Evaluation of the IMPACT of a decision-support platform (Trapelo) on testing patterns for patients with advanced solid tumors.

Abstract

513 Background: Given the evolving precision oncology landscape, it is challenging for oncologists to stay abreast of the latest biomarker testing guidelines. Texas Oncology (TxO), a network of community oncology providers, is evaluating Trapelo, a decision-support platform that assists in the ordering of appropriate biomarker tests. The objective of this study was to assess whether implementation of Trapelo’s testing pathways improves the ordering of timely and comprehensive biomarker testing (TCBT) among patients with advanced cancer. Methods: We conducted a retrospective analysis of patients with advanced non-small cell lung (NSCLC), metastatic breast (BC), or metastatic colorectal (CRC) cancer treated at TxO between July 24, 2018, and July 24, 2021, and had at least one oncology biomarker test ordered. The ‘Intervention’ group consisted of patients diagnosed after their physician’s onboarding to Trapelo; the ‘Control’ group consisted of patients diagnosed before onboarding. We also analyzed a subset of the intervention group: patients with biomarker tests actually ordered through Trapelo (Intervention - Trapelo.) Our primary endpoint, TCBT, was defined as the ordering of tests for all actionable biomarkers (i.e., with FDA approved first-line treatment) within 60 days following an advanced diagnosis. Because Trapelo implementation occurred over time, we used Two-Way Fixed Effects regression with fixed effects for time period and physician to assess the impact of the Intervention on TCBT. Results: In descriptive analysis, the rates of TCBT in the NSCLC cohort were 26, 28%, and 37% in the Control (n = 1023), Intervention (n = 554), and Intervention-Trapelo (n = 315) groups, respectively. In the CRC and BC cohorts, the rates were 32%, 40%, 55%, and 66%, 55%, and 45%, respectively. In the regression analysis, Intervention was associated with increases in TCBT rates of 7.5 ( p = 0.4) and 6.5 ( p = 0.4) percentage points in the CRC and NSCLC cohorts, respectively. The lack of statistical significance could be attributable to small sample sizes and clustered data by physicians. No differences were observed in the BC cohort. Conclusions: While trends observed in our study indicate that implementation of a decision support tool can improve rates of TCBT in a community oncology setting, studies with larger sample sizes are needed to assess the statistical significance of these findings. Despite these potential improvements, overall rates of TCBT, a critical aspect of optimal treatment in oncology, remained low. Further, over 43% of the NSCLC patients in the Intervention group did not have their orders placed through Trapelo, indicating an opportunity for improving post-onboarding adoption of Trapelo. Future analyses will investigate the impact of Trapelo utilization (as opposed to implementation/onboarding) on the receipt of TCBT.