Abstract
Background
Understanding of repeat HIV testing (RHT) is limited and the impact of rural residence as a potential barrier to RHT is unknown. Rural populations are of particular interest in the Southeastern United States because of their disproportionate HIV burden.
Methods
We used HIV surveillance data from publicly funded HIV testing sites in North Carolina (NC) to assess repeat testing by transmission risk group and residential rurality in a retrospective cohort study. Linear binomial regression models were used to estimate adjusted, one-year cumulative incidences and cumulative incidence differences comparing RHT within transmission risk populations by level of rurality.
Results
In our total study population of 600,613 persons, 19,275 (3.2%) and 9,567 (1.6%) self-identified as men who have sex with men (MSM) and persons who inject drugs (PWID), respectively. 13,723 (2.3%) resided in rural ZIP codes. MSM were most likely to repeat test (unadjusted, one-year cumulative incidence after an initial negative test = 16.4%), compared to PWID (13.2%) and persons who did not identify as either MSM or PWID (13.6%). The greatest effect of rurality was within PWID; the adjusted, one-year cumulative incidence of RHT was 6.4 (95% CI: 1.4, 11.4) percentage points higher among metropolitan versus rural PWID.
Conclusions
One-year cumulative incidence of RHT was low among all clients of publicly funded HIV testing sites in NC, including MSM and PWID for whom annual testing is recommended. Our findings suggest a need for public health efforts to increase access to and support for RHT, particularly among rural PWID.
Keywords: HIV testing, rurality, repeat testing, annual testing, surveillance
INTRODUCTION
An estimated 15% of persons living with HIV (PLWH) in the United States (US) remain unaware of their infection (1), despite national recommendations that those at high risk of infection be tested at least annually (2, 3). High-risk populations include persons who inject drugs (PWID) and their sex partners, persons who practice transactional sex, sex partners of PLWH, men who have sex with men (MSM), and persons who are infected with or are seeking testing for sexually transmitted infections (2, 3). Regular HIV testing allows for early detection of infection and serves as the gateway to care, treatment, and viral suppression (1, 4). However, current understanding of testing patterns is based on analyses restricted to cross-sectional data, self-reported testing histories, and/or individual risk groups (5–9).
Residence in a rural area is an important barrier to HIV testing among at-risk populations due to transportation challenges, reduced numbers or availability of nearby HIV testing sites, and increased perceived stigma towards PLWH (10, 11). Although rural populations test later in the HIV disease course and have lower lifetime and past-year testing prevalences than urban populations (10, 12), little is known about repeat HIV testing behaviors by rurality. Any effects of rurality on repeat HIV testing may be especially relevant in the Southeastern US, where the proportion of the rural population living with HIV is three times that of other regions (13).
We conducted a surveillance-based analysis of HIV testing patterns and the effects of rurality on repeat testing in North Carolina (NC), where 20% of new HIV diagnoses occur in rural regions, compared with 8.9% of new diagnoses across the US (14). We aimed to characterize the one-year cumulative incidence of repeat HIV testing by transmission risk group and to estimate the effect of rurality on repeat HIV testing within each population.
MATERIALS AND METHODS
Study Design, Setting, and Population
We conducted a retrospective cohort study of adults with a negative result on an initial (index) HIV test at any publicly funded testing site in NC between July 2005 and December 2012, examining surveillance records in the subsequent year to estimate the cumulative incidence of a second HIV test at a publicly funded site within one year of the index test.
Data were drawn from the NC Division of Public Health Counseling, Testing, and Referral Services (CTRS) database, which tracks conventional (non-point-of-care) HIV tests performed at 199 publicly funded testing sites in NC. These sites include health departments – which house HIV testing sites and STD, family planning, and obstetrics & gynecology clinics – as well as community-based organizations, community health centers, drug treatment facilities, and county jails. The CTRS database contains data on more than 200,000 unique tests each year and is used for program planning and surveillance at the state level. CTRS data include test setting and client-reported demographics, residential ZIP code and county, HIV testing history, risk factors for HIV infection, and reasons for testing. These data are recorded by HIV testing personnel and sent to the NC State Laboratory of Public Health for entry into the CTRS database. Testing events between July 1, 2005 and December 31, 2013 were linked by unique client in a prior study (15), allowing reconstruction of individual testing histories.
