of treatment risks and benefits is essential for shared decision making (SDM) in prostate cancer (PC). However, the quality of physician communication in real-world consultations remains poorly characterized. We assessed variation in quality of risk communication in PC consultations and its association with patient, tumor, and provider characteristics.

We conducted a prospective cohort study of 50 PC consultations in a multispecialty cohort of physicians (3 urologists, 3 medical oncologists, 3 radiation oncologists), with an analytic sample comprising 19,388 sentences. Transcripts were coded for five AUA-endorsed SDM content areas: cancer severity, life expectancy (LE), oncologic endpoints, baseline function, and treatment side effects (SEs). Analysts extracted all statements on these topics and scored communication quality using a validated 6-point scale (0 = omission, 1 = name only, 2 = binary assessment of risk, 3 = imprecise quantification, 4 = specific quantification, and 5 = patient-specific estimate). For each consultation, we calculated overall (weighted average) and content-specific scores. We then conducted an exploratory analysis of variation in communication quality by patient, tumor, and provider characteristics.

Communication quality varied substantially across consultations and content areas. Cancer severity had the highest median score (5, IQR 4-5), indicating this information was typically communicated using patient-specific estimates. LE had a median of 3 (IQR 2-5), indicating imprecise quantification. Oncologic endpoints also had a median of 3 (IQR 2-3). Baseline function and SEs had the lowest median of 1 (IQRs 1-2 and 0-2), indicating name-only communication. Exploratory analyses revealed that overall communication quality was lower for Black patients, higher among urologists, increased with higher tumor risk, and varied by physician. Content-specific quality did not improve in subgroups where the content was clinically relevant (e.g., LE quality was not higher for patients with limited LE).

Quality of physician communication in PC consultations varies across content areas and may be influenced by patient, tumor, and provider characteristics.