A study using the 2015 Behavioral Risk Factor Surveillance System found that rural residents of the United States had lower odds of lifetime and past-year human immunodeficiency virus testing.
In 2014, the Joint United National Program on HIV/AIDS released the 90–90–90 targets which stated that by 2020, 90% of people living with HIV will know their serostatus, 90% of all HIV-positive individuals will receive treatment, and 90% of all HIV-positive individuals will achieve viral suppression.1 In the United States, an estimated 15% of the 1.1 million people living with HIV in 2014 were undiagnosed,2 indicating that HIV testing remains a significant barrier to achieving these targets. The Centers for Disease Control and Prevention (CDC) estimates that 40% of new HIV infections each year are transmitted by people who are unaware of their serostatus.3 Therefore, early detection of HIV is key to decreasing morbidity and mortality of the disease,4 and to preventing future transmission.3
In an effort to include HIV testing as part of routine health care, the CDC issued new HIV testing guidelines in 2006 that recommended opt-out HIV screening in health care settings for all individuals ages 13-64.5 After 2006, 1 study reported that there was an immediate increase in HIV testing rates, but it was not sustained over time.6 This finding is supported by other studies that reported a decreasing temporal trend in HIV testing rates among high-risk men,7 young adults,8 and among Georgia residents.9 These studies, however, largely concentrated on HIV testing in urban areas of the United States.
Although the HIV epidemic still predominately affects urban areas, the incidence of HIV in rural areas has been increasing over the past decade, particularly in southern and midwestern states of the United States.10 In addition, rural individuals face several disparities along the HIV continuum of care, including being less likely to report lifetime HIV testing and more likely to be diagnosed with HIV at a later stage of infection than those who live in urban areas.11 Before 2006, lifetime testing rates for HIV among rural residents were estimated to be 32.2% compared to 43.6% in urban residents.12 There is evidence that the opt-out methods can reverse the disparities in rural HIV testing rates; a 2011 study reported that an emergency department serving a semirural population had a 91% acceptance rate in an opt-out screening program.13 However, other studies show that only 10% of physicians adhered to the CDC's guidelines.14
In the past few years, there have been limited nationally-representative studies evaluating the frequency of HIV testing in rural versus urban areas. Thus, there is a need to evaluate the effectiveness of recent HIV screening guidelines, particularly in rural populations. The purpose of this study was to (1) measure and compare self-reported lifetime and past-year HIV testing in rural versus urban populations of the United States and (2) determine if differences exist in testing site location between urban and rural residents.
MATERIALS AND METHODS
Data Source and Collection
Data from the 2015 Behavioral Risk Factor Surveillance System (BRFSS) were analyzed. The BRFSS is a cross-sectional survey administered annually by the CDC to all 50 US states, the District of Columbia, and 3 US territories (Puerto Rico, Guam, and the US Virgin Islands). The purpose of the survey was to collect data on people's health-related risk behaviors, chronic health conditions, and use of preventive services.15 The survey was administered using random-digit-dial sampling methods to landline and cellular telephones. State-level data are combined and weighted to create a national representative sample of civilian, noninstitutionalized US adults 18 years and older. The response rates for all states and the District of Columbia in 2015 was 47.2%, ranging from 33.9% to 61.1%.3 The BRFSS data are publicly available for secondary analysis at: https://www.cdc.gov/brfss/annual_data/annual_data.htm.
The 2 primary outcome variables, lifetime HIV testing and past-year HIV testing, were ascertained from the survey question “Have you ever been tested for HIV?” Response options included “yes,” “no,” “don't know/not sure,” or “refused.” For this analysis, data that indicated refusal to answer or uncertainty regarding prior testing were considered missing. This approach is consistent with prior analyses using CDC survey data of HIV testing.12 Respondents who answered “yes” for “ever been tested” were asked “In what month and year was your last HIV test?” If the month and year given by the respondent was less than 365 days before the date of the interview, they were considered to have been tested for HIV in the past year.
A secondary outcome variable included the location of a respondent’s last HIV test. Respondents who answered “Yes” for “ever been tested” were asked “Where did you have your last HIV test—at a private doctor or HMO office, at a counseling and testing site, in the emergency room, as an inpatient in a hospital, at a clinic, in a jail or prison, at a drug treatment facility, at home, or somewhere else?” For this analysis, receiving HIV testing in the emergency room (ER) or as an inpatient in a hospital was collapsed into a single category (“ER/hospital inpatient”). Similarly, receiving HIV testing at a counseling and testing site, at a clinic, in a jail or prison, at a drug treatment facility, at home, or somewhere else were collapsed into a single category (“clinic/other”). The final category for this variable was “private doctor/HMO”.