We excluded testing clients who reported being pregnant at the index test because NC prenatal care guidelines require opt-out HIV testing at the first visit and during the third trimester of pregnancy (16), and our interest was in repeat testing behaviors outside of this context. We also excluded clients who did not report a ZIP code and county of residence in NC at the index test, clients with a positive result on the index test, pediatric clients (age < 18) and those missing age, and clients with an index test after December 31, 2012.
Outcome, Exposure, and Covariate Definitions
Our outcome of interest was repeat testing, which we defined as any HIV test captured by CTRS within one year following the index test to correspond with annual testing recommendations (2, 3). Outcome assessment was based only on the first repeat test, as the intent of the current analysis was only to determine whether testing occurred again within a year; description of longitudinal testing patterns will be the focus of future work. We relaxed the assessment period to fifteen months in a sensitivity analysis to allow flexibility in assessing testing guideline adherence. In the main analysis, we included repeat tests performed in the context of pregnancy in order to avoid censoring by the competing risk of pregnancy. We performed an additional sensitivity analysis to assess the impact of this decision, excluding all repeat tests conducted during a reported pregnancy or at a pregnancy/obstetrics testing site.
We based the main exposure, rurality of residence, on client-reported residential ZIP code at the index test, collapsing the U.S. Department of Agriculture’s 2010 Rural-Urban Commuting Areas (RUCA) classifications (17) into metropolitan (RUCA score = 1–3), micropolitan (RUCA score = 4–9), and rural (RUCA score = 10). The RUCA scheme classifies census tracts into ten levels of rurality/urbanicity based on measures of population density, urbanization, and daily commuting, and then further assigns rurality at the ZIP code level according to census tract classifications within each ZIP code. In a sensitivity analysis, we assigned rurality of residence based on residential county under the 2013 NCHS Urban-Rural Classification Scheme (URCS) (18). This framework classifies counties according to the Office of Management and Budget’s 2013 delineation of metropolitan and micropolitan statistical areas within each county. We collapsed the URCS scores in an analogous manner to the RUCA classifications defined above as metropolitan, micropolitan, and rural.
All stratification and confounder variables were generated from client-reported values at the index test. We created four risk group categories according to MSM and PWID status: clients identifying as MSM but not PWID (MSM-only), clients identifying as PWID but not MSM (PWID-only), clients identifying as both MSM and PWID (MSM+PWID), and clients not identifying as either MSM or PWID (neither MSM nor PWID). In multivariable models (see below), we dichotomized gender as female and male because there were no rural testers who identified as transgender. We coded race/ethnicity with disjoint indicators as White/non-Hispanic, Black/non-Hispanic, Hispanic, Asian or Pacific Islander/non-Hispanic, Native American/non-Hispanic, and Other/Multi-racial, with White/non-Hispanic serving as the referent group. Age was represented with linear and quadratic terms in multivariable models.
Statistical Analyses
We first examined covariate distributions overall and by the full set of four transmission risk groups (MSM-only, PWID-only, MSM+PWID, neither MSM nor PWID). Due to small numbers of MSM+PWID, we conducted all subsequent analyses with a reduced set of three transmission risk group categories: MSM, PWID, and neither MSM nor PWID. Clients who reported being both MSM and PWID (MSM+PWID) were included both as MSM in analyses restricted to MSM and as PWID in analyses restricted to PWID.
