Background: The rate of new HIV diagnoses has ceased to increase over the last decade, yet those rates have remained stable but not decreasing. Paramount to HIV prevention is the accurate and systematic surveillance systems capturing timely reports of new HIV diagnoses. This study seeks to determine whether differences between county-level reported new HIV diagnoses can be explained by demographic factors.
Methods: Using publicly available HIV diagnosis data from 2008-2011 and socio-demographic factors potentially associated with HIV, we created models stratified by Maryland and North Carolina combined and Virginia. We used coefficients from the models to estimate new HIV diagnoses in Virginia and Maryland/North Carolina.
Results: The 134 counties of Virginia had 4,466 new HIV diagnoses from 2008-2011, with 3,804 occurring in 29 counties with 20 or more total cases (unsuppressed). The 122 counties of Maryland/North Carolina combined had 14,400 cases with 13,854 occurring in 70 unsuppressed counties. Our final reduced models fit Maryland/North Carolina and Virginia. After mapping the expected HIV diagnoses, we found that diagnoses in 6 counties on the Virginia border with North Carolina had the largest differences. These counties had reported between 0 and 20 diagnoses (suppressed counties), but after estimation have 20-50 HIV diagnoses each.
Conclusion: After accounting for social demographic characteristics of the counties, the Virginia socio-demographic data predicted different diagnoses counts than what was reported, indicating the significance of unmeasured factors. We hypothesize that the unmeasured factors are underreporting or HIV diagnosis issues. Moving forward, additional research should be conducted to assess the extent of reporting bias and determine the steps to mediate the problem.
Table of Contents
TABLE OF CONTENTS
CHAPTER 1: LITERATURE REVIEW 1
CHAPTER 2: MANUSCRIPT 8
TABLES AND FIGURES 20
CHAPTER 3: PUBLIC HEALTH IMPLICATIONS AND SUGGESTIONS 27
About this Master's Thesis
|Committee Chair / Thesis Advisor|
|Using County-Level Socio-Demographics to Estimate HIV Diagnoses in Maryland, North Carolina and Virginia ()||2018-08-28||