Many Black/African Americans and Hispanic/Latinx patients are much more likely to receive low quality end-of- life medical care than White individuals– in fact, they are 3 times more likely than white Americans to die after a lengthy intensive care unit stay. Advance care planning (ACP)– the process of discussing one’s wishes with loved ones and clinicians, and then documenting them in an advance directive (AD)– can help reduce these health inequities by preventing costly/burdensome treatments that are unlikely to reduce suffering or improve quality of life. Though ~60% of Americans engage in ACP, <25% of underserved populations have done so– in large part due to distrust of the healthcare system/clinicians, and reluctance to discuss death and dying. This study harnesses underserved communities’ existing, trusted social networks to deploy two community-based ACP interventions and study their mechanisms of action. By identifying which interventions increase engagement in ACP in underserved communities (and why), this project will help improve quality of end-of-life care, reduce unnecessary suffering, and end-of-life healthcare costs which conserves public health resources. Over the past 5 years, our team has studied an inexpensive and readily scalable serious game called Hello that prompts discussion of sensitive end-of-life issues. Across multiple studies, participants report that playing Hello is enjoyable, eye-opening, and motivating —98% of participants subsequently performed at least 1 ACP behavior. Recently, we developed a nationwide community-based delivery model for ‘Hello’ and confirmed the game’s acceptability by engaging 53 underserved communities in ACP (n=1,165). We now propose a 3- armed, cluster, randomized controlled trial (RCT) in diverse, underserved populations to compare the efficacy of ‘Hello’ (Group 1) with a nationally promoted structured workbook ‘The Conversation Project Starter Kit,’ (Group 2), and a non-ACP game called ‘Table Topics,’ (placebo/attention control; Group 3). We will randomize 75 underserved communities across the US (20 participants/site; n=1,500). The primary outcome is completion of a visually verified AD; secondary outcomes include performance of other ACP behaviors. This study will provide key scientific advancements by: 1) providing efficacy data on two widely used and easily scalable but not yet evidence-based interventions; 2) advancing the science of interventional design by examining the interventions’ potential mechanisms of action (i.e. quality of communication, and role of socio- cultural environment); and 3) assessing how and why our community-based delivery model engages underserved communities in ACP. Should the RCT have negative findings, we still will have gained a robust understanding of the sociocultural environment’s role in population health research. If successful, this project will provide an evidence-based model for engaging underserved communities in ACP, along with a robust understanding of how to design and deliver community-based initiatives relevant for other population health- based research or initiatives for the underserved.

PROJECT NARRATIVE Compared to the general population, individuals from underserved communities are more likely to receive low quality end-of-life care and unwanted, costly and burdensome treatments due in part to a lack of advance care planning (ACP; the process of discussing wishes for end-of-life care with loved ones/clinicians and documenting them in advance directives). This study will use existing, trusted, and respected social networks to evaluate two conversation-based tools intended to engage underserved individuals in discussions about end-of-life issue and motivate them to carry out ACP behaviors. Through this study we will learn how best to engage underserved populations in ACP so as to: 1) increase the likelihood that patients from underserved communities will receive high-quality end-of-life care; 2) address health disparities related to end-of-life treatments; and 3) reduce unnecessary suffering for patients and their families.