We calculated unadjusted one-year cumulative incidences of repeat testing in each of the transmission risk groups in our reduced set (MSM, PWID, and neither MSM nor PWID) by rurality of residence and several relevant covariates. To assess the effect of rurality on repeat testing within each group, we then constructed distinct linear binomial regression models tailored to each population. We selected minimum sufficient adjustment sets using directed acyclic graphs (19) following a comprehensive literature review, adjusting for age and race/ethnicity in the model restricted to MSM; for age, race/ethnicity, gender, and MSM status in the model restricted to PWID; and for age, race/ethnicity, and gender in the model restricted to neither MSM nor PWID. These models were used to estimate the adjusted, one-year cumulative incidences and 95% confidence intervals (95% CIs) of repeat HIV testing by level of rurality within each population, along with the corresponding cumulative incidence differences (CIDs) and 95% CIs comparing micropolitan and rural residents (separately) to metropolitan residents within each population.
The Institutional Review Board at the University of North Carolina at Chapel Hill approved this secondary analysis of public health surveillance data. All statistical analyses were conducted in SAS 9.4 (SAS Institute Inc., Cary, NC).
RESULTS
Of 858,501 persons testing during the study period, 600,613 were eligible for analysis (Figure 1). The majority were female (57%) (Table 1). Clients in the “neither MSM nor PWID” group were most likely to be Black/non-Hispanic (46%), while MSM-only, PWID-only, and MSM+PWID clients were most likely to be White/non-Hispanic. PWID-only clients tended to be older compared to other testing populations, and the percentage of white/non-Hispanic clients was particularly high (70%) in this population. Testing site type varied widely across risk groups, although STD testing and treatment sites accounted for the largest number of index tests in each group. The vast majority of repeat tests across all risk groups occurred at the same testing site type as a client’s initial test. A large majority of all clients resided in metropolitan ZIP codes at index testing, with slightly larger proportions of MSM-only, PWID-only, and MSM+PWID clients residing in metropolitan areas compared to neither MSM nor PWID clients. MSM and PWID made up 3.2% and 1.6% of testing clients, with MSM+PWID clients contributing to both populations. Transmission risk group, including MSM and PWID status, was missing for 8.3% of male clients and 10.0% of all clients, respectively, suggesting underreporting of these identifiers. No other substantial missingness was present.
Figure 1.
Application of inclusion and exclusion criteria to form study population. Of 858,501 testing clients, we excluded pediatric clients and those missing age, clients with an index test after December 31, 2012, clients with a positive result on the index test, clients who reported being pregnant at the index test, and clients who did not report a NC ZIP code and county of residence at the index test, for a total of 600,613 included clients.
Table 1.
Characteristics by transmission risk group of persons undergoing an HIV test at publicly funded testing sites in North Carolina at first (index) visit, 2005–2012.
| Characteristics | MSM*-only N (%)‡ |
PWID†-only N (%)‡ |
MSM+PWID N (%)‡ |
Neither MSM nor PWID N (%)‡ |
Tota N (%)‡ |
|---|---|---|---|---|---|
| Gender | |||||
| Male | 18252 (100) | 4642 (55) | 1023 (100) | 234214 (41) | 258131 (43) |
| Female | 0 (0) | 3755 (45) | 0 (0) | 334469 (59) | 338224 (57) |
| Transgender§ | 0 (0) | 1 (0) | 0 (0) | 131 (0) | 132 (0) |
