Date of Award

January 2020

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Tamar H. Taddei


Current trends in liver disease are changing due to increased access to pharmacotherapies and the worsening burden of obesity. Large, observational cohort studies are a common tool in clinical and epidemiological research to determine risk for patients that may have experienced an exposure of interest. Past studies have largely examined risk factors for liver disease in patients with cirrhosis or fatty liver disease. We utilized retrospective cohort studies in a disease-agnostic approach and aimed to characterize risk for two increasingly prevalent populations: patients with severe acute liver injury and those with liver fibrosis progressing to hepatocellular carcinoma (HCC).

First, we studied the Nationwide/National Inpatient Sample (NIS), the largest publicly available United States all-payer inpatient database of over eight million hospital stays, to assess for a meaningful association between weekend admissions and in-hospital mortality for severe acute liver injury. Descriptive statistics were run for patient and hospital factors, and complete-case univariate and multiple logistic regression modeling of in-hospital mortality by weekend admission and other covariates was performed. Our findings did not detect a difference in outcomes for patients hospitalized for severe acute liver injury on weekends compared to weekdays. However, our analysis revealed that racial differences and facility factors may be associated with worse outcomes.

Next, we analyzed the Veterans Affairs (VA) electronic health record data for all veterans born between 1945-1965 (the “birth cohort”) to characterize risk factors for HCC in over two million patients without viral hepatitis (hepatitis B and/or C). By stratifying patients based on Fibrosis-4 Index (FIB-4) for liver fibrosis, we investigated the impact of lifestyle factors on HCC risk. Bivariate analysis revealed increased HCC events among veterans with a low likelihood of cirrhosis (FIB-4<3.25); additionally, ongoing smoking is a significant risk factor in this population. Multivariate analysis revealed that male gender, age >60 years, Hispanic ethnicity, diabetes, excessive drinking, and obesity were significantly associated with HCC, while younger age and black race were associated with decreased HCC risk.

In conclusion, the profile of liver disease is evolving due to multiple sociodemographic factors. Retrospective cohort studies are a useful tool to study trends in patient outcomes and to identify potential areas to mitigate risk.


This thesis is restricted to Yale network users only. It will be made publicly available on 09/10/2022