Patient and clinician perspectives on the timing of serious illness conversations in oncology.

223 Background: Serious illness conversations (SICs) about patients’ values and care preferences are tied to improved outcomes and quality of life. However, SICs can be hard to implement early in the cancer care journey, and questions remain about how and when to start. Following a pragmatic trial testing behavioral nudges to promote SICs in oncology (NCT04867850), we used an explanatory-sequential mixed methods design to gauge patient and clinician perspectives on the best communication strategy for SICs. Methods: Thirty patients with cancer at high mortality risk and 16 oncology clinicians were recruited from our academic cancer center. Guided by the Consolidated Framework for Implementation Research, our interviews systematically assessed multi-level factors shaping SIC implementation. In addition to interviews, patients completed a structured survey with validated measures on patient-focused communication and care planning. Survey data were studied descriptively, interviews were analyzed thematically, and both were integrated to triangulate qualitative and quantitative results. Results: Participants described two SIC types, which we classified as event-triggered and phased . Clinicians identified event-triggered SICs as prompted by acute incidents (e.g., disease stage transition) and more likely to lead to EHR documentation. However, they were seen as prohibitively time-consuming, and as such, did not become a top priority until a crisis arose. Patients were also uncomfortable with extensive early SICs, wanting to keep hope and avoid information overload soon after a diagnosis. They preferred not to discuss “turns for the worse” until they faced an acute challenge, at which point they valued thorough SICs. In surveys, almost all patients reported asking about benefits (96%), risks (93%), and quality of life (81%) tied to treatment, but few raised sensitive topics like planning for if things got worse (44%). By contrast, phased SICs were described as ongoing, progressive conversations not prompted by specific events and that arose early but deepened over time as a patient and care team nurtured their relationship. Patients and clinicians described phased SICs as better at promoting person-centered care and an essential foundation for SICs prompted by potential transitions. However, since these discussions occur gradually, clinicians reported challenges documenting them, reducing their value at informing future decisions and the ability to know whether conversations occurred at all. Conclusions: Interventions to increase SICs often focus on acute moments in the care trajectory. However, our study suggests that ongoing, staged approaches to SICs are vital for goal-concordant care as well and may be more easily integrated into busy clinical workflows. Further work should investigate ways to support clinicians in both SIC types, such as flexible documentation tools aligned with each approach.