| Race/Ethnicity | |||||
| White, non-Hispanic | 8598 (49) | 5790 (70) | 474 (48) | 189012 (35) | 203874 (36) |
| Black, non-Hispanic | 6423 (37) | 1841 (22) | 376 (38) | 251104 (46) | 259744 (45) |
| Hispanic | 2012 (12) | 510 (6) | 124 (13) | 89298 (16) | 91944 (16) |
| Asian/Pacific Islander | 230 (1) | 43 (1) | 5 (1) | 7924 (2) | 8202 (1) |
| American Indian/ Alaska Native | 156 (1) | 112 (1) | 6 (1) | 7284 (1) | 7558 (1) |
| Other/multi-racial | 17 (0) | 7 (0) | 0 (0) | 644 (0) | 668 (0) |
| Age | |||||
| 18–24 | 7843 (43) | 2310 (27) | 344 (34) | 232117 (41) | 242614 (40) |
| 25–34 | 5129 (28) | 2728 (32) | 301 (29) | 174227 (30) | 182385 (30) |
| 35–44 | 2868 (16) | 1803 (21) | 176 (17) | 92327 (16) | 97174 (16) |
| ≥45 | 2412 (13) | 1703(20) | 202 (20) | 74123 (13) | 78440 (13) |
| Initial Testing Site | |||||
| LHD/CBO¶ | 2516 (15) | 644 (8) | 70 (7) | 30759 (6) | 33989 (6) |
| STD Testing & Treatment | 8878 (51) | 2310 (28) | 517 (53) | 254954 (47) | 266659 (47) |
| Drug Treatment | 206 (1) | 1471 (18) | 54 (6) | 7710 (1) | 9441 (2) |
| Family Planning/Sexual Health | 158 (1) | 660 (8) | 2 (0) | 119769 (22) | 120589 (21) |
| CHC/PHC\\ | 638 (4) | 220 (3) | 30 (3) | 16166 (3) | 17054 (3) |
| Jail | 362 (2) | 1341 (16) | 63 (7) | 34200 (6) | 35966 (6) |
| Other | 4631 (27) | 1572 (19) | 232 (24) | 80690 (15) | 87125 (15) |
| Repeat Test at Same Site Type** | |||||
| Yes | 6572 (97) | 2901 (98) | 356 (98) | 207637 (99) | 595786 (99) |
| No | 204 (3) | 62 (2) | 8 (2) | 1104 (1) | 4827 (1) |
| Rurality of residence | |||||
| Metropolitan | 15194 (83) | 6833 (80) | 813 (80) | 435531 (76) | 458371 (76) |
| Micropolitan | 2827 (16) | 1491 (18) | 186 (18) | 124015 (22) | 128519 (21) |
| Rural | 231 (1) | 220 (3) | 24 (2) | 13248 (2) | 13723 (2) |
| Repeat HIV testing | |||||
| No repeat test | 11476 (63) | 5581 (65) | 659 (64) | 364053 (64) | 381769 (64) |
| Repeat test ≤ 1 year | 3361 (18) | 1247 (15) | 142 (14) | 88106 (15) | 92856 (16) |
| Repeat test > 1 year | 3415 (19) | 1716 (20) | 222 (22) | 120635 (21) | 125988 (21) |
|
| |||||
| Total | 18252 | 8544 | 1023 | 572794 | 600613 |
Men who have sex with men (MSM);
Persons who inject drugs (PWID);
Missing values are not included in calculation of percentage;
Includes male-to-female and female-to-male transgender clients;
Local health department/community based organization;
Community health center/primary health care;
Includes first repeat test following the index test, regardless of time elapsed between tests. Percentage is out of only those with at least one repeat test.
In univariate analyses, 16.4% of MSM clients returned for repeat testing within one year of index testing, compared to 13.2% of PWID clients and 13.6% of neither MSM nor PWID clients. We refer the reader to Supplemental Table 1 for a summary of these cumulative incidences by selected covariates.
Across all sub-populations of interest, metropolitan MSM and PWID clients were most likely to return for repeat testing (adjusted, one-year cumulative incidence: 0.170, 95% CI: 0.16, 0.18 in both populations) and rural PWID clients were the least likely (adjusted, one-year cumulative incidence: 0.102, 95% CI: 0.05, 0.15) (Table 2). Among MSM, rural and micropolitan clients showed adjusted, one-year cumulative incidences of repeat testing that were 2.4 (95% CI: −2.0, 6.8) and 3.2 (95% CI: 1.9, 4.6) percentage points lower than those of their metropolitan peers, respectively. Similar trends were observed among PWID and neither MSM nor PWID. The greatest effect of rurality on repeat testing was within PWID; the one-year cumulative incidence of repeat testing was 6.4 (95% CI: 1.4, 11.4) percentage points lower among rural versus metropolitan PWID.
Table 2.
