Conducting a Multicenter Randomized Trial in Hospice-Eligible Individuals with Dementia: Recruitment Strategies, Ethical/Operational Challenges, and Implementation Lessons from the LiBBY Study

Arianne Fritts, MS
Medical University of South Carolina

Co-authors: Jacobo Mintzer¹, Brigid Reynolds^2
1Medical University of South Carolina, ²Georgetown University

Background: Approximately 45% of patients with dementia experience agitation near the end of life, and an estimated 35% of those individuals continue to experience symptoms despite treatment with benzodiazepines, opioids, and atypical antipsychotics. Despite the prevalence and burden of agitation, there is little evidence to guide the management of end-of-life agitation in dementia. Emerging literature suggests that cannabis-based therapies may offer a safe and effective treatment approach. The Life’s End Benefits of cannaBidiol and tetrahYdrocannabinol (LiBBY) Study was designed to evaluate this hypothesis. Conducting randomized clinical trials in hospice-eligible populations presents challenges related to recruitment, informed consent, caregiver engagement, home-based data collection, and medication administration. Additional complexities arose from the use of a Schedule I investigational product, which was not available in the U.S. when the study was initiated and required extensive regulatory oversight.

Objectives: To describe the operational, regulatory, ethical, and methodological challenges encountered in conducting a multicenter randomized clinical trial of cannabis-based therapies for agitation in hospice-eligible individuals with dementia and to identify strategies that enabled successful study implementation.

Methods: This presentation will review key lessons learned during the design and conduct of the LiBBY Study. Topics will include recruitment feasibility, definition of the target syndrome, selection and delivery of the investigational product, implementation of home-based assessments, caregiver engagement, management of concomitant medications, regulatory requirements, and the standardization of study procedures across multiple clinical and hospice settings.

Results: Recruitment was completed in November 2025, with 120 participants randomized across diverse care settings, including a high proportion of individuals from underrepresented ethnoracial backgrounds (58%) and participants living in the community (75%). Key recruitment insights include successful enrollment through partnerships with hospice organizations and non-traditional pathways. Consent processes required careful consideration of capacity, surrogate consent, and sensitivity to caregiver burden in an end-of-life context. Study procedures were adapted to both home and hospice settings. Caregiver involvement was essential for study conduct and outcome assessment. Standardized training approaches were required to ensure protocol consistency across clinical and hospice settings. Study medication had to be imported from Canada, and each site needed to be qualified, registered, and licensed by the DEA to handle Schedule I medications.

Conclusion: The LiBBY study demonstrates that randomized clinical trials in hospice-eligible populations are feasible with tailored implementation strategies. Lessons learned regarding recruitment pathways, hospice partnerships, ethical consent processes, and caregiver-centered approaches provide a practical framework for future trials in advanced dementia and other end-of-life populations.