Geographic area (rural/urban) of the respondents was the main independent variable. Rural/urban designation was determined based on the metropolitan status code. A designation of “urban” was given if the metropolitan status code was “in the center city of a metropolitan statistical area (MSA)”, “outside the center city of an MSA but inside the county containing the center city”, or “inside a suburban county of the MSA”. A designation of “rural” was given if the metropolitan status code was “not in an MSA”.
Other sociodemographic variables that were adjusted for in the analysis included: sex (male, female), age group (18–34 years, 35–54 years, ≥ 55 years), race and ethnicity (non-Hispanic black, non-Hispanic white, or other), marital status (married/unmarried couple, never married, divorced/separated/widowed), education (high school or less, some post-high school, college graduate), household income (< US $25,000, US $25,000–50,000, ≥ US $50,000), census region (northeast, midwest, south, west), health insurance (yes/no), and sexual or gender minority (yes/no).
The proportion of missing data for sociodemographic variables was less than 2%, except for household income and sexual/gender minority status. Because only 22 states completed the optional sexual orientation and gender identity module in 2015, the sexual orientation/gender identity variable was removed from further analysis. A sensitivity analysis revealed no differences in outcomes after the exclusion of this variable. The proportion of missing data for outcome variables was less than 1%, except for lifetime HIV testing. A χ2 test was performed to compare the characteristics of individuals who had missing lifetime HIV testing data compared with those who did not. To reduce potential bias, missing values were treated as a separate category in the outcome variables in the analysis.
All analyses were conducted in SAS System version 9.4 (SAS Institute Inc., Cary, NC). All associations were deemed statistically significance with P less than 0.05.
To account for unequal probability of respondent selection, nonresponse, and telephone noncoverage, the data were weighted using a final weight included with the BRFSS data set. Bivariate analyses were used to determine differences between rural and urban residents in demographic characteristics (age, sex, race/ethnicity, census region, education, income, marital status), presence of health insurance, and self-reported lifetime and past-year HIV testing. The χ2 tests were used to compare the categorical measures. Multinomial logistic regression models were used to evaluate the association between lifetime and past-year HIV testing and respondents' urban/rural residence while controlling for the effect of other sociodemographic and predictor variables. Multinomial logistic regression was also used to compare HIV testing locations by urban/rural status. Finally, a χ2 test was used to assess differences in sociodemographics between rural residents who tested at a private doctor/HMO, hospital/ER, or a clinic.
The population with missing data differed significantly from the population with no missing HIV testing data. A greater percentage of respondents with missing data were male, urban residents, with younger age (18–54 years), non-Hispanic Black and other race, single (never married, divorced, separated, or widowed), had high school education or less, had annual household income < US $25,000, lived in the northeast, south, or west, and did not have health insurance (see Supplemental Digital Content 1, http://links.lww.com/OLQ/A289 for table).
Differences in respondents' sociodemographic characteristics, presence of insurance, and past-year and lifetime HIV testing frequencies by urban/rural status are summarized in Table 1. Among 250,579 respondents, 76,530 (30.5%) reported having a rural residence while 174,049 (69.5%) reported having an urban residence. Compared to urban individuals, rural individuals were more likely to be older, non-Hispanic white, live in the Midwest, and less likely to be never married, have a college degree, have income over US $50,000 per year, and less likely to report past-year or lifetime HIV testing. For urban residents, 26.9% reported ever having received an HIV test in their lifetime and 24.5% reporting having received an HIV test in the past year. On the other hand, 21.5% of rural residents reported having received an HIV test in their lifetime and 20.2% reported having been tested for HIV in the past year. Respondent sex and presence of health insurance did not differ between urban and rural residence (Table 1).
The results of the multinomial logistic regression analyses of lifetime and past-year HIV testing of rural compared with urban residents are shown in Table 2. After adjusting for sociodemographic and other predictor variables, living in a rural area was associated with lower odds of lifetime HIV testing compared with never testing (odds ratio [OR], 0.85; 95% confidence interval [CI], 0.81–0.90). Additionally, living in a rural area was associated with lower odds of having missing lifetime HIV testing data compared with never testing (OR, 0.83; 95% CI, 0.77–0.90). Rural residents also had lower odds than urban residents of reporting past-year HIV testing compared with not testing in the past-year (OR, 0.84; 95% CI, 0.74–0.95).