Adjusted one-year cumulative incidence of repeat HIV testing within transmission risk group of persons undergoing an HIV test at publicly funded testing sites in North Carolina by rurality of zip code of residence at first (index) visit, 2005–2012.
| MSM*§ | PWID†|| | Neither MSM nor PWID¶ | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| Rurality of residence | Cumulative Incidence (95% CI) | CID‡ (95% CI)** | Cumulative Incidence (95% CI) | CID‡ (95% CI)** | Cumulative Incidence (95% CI) | CID‡ (95% CI)** |
|
|
||||||
| Metropolitan | 0.17 (0.16, 0.18) | -- | 0.17 (0.16, 0.18) | -- | 0.16 (0.15, 0.16) | -- |
| Micropolitan | 0.14 (0.13, 0.15) | −0.03 (−0.05, −0.02) | 0.13 (0.11, 0.15) | −0.03 (−0.05, −0.01) | 0.14 (0.13, 0.14) | −0.02 (−0.02, −0.02) |
| Rural | 0.15 (0.10, 0.19) | −0.02 (−0.07, 0.02) | 0.10 (0.05, 0.15) | −0.06 (−0.11, −0.01) | 0.11 (0.11, 0.12) | −0.04 (−0.04, −0.04) |
| Unadjusted Total | 0.164 | 0.132 | 0.136 | |||
Men who have sex with men (MSM);
Persons who inject drugs (PWID);
Cumulative incidence difference (CID);
Model conducted only among MSM (including MSM+PWID) clients and adjusts for age and race/ethnicity;
Model conducted only among PWID (including MSM+PWID) clients and adjusts for age, race/ethnicity, gender, and MSM status;
Model conducted only among neither MSM nor PWID clients and adjusts for age, race/ethnicity, and gender ;
Referent group is metropolitan clients within transmission risk group (MSM, PWID, or neither MSM nor PWID)
Sensitivity analyses allowing fifteen months of observation showed slightly larger adjusted cumulative incidences of repeat testing across all populations and levels of rurality (Supplemental Table 2), reflecting an uptick in repeat testing between 12 and 15 months (Supplemental Figure 1). The increased cumulative incidences were smaller among PWID than the other testing populations. CIDs in the fifteen-month analyses were very similar to those in the main analyses. Sensitivity analyses with rurality assigned by county of residence at the index test shifted approximately 4% of clients from the micropolitan classification to both the metropolitan and rural classifications and yielded effect estimates within approximately one percentage point of the absolute effect estimates identified in the primary analyses (Supplemental Table 3). A final sensitivity analysis excluding all repeat tests that occurred during pregnancy or were conducted at pregnancy/obstetrics testing sites showed slightly lower cumulative incidences among PWID and neither MSM nor PWID clients, but CIDs were again similar to the main results (Supplemental Table 4).
DISCUSSION
Although regular HIV testing among at-risk groups is essential to maximize clinical and public health benefits, little is known about repeat HIV testing based on objective measures across multiple risk groups, and the effect of rurality as a potential barrier to HIV testing is unknown. In this study, we examined these measures and effects using surveillance data from publicly funded testing sites in a large, diverse population. Despite recommended annual testing for all high-risk individuals, including MSM and PWID (2, 3), the one-year cumulative incidence of repeat testing in our setting was only 16% among MSM and 13% among PWID. After inclusion of repeat tests that occurred within 15 months of the index test, cumulative incidence remained suboptimal (21% among MSM and 18% among PWID). These results indicate a clear need for public health efforts to increase access to and support for regular HIV testing in NC.
The rural/urban disparities that we observed in repeat HIV testing mirror disparities that have been identified previously with other HIV testing metrics (11–14, 20–24). Lifetime testing prevalence estimates range from 43.6% among residents of the most urban areas of the US to 32.2% among residents of the most rural areas, while past-year testing prevalence ranges from 13.5% to 7.3% in those areas, respectively (14). Of note, our estimated one-year cumulative incidences of testing among persons who had already had a negative HIV test are slightly higher at all rurality levels than corresponding past-year estimates calculated irrespective of prior testing status (14), suggesting that previously reported rural-urban disparities in testing prevalences and disease stage at diagnosis (14, 23) may be largely due to differences in first-time testing behaviors rather than repeating testing behaviors. Although repeat HIV testing should be encouraged among all at-risk clients, our results indicate that interventions aiming to reduce rural-urban disparities within the full population may be best targeted towards introduction of new rural testing clients to the existing HIV testing infrastructure in NC.
