Abstract
Effect of real-time data-driven physician engagement on appropriate precision oncology testing.
Author
person
Ying Liu
Diaceutics, Parsippany, NJ
info_outline
Ying Liu, Howard L. McLeod, Jeff Schreier, Nirmala Bhogal, Jordan Clark, Eve Thompson, Arun Rompicherla, Gemma Little, Joshua McKenna, Anjum Ismail, Sarah Varghese, Bethany Michelle Slifko
Full text
Authors
person
Ying Liu
Diaceutics, Parsippany, NJ
info_outline
Ying Liu, Howard L. McLeod, Jeff Schreier, Nirmala Bhogal, Jordan Clark, Eve Thompson, Arun Rompicherla, Gemma Little, Joshua McKenna, Anjum Ismail, Sarah Varghese, Bethany Michelle Slifko
Organizations
Diaceutics, Parsippany, NJ, Utah Tech University, St. George, UT, Diaceutics Plc, Belfast, United Kingdom, Diaceutics PLC, Belfast, United Kingdom, Diaceutics, Belfast, United Kingdom
Abstract Disclosures
Research Funding
No funding sources reported
Background:
Precision medicine is a paradigm shift in healthcare: 24% of all FDA approvals are precision medicines, with 67% of new approvals relying on companion or complementary testing or on diagnostics prior to clinical use. Yet a recent publication indicates that approximately 64% of potentially eligible patients could be lost due to various clinical practice gaps along patient journey (Sadik et al JCO 2022). Low prevalence biomarker testing is proven to be one of the critical determinants of missed therapy, potentially due to poor physician awareness. The increasingly advanced machine learning (ML) and natural language processing (NLP)-based ability to label, analyze, and segment data can be utilized to monitor and influence testing behavior in real-time. We hypothesize that a real-time data-driven physician-targeted digital engagement can increase biomarker testing rate and in turn enhance patient identification for precision therapies.
Methods:
Unstructured real-world, real-time biomarker testing reports were collected from 494 laboratories across the US. These testing reports were first curated, labeled and extracted via ML and NLP-based technologies to allow data analysis and reveal testing behavior. Suboptimal testing behavior was defined as no requisition of the selected novel biomarker as part of the diagnostic workups. Physicians who identified as having suboptimal testing behavior were selected as candidates to receive a personalized digital engagement over a 4-week period. The digital engagement's purpose was to raise clinical awareness towards the novel biomarker testing. The follow-up testing behavior was analyzed again over laboratory reports collected from 26-weeks’ time after the initial engagement.
Results:
822 physicians were identified with suboptimal testing behavior and selected to receive 4 weeks' follow-up digital engagement. 271 out of 822 (33%) recipients were successfully engaged shown as open and read the digital communication. 42 (~15%) physicians reacted to the communications and ordered the new test at least once during this period. The initial 4 weeks' engagement was observed to be the most impactful, when 23 out of the 42 (~55%) physicians ordered the novel biomarker test for the first time.71 new therapy-eligible patients were identified via data-driven targeted physician digital engagement during the 26 weeks’ follow-up period.
Conclusions:
This study illustrates the value of data in influencing or reinforcing testing behaviors and emphasizes the value of data-driven physician-targeted digital engagement in addressing clinical practice gaps which in turn, drive patient identification to match potential eligible precision therapies.
8 organizations
Organization
DiaceuticsOrganization
ParsippanyOrganization
NJOrganization
Utah Tech UniversityOrganization
St. George's University of LondonOrganization
UTC TherapeuticsOrganization
Diaceutics PlcOrganization
Belfast Health and Social Care Trust