The type of HIV testing site differed for urban and rural respondents (P < 0.001). Over half (50.8%) of urban respondents reported receiving an HIV test at a private doctor or HMO, but only 42.2% of rural respondents reported receiving an HIV test at those locations. There were more rural respondents testing in hospital/ER (14.5%) or clinic/other locations (43.2%) compared with urban respondents (hospital/ER, 11.7%; clinic/other locations, 37.5%). In multinomial logistic regression, urban and rural respondents differed significantly in testing site (Table 2). Living in a rural area was associated with higher odds of HIV testing at the hospital/ER compared with a private doctor’s office/HMO (adjusted OR [aOR], 1.41; 95% CI, 1.23–1.62), and higher odds of testing at a clinic or other location compared to a private doctor’s office/HMO (aOR, 1.21; 95% CI, 1.02–1.24).
The sociodemographic characteristics of rural residents differed among the 3 testing site locations (Table 3). Rural residents who reported testing at a hospital/ER were more likely to be female, age 55 years or older, non-Hispanic white, have a high school diploma or less, have income less than US $25,000 per year, live in the South, and not have insurance. Notably, a larger proportion of male respondents and people of color reported testing at a clinic or other location compared to a private doctor/HMO or hospital/ER. Marital status did not differ significantly between respondents from each testing site location.
After adjusting for differences in sociodemographic variables, rural individuals had 15% lower odds of having been tested for HIV in their lifetime and 16% lower odds of reporting past-year testing. These findings are consistent with a prior study that used BRFSS data from 2005 and 2009.12 However, the frequency of lifetime HIV testing in this study is low compared with studies using other data sets. Although we found that 26.9% of urban and 21.5% of rural residents reported lifetime HIV testing, a prior study using the National Health and Nutrition Examination Survey found that lifetime HIV testing was 44.5% overall from 2007 until 2010.16 This disparity may reflect a diminishing enthusiasm for promoting the opt-out method of HIV testing.
We also found that although there were no differences in insurance coverage between rural and urban residents, individuals in rural areas were more likely to report receiving an HIV test at the hospital/ER or clinic compared with a private doctor. These results indicate a different uptake of HIV testing practices by providers compared with clinics in rural settings and highlight an opportunity to increase HIV testing in rural areas by targeting interventions toward rural providers. A 2011 study in a nonurban setting reported that only 10% of physicians adhered to the opt-out guidelines and cited barriers, such as policies, stigma, finances, attitudes toward testing, and patient acceptance.13 However, there is considerable data showing that HIV testing is rarely declined by patients, but often not done because it is not offered by providers.17,18 It is imperative that providers in rural areas routinely offer HIV testing to their patients. The population gap identified in this study can be used to support future communication strategies between health care providers and their patients to increasing HIV screening behaviors.
In addition to increasing the knowledge and awareness of HIV screening practices by rural providers, HIV testing may also be increased in rural areas by removing several structural barriers. Long distances to care are often cited as a significant barrier to care in rural settings19 and may also explain why rural residents were less likely to receive HIV testing at a private doctor. Expanding the availability of community-based clinics and the use of nonclinical screening tools, such as mobile HIV testing units, may help overcome the challenge of limited transportation or long distances to providers. Future research should investigate the acceptability of mobile HIV testing units in rural areas of the United States.
Furthermore, there were no differences in insurance coverage between rural and urban residents, indicating lower HIV testing rates in rural settings may not be primarily due to cost or insurance coverage. Instead, there may be less access to publically funded or easily accessible HIV testing for the uninsured in rural settings. Our results showed that of the testing sites serving rural individuals, clinics had the highest proportion of uninsured patients. Therefore, expanding access to and availability of free HIV testing at clinics serving the uninsured or underinsured may significantly improve HIV testing rates in rural settings.
This study has several limitations. First, the results should be interpreted in light of the missing data. We found that living in a rural setting was associated with lower odds of having missing lifetime HIV testing data compared with never testing (OR, 0.83; 95% CI, 0.77–0.90). However, because missing values were treated as a separate category in the outcome variables in the multinomial logistic regression analyses, it is unlikely that the magnitude or directionality of the results would change significantly had there been no missing data.
Second, the validity of self-reported HIV testing behaviors has been challenged in several studies.20–22 Comparisons of self-reported and medical record HIV testing has shown good agreement between date of first positive HIV test but poor agreement between date of last negative HIV test.20 Self-reported HIV testing behaviors may be subject to recall bias, particularly if testing is not a salient issue. The relatively small difference between lifetime and past-year testing frequencies may indicate that self-report data are subject to forward telescoping, which is a tendency for people to perceive recent events as being more recent.23
Third, this study is generalizable to a large portion of the US population, with the exception of individuals younger than 18 years, or those who are homeless, incarcerated, in the military or in nursing homes or other long-term care facilities. Future research should aim to address disparities in HIV testing particularly among homeless or incarcerated individuals. The BRFSS survey also excludes individuals who do not have either a landline telephone or a cellphone. Finally, this study was cross-sectional by design, and with the lack of temporality, causal inference cannot be estimated. We are also unable to follow participants to record changes over time, which may be especially important for measuring HIV testing behaviors. Despite such limitations, our study is one of the few in the literature to have examined recent HIV testing behaviors stratified by rural/urban status.