Although metropolitan MSM clients were more likely to repeat-test within one year than any other testing population, the one-year cumulative incidence of repeat HIV testing among MSM clients in NC across all levels of rurality was lower than might be expected from previous estimates. In one study among HIV-uninfected Black MSM ages 18–30 in NC, 83% reported ever receiving an HIV test. Among those subjects who had previously tested, 71% of this particularly at-risk population reported testing at least once every six months and 40% reported testing at least quarterly (8). Further, 21% of a sample of MSM with HIV-negative primary partners in the US reported testing at least once every six months and 29% reported testing at least once per year (11). The apparent discrepancies between our results and those of prior studies may be due to differences in study populations and measures, our capture of only those tests performed at publicly-funded testing sites, and/or earlier studies’ reliance on self-reported HIV testing.
Although rural and micropolitan populations represent a minority of the total, at-risk, and testing populations in NC, the substantial burden placed on rural areas by the recent opioid epidemic heightens the need for improved understanding of rural HIV testing behaviors, especially among PWID. Only 10% of PWID in rural areas in our setting returned for repeat testing within a year of their index test, a substantially smaller percentage compared to their metropolitan peers. This difference may be due to greater accessibility of resources for PWID in metropolitan compared to micropolitan, and particularly, rural areas. A lack of harm reduction resources for PWID, including HIV testing, has previously been linked to anomalous HIV outbreaks in rural areas (25), highlighting the potential hazard of infrequent HIV testing among rural PWID.
More generally, the low one-year cumulative incidence of repeat testing observed among PWID overall underscores a need for increased repeat HIV testing among PWID in NC, regardless of region of residence. To this end, the 2016 legalization of syringe service programs (SSPs) in NC offers improved opportunities to reach PWID with harm reduction services, including HIV testing (26). Although we did not examine hepatitis C virus (HCV) testing, our results may have similar implications for the provision of HCV testing, which has substantial overlap with HIV testing in NC. Integration of point-of-care HIV and HCV testing into SSPs, just as these tests have already been introduced into many other non-traditional HIV testing sites, could improve early detection and treatment of both diseases among the high-risk PWID populations served by these programs. Integration of services may be particularly beneficial in rural regions where transportation is an obstacle to accessing testing and other harm reduction resources (24).
Our reliance on CTRS data results in underestimation of the true one-year cumulative incidence of repeat testing because secondary tests performed outside of publicly funded testing sites in NC and point-of-care secondary tests are not captured. There is a clear need for increased collaboration, including data sharing, between public and private organizations offering testing services in order to fully characterize HIV testing patterns in NC. The reliance on CTRS data may also bias rurality effect estimates if the likelihood of receiving a secondary test at a publicly vs. privately funded testing site is related to rurality. In one prior study, residents of rural areas were more likely to report testing in a hospital, while residents of urban areas were more likely to report testing at publicly funded testing centers (12). Exclusion of persons with missing age or county/ZIP of residence may have resulted in some selection bias, but it is difficult to speculate on the magnitude or direction of any such bias.
Substantial missing data for both MSM and PWID risk behaviors indicate that both factors were likely underreported, potentially in part due to stigma, which limits inference from our findings in these groups to those who disclose MSM and PWID status to testing personnel. One recent study of acute HIV cases diagnosed in NC showed substantially increased reporting of MSM behaviors following diagnosis compared to reporting at the time of testing (27), supporting the likelihood that some MSM and PWID were misclassified in our analysis as not belonging to either group. Additionally, the small number of rural MSM and PWID clients resulted in comparatively imprecise estimates among these subpopulations, and we were unable to separately examine repeat testing in the population of MSM+PWID clients. Further, this analysis is unable to distinguish between true first-time testing and those who had received HIV tests at publicly funded testing sites prior to initiation of data collection, as well as confirmatory repeat testing. Only 1.5% of repeat tests occurred within one month of the index test, suggesting that testers seeking confirmation of prior negative results had a minimal effect on our estimates. Finally, CTRS data are limited in scope and were collected for surveillance rather than research. Thus, there may be uncontrolled confounding by variables not available in the database, including socioeconomic status. The restricted nature of CTRS data also limited us to evaluating populations of testing clients in broad strokes instead of by more nuanced individual risk factors for infection.
