Date of Award

Fall 12-15-2017

Degree Type

Thesis

Degree Name

Master of Public Health (MPH)

Department

Public Health

First Advisor

Dora Il'yasova, PhD

Second Advisor

William Thompson, PhD

Abstract

Introduction:

Hepatitis C Virus (HCV) Infection is the most common blood-borne infection in the world with a global prevalence of ~3%. It also represents an underestimated and under-recognized viral infection because it is asymptomatic during the initial period of infection, which tends to span several decades. However, after establishing itself as a chronic state, HCV infection often leads to severe debilitating liver conditions such as cirrhosis, hepatocellular carcinoma to name a few resulting in poor quality of life, increased healthcare costs and mortality. Several earlier studies have examined risk factors associated with HCV. However, a comprehensive study that has simultaneously evaluated a wide range of addictive risk behaviors associated with HCV has not been conducted to date. This type of investigation will help to identify at-risk populations for HCV and provide valuable information regarding how one might efficiently link them to appropriate treatment and care.

Aim:

The primary aims of this study were 1). To estimate chronic HCV infection (CHI) prevalence in non-institutionalized U.S. adult population from 2003-2014 2). To perform a multivariate examination of all known behavioral risk factors significantly associated with CHI and 3). To identify less invasive questions regarding risk behaviors associated with CHI that could be used to predict state-level CHI prevalence using other state-specific data sources such as The Behavioral Risk Factor Surveillance System (BRFSS).

Methods:

The study utilized nationally representative data from National Health and Nutrition Examination Survey (NHANES) for the years 2003-2014. HCV RNA positive persons were CHI positive population. Bivariate analyses were performed to examine the frequency distributions of the study’s primary dependent variable (CHI) and all independent variables (demographical + behavioral risk factor variables). The analysis sample included 11,596 adults aged 20-59 years. Risk factors for CHI were examined using both bivariate and multivariate logistic regression analyses. We first conducted weighted bivariate logistic regression analyses to examine the relationships of dependent and independent variables without controlling for potential confounders. We then conducted weighted multivariate adjusted logistic regression analyses to examine the relationships between the dependent and independent variables while controlling for potential confounders. Odds ratios (OR), 95% confidence limits (CL) and p-values were calculated. A p-value < 0.05 was considered statistically significant. SAS 9.4 was used for all statistical analyses.

Results:

The estimated number of CHI adults aged >/= 20 years in 2014 was 1.93 million leading to an estimated CHI prevalence of 0.7%. Injection drug users (IDU) had the highest CHI prevalence of 30.24% by bivariate analyses. Multivariate logistic regression analysis in the chronic HCV full model indicated that age categories 40-49 y (OR: 7.9, 95% CL: 3.8-16.2) and 50-59 y (OR: 8.0, 95% CL: 3.5-18.2); non-Hispanic blacks (OR: 2.4, 95% CL: 1.3-4.1); less than high school education (OR: 2.6, 95% CL: 1.5-4.8); < 2.0 times the poverty level (OR: 3.5, 95% CL: 1.9- 6.6); heroin consumers (OR: 2.3, 95% CL: 1.1-4.6); IDU (OR: 8.1, 95% CL: 3.1-21); blood transfusion recipients (OR: 2.9, 95% CL: 1.4-5.7) and >/= 10 lifetime sex-partners (OR: 5.5, 95% CL: 1.5-19.7) were significantly associated with CHI.

Multivariate logistic regression analysis in the chronic HCV risk factor model indicated that persons in moderate (OR: 2.5, 95% CL: 1.1-5.5) and high (OR: 30.3, 95% CL: 12.1-76) substance abuse risk factor categories were significantly associated with CHI. Multivariate logistic regression analysis in chronic HCV BRFSS model indicated that age categories 40-49 y (OR: 7.5, 95% CL: 3.5-15.9) 50-59 y (OR: 8.7, 95% CL: 4.2-18); males (OR: 3.1, 95% CL: 1.5-6.4); non-Hispanic black (OR: 1.8, 95% CL: 1.1-2.9); less than high school education (OR: 2.2, 95%CL: 1.3-3.8); < 2.0 times the poverty level (OR: 3.7, 95% CL: 2.0-6.8); alcohol consumers (OR: 1.7; 95% CL: 1.0-2.9) and smokers (OR: 3.7, 95% CL: 1.8-7.6) were significantly associated with CHI. c-statistic of the three models were 0.94, 0.92 and 0.88 respectively thereby implying that all three models were strong models with a higher predictive accuracy of CHI.

Conclusions:

We conclude that the estimated prevalence of CHI in this analysis sample is 0.7%, however the true prevalence estimates of CHI are likely to be significantly higher if incarcerated, homeless and other population not presented in NHANES are included. IDU continues to be the strongest risk factor for CHI. Persons with two or more addictive behavioral risk factors have significant associations with CHI. Results from this study will enable identification of at-risk population for CHI and provide valuable information for linking them to appropriate treatment and care.

Share

COinS