1. Joint UN Programme on HIV/AIDS (UNAIDS). 90–90-90 An ambitious treatment target to help end the AIDS epidemic. Available at: http://www.unaids.org/sites/default/files/media_asset/90-90-90_en_0.pdf
. Accessed April 4, 2018.
2. Satcher Johnson A, Song R, Hall HI. State-level estimates of HIV incidence, prevalence and undiagnosed infection. Presented at: Conference on Retroviruses and Opportunistic Infections [P0260] 2017; Seattle.
3. Centers for Disease Control and Prevention. HIV Testing. Available at: http://www.cdc.gov/hiv/testing/index.html
. Accessed April 4, 2018.
4. Sun J, Liu L, Shen P, et al. Trends in baseline CD4 cell counts and risk factors for late antiretroviral therapy initiation among HIV-positive patients in Shanghai, a retrospective cross-sectional study. BMC Infect Dis 2017; 17:285.
5. Branson BM, Handsfield HH, Lampe MA, et al. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep 2006; 55:1–17.
6. Ford C, Mulatu M, Godette D, et al. Trends in HIV testing among US older adults prior to and since release of CDC's routine HIV testing recommendations: national findings from the BRFSS. Public Health Rep 2015; 130:514–525.
7. Kwan C, Rose C, Brooks J, et al. HIV testing among mean at risk for acquiring HIV infection before and after the 2006 CDC recommendations. Public Health Rep 2016; 131:311–319.
8. Van Handel M, Kann L, Olsen EM, et al. HIV testing among us high school students and young adults. Pediatrics 2016; 137:e20152700.
9. Ansa B, White S, Chung Y, et al. Trends in HIV testing among adults in Georgia: analysis of the 2011–2015 BRFSS data. Int J Environ Res Public Health 2016; 13:1126.
10. Centers for Disease Control and Prevention. HIV in the United States by Geography. https://www.cdc.gov/hiv/statistics/overview/geographicdistribution.html
. Accessed April 4, 2018.
11. 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; 26:105–112.
12. Ohl M, 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.
13. Sattin RW, Wilde JA, Freeman AE, et al. Rapid HIV testing in a southeastern emergency department serving a semiurban-semirural adolescent and adult population. Ann Emerg Med 2011; 58(suppl 1):S60–S64.
14. White BL, Walsh J, Rayasam S, et al. What makes me screen for HIV? Perceived barriers and facilitators to conducting recommended routine HIV testing among primary care physicians in the Southeastern United States. J Int Assoc Provid AIDS Care. 2015; 14:127–135.
15. Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System: About BRFSS. Available at: http://www.cdc.gov/brfss/about/index.htm
. Accessed August 30, 2017.
16. Woodring JV, Kruszon-Moran D, Oster A, et al. Did CDC's 2006 revised HIV testing recommendations make a difference? Evaluation of HIV testing in the US household population, 2003-2010. J Acquir Immune Defic Syndr 2014; 67(3):331–340.
17. Baumann KE, Hemmige V, Kallen MA, et al. Whether patients want it or not, physician recommendations will convince them to accept HIV testing. J Int Assoc Provid AIDS Care 2018; 17:1–6.
18. Al Mohajer M, Lyons M, King E, et al. Internal medicine and emergency medicine physicians lack accurate knowledge of current CDC HIV testing recommendations and infrequently offer HIV testing. J Int Assoc Physicians AIDS Care 2012; 11:101–108.
19. Reif S, Golin CE, Smith SR. Barriers to accessing HIV/AIDS care in North Carolina: rural and urban differences. AIDS Care 2005; 17:558–565.
20. An Q, Chronister K, Song R, et al. Comparison of self-reported HIV testing data with medical records data in Houston, TX 2012-2013. Ann Epidemiol 2016; 26:255–260.
21. Fisher DG, Reynolds GL, Jaffe A, et al. Reliability, sensitivity and specificity of self-report of HIV test results. AIDS Care 2007; 19:692–696.
22. Yudin MH, Barbara AM, Guenter D, et al. Medical records and women's self-report are not reliable sources for determining whether prenatal HIV testing was done. J Obstet Gynaecol Can 2006; 28:867–872.
23. Janssen SM, Chessa AG, Murre JM. Memory for time: how people date events. Mem Cognit 2006; 34:138–147.