Conclusions
One-year cumulative incidence of repeat HIV testing was low among all clients of publicly-funded testing centers in NC, including populations of MSM and PWID in whom annual testing is recommended. This study also suggests that the effect of rurality on one-year cumulative incidence of repeat HIV testing is substantively small overall and among MSM in North Carolina. However, this effect is slightly more pronounced within PWID. There is a need for increased repeat HIV testing, as well as general STD screening, across all at-risk testing populations in NC, with perhaps the greatest need among rural populations of PWID.
Supplementary Material
Acknowledgments
Sources of Support: Supported by an administrative supplement to the UNC Center for AIDS Research, an NIH-funded program (P30 AI50410), and an NIH K23 award (K23MH099941).
This research was supported by an administrative supplement to the UNC Center for AIDS Research, an NIH-funded program (P30 AI50410), and an NIH K23 award (K23MH099941). All data used in this study were provided by the NC Department of Health and Human Services, Division of Public Health (DPH). The NC DPH does not take responsibility for the scientific validity or accuracy of methodology, results, statistical analyses, or conclusions presented.
Footnotes
Conflicts of Interest: The authors declare that they have no conflicts of interest.
References
- 1.Centers for Disease Control and Prevention. Monitoring selected national HIV prevention and care objectives by using HIV surveillance data – United States and 6 dependent areas, 2015. HIV Surveillance Supplemental Report. 2017;22(2) http://www.cdc.gov/hiv/library/reports/hiv-surveillance.html. [Google Scholar]
- 2.Branson BM, Handsfield HH, Lampe MA, et al. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. Morb Mortal Wkly Rep. 2006;55:1–17. [PubMed] [Google Scholar]
- 3.Moyer VA. Screening for HIV: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2013;159:51–60. doi: 10.7326/0003-4819-159-1-201307020-00645. [DOI] [PubMed] [Google Scholar]
- 4.Ambrosioni J, Calmy A, Hirschel B. HIV treatment for prevention. J Int AIDS Soc. 2011;14:28. doi: 10.1186/1758-2652-14-28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.North Carolina HIV/STD Surveillance Unit. 2015 North Carolina HIV/STD surveillance report. North Carolina Department of Health and Human Services; Raleigh, North Carolina: 2016. [Google Scholar]
- 6.Hoenigl M, Anderson CM, Green N, et al. Repeat HIV-testing is associated with an increase in behavioral risk among men who have sex with men: A Cohort Study. BMC Med. 2015;13:218–227. doi: 10.1186/s12916-015-0458-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hightow LB, Miller WC, Leone PA, et al. Predictors of repeat testing and HIV seroconversion in a sexually transmitted disease clinic population. Sex Transm Dis. 2004;31:455–459. doi: 10.1097/01.olq.0000135984.27593.5f. [DOI] [PubMed] [Google Scholar]
- 8.Hurt CB, Soni K, Miller WC, et al. Human immunodeficiency virus testing practices and interest in self-testing options among young, black men who have sex with men in North Carolina. Sex Transm Dis. 2016;43:587–593. doi: 10.1097/OLQ.0000000000000484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Centers for Disease Control and Prevention. HIV infection risk, prevention, and testing behaviors among men who have sex with men—national HIV behavioral surveillance, 20 U.S. Cities, 2014. [Accessed September 9, 2017];HIV Surveillance Special Report. :15. Available at: http://www.cdc.gov/hiv/library/reports/surveillance/#panel2. Published January 2016.
