Date of Award

January 2026

Document Type

Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

Tamar Taddei

Abstract

Abstract

PREDICTORS AND OUTCOMES OF PALLIATIVE CARE CONSULTATION IN VETERANS WITH ALCOHOL-ASSOCIATED HEPATITIS

Authors: Olivia Belliveau, Catherine Mezzacappa (Department of Medicine, Division of Digestive Diseases, Yale School of Medicine, New Haven, CT), Arpan Patel (Department of Medicine, Division of Digestive Diseases, David Geffen School of Medicine at University of California, Los Angeles, California), Jennifer Kapo (Palliative Care Program, Yale School of Medicine, New Haven, CT), David Kaplan (Department of Medicine, Division of Gastroenterology and Hepatology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA), Tamar Taddei (Department of Medicine, Division of Digestive Diseases, Yale School of Medicine, New Haven, CT)

Background: Acute alcohol-associated hepatitis (AH) carries a high mortality risk. Patients with advanced chronic liver disease often receive intensive healthcare interventions at the end of life. Palliative care aims to clarify patients’ goals, align their care to their values, and address patients’ physical and emotional symptoms while facing life-limiting illnesses.

Research Aims: Little is known about utilization of palliative care (PC) consultation in AH. In a cohort of Veterans hospitalized with AH, we aimed to characterize predictors of specialty PC consultation and short-term mortality.

Hypothesis: We hypothesized that higher MELD 3.0 score, comorbid cancer diagnosis, and pre-hospital PC engagement would be associated with receipt of a specialty PC consult among Veterans hospitalized with AH, and that short-term mortality would be high across both groups.

Methods: A validated algorithm using laboratory measures and diagnosis codes was used to identify first hospitalizations with AH between 2008-2022 from a cohort of Veterans with abnormal liver enzyme measurements. Baseline was defined as 30 days prior to hospitalization. In multivariable analyses, we identified factors associated with PC consultation, in-hospital mortality, and post-discharge mortality. Logistic regression was used to predict PC consult receipt. Mortality during and after hospitalization was compared across PC consult receipt, assessed by logistic regression (in-hospital mortality), Cox proportional hazards regression, and accelerated failure time modeling (60 day post-discharge mortality).

Results: We identified 4,098 first AH hospitalizations. At baseline, 33.9% had a diagnosis of cirrhosis, 1.4% had hepatocellular carcinoma, and 3% had metastatic cancer. Overall, 743 patients (18.1%) received a PC consult, which did not significantly change over time. Among these consults, 65.1% occurred late (greater than 3 days post-admission). In multivariable analysis, independent predictors of receiving a PC consult included higher age at admission (odds ratio [OR] for 5-year increase [95% CI]: 1.23 [1.16-1.30], p<0.001), admission MELD 3.0 score (OR for 5-point increase [95% CI]: 2.25 [2.06-2.45], p<0.001), HCC at baseline (OR [95% CI]: 2.10 [1.07-4.12], p=0.03), and pre-hospital PC engagement (OR [95% CI]: 2.54 [1.75-3.68], p<0.001). Black Veterans hospitalized with AH were significantly less likely to receive a PC consult compared to White Veterans (OR [95% CI]: 0.48 [0.35-0.66], p<0.001)

Median time from PC consult to death was 12 days (IQR: 5-38). Veterans who received a PC consult were more likely to die while hospitalized (29.6% vs 3.7%, p<0.001) and within 60 days of discharge (39% vs 11.7%, p<0.001). In multivariable analysis, predictors of in-hospital mortality included receipt of PC consultation (OR [95% CI]: 5.84 [4.33-7.87], p<0.001), age at admission (OR for 5-year increase [95% CI]: 1.12 [1.04-1.21], p=0.004), and higher admission MELD 3.0 score (OR for 5-point increase [95% CI]: 1.56 [1.40-1.74], p<0.001). Hispanic Veterans were less likely to die while hospitalized than non-Hispanic White Veterans (OR [95% CI]: 0.49 [0.28-0.85], p=0.01). In the multivariable proportional hazards mode, predictors of increased risk of death after discharge included receipt of PC consultation (hazard ratio [HR] [95% CI]: 3.84 [3.18-4.65], p<0.001), age at admission (HR for 5-year increase [95% CI]: 1.17 [1.12-1.23], p<0.001), other or unknown race or ethnicity (HR [95% CI]: 1.55 [1.22-1.96], p<0.001), and admission MELD 3.0 score (HR for 5-point increase [95% CI]: 1.50 [1.40-1.60], p<0.001).

Impact on Communities of Interest: Receipt of specialty PC consultation was associated with greater burden of illness, including higher MELD score, concomitant HCC, and older age. Receipt of specialty PC consultation was associated with mortality and occurred close in time to death, suggesting that primary teams consider PC for the sickest patients and primarily as end-of-life care, which may limit its potential benefits to other patients. The findings can help shape VHA practices around specialty palliative care. Workforce expansion or specific training for hepatologists in primary PC skills may be required to improve access to PC for more patients in the VHA and other healthcare settings. Additionally, Black Veterans were less likely to receive a PC consult; racial and ethnic disparities in receipt of PC require further investigation.

Comments

This thesis is restricted to Yale network users only. It will be made publicly available on 07/14/2027

Download

Share

COinS