Impact of an EHR-Embedded Clinical Decision Support (CDS) Tool on Guideline Adoption: A Retrospective Analysis of Clinician Adaptation to Treatment Intensification in Prostate Cancer

Flatiron Assist^TM (FA) is a customizable clinical decision support (CDS) tool embedded in the electronic health record (EHR) to facilitate selection and documentation of National Comprehensive Cancer Network (NCCN) concordant treatment regimens, undergoing monthly updates aligned with NCCN Guidelines. In prior research, CDS tool users showed accelerated uptake of new therapies compared to non-users.¹ This retrospective study explores how CDS tools affect clinician adaptation to a guideline update. Single-agent docetaxel was US Food and Drug Administration (FDA) approved in 2004 and recommended in NCCN Guidelines until deprecation in the 1.2023 update; replaced by two new combination regimens for metastatic castrate-sensitive prostate cancer.

This study used the longitudinal nationwide Flatiron Health EHR-derived de-identified database comprising de-identified patient-level structured and unstructured data, which was curated via technology-enabled abstraction. Ordering patterns between users and non-users of the CDS tool over a 16-month period were compared, eight months before and after the regimen update (April 2022-August 2023). Orders for analysis included docetaxel alone and in combination with darolutamide or abiraterone and steroids, identified using regular expression-based filtering on generic and brand names.

Prostate cancer orders (24 650) placed by 2049 providers were analyzed across 151 community oncology practices, 16 CDS tool-enabled sites (median 43.5 providers [range 2-227]) and 135 non-enabled sites (5 [1-57]). A logistic regression model predicting the use of docetaxel alone vs combinations was fit on binary indicators for the order occurring after the guideline update, whether the provider was using the CDS tool, and the interaction between these two indicators.

The model was adjusted for the number of ordering providers per practice. The interaction term was statistically significant (OR 0.11, P < 0.001), suggesting a greater reduction in the use of single agent docetaxel after the guideline update at CDS tool enabled sites (Table 1). CDS tool users ordered the new combination regimens, with a median of 69.3 and 152.1 days (average 110.7), sooner than non-users.

Table 1. Proportion of Orders for Single-Agent Docetaxel vs Docetaxel/Darolutamide or Docetaxel/Abiraterone/Steroid

Practices utilizing the CDS tool adopted NCCN Guideline updates sooner and more broadly than non-users, highlighting CDS tools as an effective tech-enabled driver in behavior transformation. The median size of practices using CDS tools was different from that of non-users, and future research is needed to examine how practice size factors into ability and willingness to engage with technological innovation.

Authors:

Rebecca Maniago¹, PharmD, BCOP; Anna Bolha¹, MPAS, PA-C; Taylor Dias-Foundas¹; Spencer Langerman¹; James Hamrick, MD, MPH¹; Sylvia Richey, MD²; Colin Harvey¹

Affiliations:

¹Flatiron Health, New York, NY 
²West Cancer Center, Germantown, TN