- 10.Mitchell JW, Horvath KJ. Factors associated with regular HIV testing among a sample of US MSM with HIV-negative main partners. J Acquir Immune Defic Syndr. 2013;64:417–423. doi: 10.1097/QAI.0b013e3182a6c8d9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.McKinney MM. Variations in rural AIDS epidemiology and service delivery models in the United States. J Rural Health. 2002;18:455–466. doi: 10.1111/j.1748-0361.2002.tb00910.x. [DOI] [PubMed] [Google Scholar]
- 12.Gonzalez A, Miller CT, Solomon SE, et al. Size matters: community size, HIV stigma, & gender differences. AIDS Behav. 2008;13:1205–1212. doi: 10.1007/s10461-008-9465-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Irene HH, Li J, Mckenna MT. HIV in predominantly rural areas of the United States. J Rural Health. 2005;21(3):245–253. doi: 10.1111/j.1748-0361.2005.tb00090.x. [DOI] [PubMed] [Google Scholar]
- 14.Ohl ME, Perencevich E. Frequency of human immunodeficiency virus (HIV) testing in urban vs. rural areas of the United States: results from a nationally-representative sample. BMC Public Health. 2011;11:681–687. doi: 10.1186/1471-2458-11-681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cope A. Diss. University of North Carolina; Chapel Hill: 2014. Assessment of HIV transmission and diagnosis patterns in North Carolina. [Google Scholar]
- 16.North Carolina Division of Health and Human Services. Rule 10A NCAC 41A.0202.
- 17.Documentation. USDA Economic Research Service; Oct 12, 2016. [Accessed September 9, 2017]. 2010 rural-urban commuting area (RUCA) codes. Available at: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/ [Google Scholar]
- 18.Ingram DD, Franco SJ. 2013 NCHS urban-rural classification scheme for counties. National Center for Health Statistics. Vital Health Stat. 2014;2(166):1–73. [PubMed] [Google Scholar]
- 19.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10(1):37–48. [PubMed] [Google Scholar]
- 20.Weissman S, Duffus WA, Vyavaharkar M, et al. Defining the rural HIV epidemic: correlations of 3 definitions-South Carolina, 2005–2011. J Rural Health. 2013;30(3):275–283. doi: 10.1111/jrh.12057. [DOI] [PubMed] [Google Scholar]
- 21.Reif S, Golin CE, Smith SR. Barriers to accessing HIV/AIDS care in North Carolina: rural and urban differences. AIDS Care. 2005;17(5):558–565. doi: 10.1080/09540120412331319750. [DOI] [PubMed] [Google Scholar]
- 22.Heckman TG, Somlai AM, Peters J, et al. Barriers to care among persons living with HIV/AIDS in urban and rural areas. AIDS Care. 1998;10(3):365–375. doi: 10.1080/713612410. [DOI] [PubMed] [Google Scholar]
- 23.Weis KE, Liese AD, Hussey J, et al. Associations of rural residence with timing of HIV diagnosis and stage of disease at diagnosis, South Carolina 2001–2005. J Rural Health. 2010;26(2):105–112. doi: 10.1111/j.1748-0361.2010.00271.x. [DOI] [PubMed] [Google Scholar]
- 24.Lopes BLW, Eron JJ, Jr, Mugavero MJ, et al. HIV care initiation delay among rural residents in the Southeastern United States, 1996 to 2012. J Acquir Immune Defic Syndr. 2017;76(2):171–176. doi: 10.1097/QAI.0000000000001483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Conrad C, Bradley HM, Broz D, et al. Community outbreak of HIV infection linked to injection drug use of Oxymorphone – Indiana, 2015. Morb Mortal Wkly Rep. 2015;64(16):443–444. [PMC free article] [PubMed] [Google Scholar]
- 26.General Assembly of North Carolina. Senate Bill 794: Authorize needle exchange programs. 2016. [Google Scholar]
- 27.Kuruc JD, Cope AB, Sampson LA, et al. Ten years of screening and testing for acute HIV infection in North Carolina. J Acquir Immune Defic Syndr. 2016;71(1):111–119. doi: 10.1097/QAI.0000000000000818. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.