The Effect of Race and Decision Aids on Provider Recommendation for Prostate Cancer Treatment Planning: A Randomized Study

Abstract

Racial disparities in care among patients with prostate cancer have been well documented. ASCO proposed a value framework (VF) to assist in decision making. The VF does not make a treatment recommendation but tries to present the efficacy, toxicity, and cost data in a clear format to aid in collaborative decision making. We conducted a randomized survey to assess the impact of race and VF on the decision to recommend adjuvant radiotherapy (ART) on radiation oncologists (RO), urologists (U), and medical students (MS). A secure anonymous internet-based survey with a hypothetical clinical vignette of an elderly man with pT3bN0M0 Gleason 4+3=7 prostate cancer who had recently undergone radical prostatectomy was emailed to 410 RO and 400 U at multiple institutions and 523 MS a single institution. Participants read identical clinical vignettes (with relative equipoise as whether to treat or not) but were randomly assigned a White or Black patient accompanied by a representative patient photograph. After viewing the vignette, participants were asked if they would treat the patient with ART or observation (OBS). Participants were then shown the VF of the ART study SWOG 8794 and asked again if they would treat the patient with ART or OBS. Pearson’s Chi-squared test was used to test relationships between patient race, provider specialty (RO, U, or MS), and treatment recommendation. Multivariable logistic regression analyses assessed the effect and interaction of race, provider specialty and pre- vs. post-VF on the decision to recommend adjuvant radiation. Responses from 120 RO, 98 U, and 122 MS were recorded. Overall response rate was 25.5%. Both U and MS were more likely to recommend OBS vs. ART. For both U and MS there were no differences in recommendations between those randomized to Black vs. those randomized to White patients (p>0.2). The VF also had no significant impact on U or MS recommendations. Only RO randomized to the Black patient (vs. White patient) chose ART more often than OBS (67.7% vs. 48.3%, p=0.03). After viewing the VF, RO were even more likely to choose ART for Black men (77.4% vs. 37.9%, p<0.001). There was a significant interaction between VF (pre vs. post) and RO (vs. non-RO) for RO randomized to the Black patient (Pinteraction<0.001). RO, compared to U and MS, treat Black men more aggressively than White men and this tendency was amplified by the VF. In contrast, U and MS were unaffected by the VF, both tending to recommend OBS. It is possible that the VF acts to solidify existing opinions regarding treatment decisions, especially for RO, who may be more likely to interpret the results of SWOG 8794 as favoring ART.Abstract 1226; Table 1Pre-VF % ART RecommendedPost-VF % ART RecommendedBlackWhiteP-valueBlackWhiteP-valueMedical Student46.8% (29/62)36.7% (22/60)0.2632.3% (20/62)31.7% (19/60)0.94Radiation Oncologists67.7% (42/62)48.3% (28/58)0.0377.4% (48/62)37.9% (22/58)<0.001Urologists14.3% (7/49)19.6% (10/51)0.4816.3% (8/49)23.5% (12/51)0.37 Open table in a new tab