Preliminary efficacy of VOICE, a decision support tool for older adults with advanced cancer: A pilot randomized trial.

11094 Background: Quality shared decision-making (SDM) requires integrating patient values into treatment decisions. Many older patients (pts) with cancer report that clinicians do not elicit their values when making treatment decisions. We developed a decision support tool for older adults with advanced cancer (called VOICE) with values elicitation, real-time feedback to patients, and tailored question prompt lists. Prior research suggests that variation in clinician behavior may impact the efficacy of SDM tools. Methods: We evaluated the preliminary efficacy of VOICE to improve SDM quality in a pilot double-blinded randomized trial. We enrolled pts ≥ 60 years with advanced solid cancers. Pts were randomized to receive VOICE or an American Cancer Society (ACS) educational document with general cancer-related question prompts. Pts completed 2 simulated consultations where they made a treatment decision for a fictional secondary cancer diagnosis: one values-based (VB), where clinicians initiated values elicitation, and one non-values-based (NVB), where clinicians did not. Logistic regression, Fisher’s exact and Mann-Whitney tests evaluated the relationship of trial arm and clinician behavior to pt reports of SDM quality (CollaboRATE) and perceived usefulness of intervention (PrepDM). Five medical students, trained in VB and NVB consultations, functioned as clinicians during consultations. All consultations were recorded and analyzed for fidelity. Results: Forty-four pts (ages 60-88; 50% female; 86% Caucasian, 7% Black, 7% other) with advanced cancer (34% prostate, 23% breast, 18% lung, 11% melanoma, 14% other) were randomized and completed 88 consultations. Pts were more likely to report quality SDM for VB over NVB consultations (Odds ratio [OR] 2.57, 95% Confidence Interval [CI] 1.09-6.25, p=0.03). Pts reported VOICE was more useful to inform SDM than ACS document (mean PrepDM score 65.1 v. 17.4, p<0.001). Pts in the VOICE arm were more likely to report quality SDM (OR 1.70, 95% CI 0.73-4.04) but this did not reach statistical significance (p=0.22). The difference in the percentage of patients reporting high-quality SDM between arms was larger for NVB (43% VOICE v. 22% ACS) than for VB consultations (57% VOICE v. 52% ACS) though not statistically significant. Conclusions: This study demonstrates the value of using simulated encounters for “early-phase” testing of SDM tools. In this pilot randomized trial using simulated clinical encounters, clinician initiation of a discussion of patient values resulted in the highest quality of SDM regardless of study arm. VOICE improved preparation for SDM and showed encouraging preliminary data for its potential to improve the quality of SDM for older adults with advanced cancer especially when clinicians did not engage in values elicitation. Further research is needed to improve values elicitation as part of treatment decision-making.