2020-P: Communication as a Modifiable Factor in Insulin Inertia: An Analysis of Mock Consultations Using ConversationAIly

Introduction and Objective: Delays in insulin initiation persist in type 2 diabetes (T2D) despite guideline recommendations. While patient reluctance is often cited, clinician communication at the point of prescribing may represent an underrecognized contributor. ConversationAIly™ is a clinical communication audit system designed to identify empathy gaps, stigmatizing language, and directive framing during consultations. We aimed to characterize communication patterns during insulin initiation and titration in T2D and identify modifiable language features influencing insulin use. Methods: We analyzed 197 standardized mock consultations conducted by 15 clinicians and 10 residents using ConversationAIly™. Each transcript was audited for empathy score, tone, stigmatizing versus person-first language, medical jargon density and directive versus collaborative phrasing. Results: In newly diagnosed patients, residents more often framed insulin as urgent severity, using jargon-heavy explanations with limited exploration of patient concerns, while clinicians relied on reassurance but remained largely directive. After failure of multiple oral agents, clinicians most strongly adopted failure- and escalation-based language, with higher stigma and lower empathy; residents echoed this framing with greater biomedical justification. Titration discussions shifted toward collaborative, problem-solving language with relatively higher empathy. Across scenarios, empathy differences were driven by framing and word choice rather than consultation duration. Participants acknowledged that ConversationAIly™ highlighted missed opportunities for patient-centered communication. Conclusion: Insulin inertia is partly constructed through routine clinical language at the point of prescribing. Treating communication as a measurable, modifiable process offers a pragmatic strategy to support timely insulin initiation and titration in T2D, with direct implications for clinical practice, training, and CME. Disclosure R. Mallabadi: None. S. Dash: None. J. Swain: None.