Background: Ensuring that Veterans with serious illness receive patient-centered care is a fundamental goal of VA. Advanced liver disease (AdvLD) is a serious illness that disproportionately affects Veterans and is characterized by declining health, increasing symptoms and frequent hospitalizations. Recent patient-centered models of care in other conditions, like cancer, have promoted early integration of supportive and palliative with curative care. These models can improve quality and even length of life. Integrated care in AdvLD is essential because it can reduce complications, maintain both function and well-being, and may even prolong life. Objectives: Patient-centered models of integrated AdvLD care have been slow to develop due to specific knowledge gaps. First, precise and complete understanding of disease severity and progression is lacking. Second, no studies exist that characterize Veterans' experiences and understanding of their AdvLD severity, or goals of care across the spectrum of disease severity. Third, little is known about clinicians' experiences, expectations, and perceived barriers to delivering AdvLD care. Our proposed study, conducted in close collaboration with AdvLD patients, caregivers, clinicians and clinical operations partners, will use a multi- method approach to fill these gaps so crucial to developing integrated care for AdvLD. Our specific aims are: (1) to develop risk prediction models to stratify patients into groups at different risks for key AdvLD outcomes; (2) to describe patient and caregiver experiences and goals of AdvLD care, including perceptions of illness severity, preferences for the amount and type of risk information desired, and expected health outcome goals; and (3) to identify clinicians' perceptions of opportunities and barriers to patient-centered AdvLD care. Methods: For Aim 1, we will conduct a national retrospective cohort study of Veterans with AdvLD seen in the VA between 2011 and 2015. We will use pre-existing administrative and clinical data in VA Corporate Data Warehouse to combine liver severity indices with sociodemographic (age, gender, homelessness), clinical (physical and mental health comorbidity, alcohol use), and healthcare resource use (hospitalization, emergency room visits) factors to provide risk category estimates for three patient-centered outcomes: risk of developing AdvLD complications, requiring AdvLD related hospitalizations, and overall mortality. To achieve Aim 2, we will conduct in-depth qualitative interviews with 60 patients who have AdvLD, and 30 caregivers, at 3 diverse VA facilities. Participants will be stratified by risk strata (e.g., low, intermediate or high risk of mortality from AdvLD-specific complications). To achieve Aim 3, we will conduct in-depth qualitative interviews with 30 clinicians involved in direct care of patients with AdvLD. Interviews will examine their experiences in communicating risk and making treatment plans; perceptions of their and ...