Author ORCID Identifier

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Public Health

First Advisor

Heather Bradley, Ph.D.

Second Advisor

Katherine Masyn, Ph.D.

Third Advisor

Pascale Wortley, MD


Injection drug use (IDU) behavior has increased in the U.S. during the past decade largely due to the on-going opioid epidemic. People who inject drugs (PWID) have substantial risk for HIV and hepatitis C virus (HCV) infections. This dissertation consists of three studies to understand risk for HIV and HCV among PWID and people living with HIV (PLWH) in Georgia, as well as missed opportunities to address these infections.

In study 1, we constructed a retrospective longitudinal cohort using clinical encounters with patients who had probable recent IDU behavior based on diagnostic codes in electronic medical records at a metro Atlanta hospital during 2012 – 2018. We linked cohort data with HIV and HCV surveillance records from Georgia Department of Public Health (GDPH) to examine prevalence of infections at clinical discharge and incidence of infections post-clinical encounters. Nearly 4% of patients with IDU-related clinical encounters were later diagnosed with HIV, and 17% were later diagnosed with HCV, translating to incidence rates of 9.3 per 1,000 person-years and 42.9 per 1,000 person-years, respectively. Results from Poisson models indicate that patients aged 16­-39 years at discharge were less likely than older patients to be later diagnosed with HCV (Adjusted Incidence Rate Ratio [IRR]=0.63, 95% Confidence Interval [CI]=0.41-0.97, p=0.04) The majority of HIV and HCV diagnoses post-discharge occurred among Black/African Americans and males. At the time of clinical discharge, 32.9% of patients had an HIV diagnosis and 28.1% of patients had an HCV diagnosis.

In study 2, we used IDU-related clinical encounters at an urban Atlanta hospital spanning January, 2012 – December, 2018 to estimate the frequency of HIV and HCV testing at clinical encounters. We also assessed associations between patient factors and testing using unadjusted and adjusted generalized estimating equations models. Of encounters eligible for HIV or HCV testing, testing occurred in 29.3% and 12.2%, respectively. Testing was less likely among Black/African American patients compared to white patients (HIV, adjusted odds ratio [AOR]=0.43, 95% CI, 0.29-0.63, P <0.01; HCV, AOR=0.43, 95% CI, 0.26-0.72, P <0.01). Testing was more likely to occur in encounters during 2016-2018 than in encounters during 2012-2013; (HIV, AOR=4.73, 95% CI, 2.72-8.23, P <0.01; HCV, AOR=3.74, 95% CI, 1.93-7.24, P <0.01) and in those requiring five days or longer hospital stays compared to those requiring less than five days or emergency department (ED) visits (HIV, AOR=3.70, 95% CI, 2.30-5.95, P <0.01; HCV, AOR=4.49, 95% CI, 2.78-7.25, P <0.01).

In study 3, we constructed a retrospective cohort of PLWH using matched GDPH HIV and HCV case surveillance data from persons diagnosed with HIV and/or HCV from January 1, 2014 – December 31, 2019. We estimated trends over time in HCV co-diagnoses among a cohort of PLWH by demographic characteristics and HIV care outcomes. From 2014 – 2019, 1,183 (3.8%) PLWH were co-diagnosed with HCV infection. During this time period, the percentage of PLWH newly co-diagnosed with HCV increased by 243%, from 7% to 24% (β = 0.03, P for trend <0.01) among persons born during 1980-1989, and by 900%, from 1% to 10% (β = 0.01, P for trend <0.01) among persons born in 1990 or later. The percentage of PLWH newly co-diagnosed with HCV increased by 42%, from 43% to 61% (β = 0.03, P for trend <0.01) among persons with male-to-male sexual contact (MSM).

Overall, targeted interventions for HIV/HCV prevention, diagnosis and linkage to treatment are needed to reduce incidence of new infections among high-risk groups (i.e., PWID, younger populations, PLWH and MSM). Use of novel data sources including linking surveillance and clinical data can aid in informing these strategies.


File Upload Confirmation


Available for download on Monday, October 30, 2023