HIV Testing in 50 Local Jurisdictions Accounting for the Majority of New HIV Diagnoses and Seven States with Disproportionate Occurrence of HIV in Rural Areas, 2016–2017

Since 2006, CDC has recommended universal screening for human immunodeficiency virus (HIV) infection at least once in health care settings and at least annual rescreening of persons at increased risk for infection (1,2), but data from national surveys and HIV surveillance demonstrate that these recommendations have not been fully implemented (3,4). The national Ending the HIV Epidemic initiative* is intended to reduce the number of new infections by 90% from 2020 to 2030. The initiative focuses first on 50 local jurisdictions (48 counties, the District of Columbia, and San Juan, Puerto Rico) where the majority of new diagnoses of HIV infection in 2016 and 2017 were concentrated and seven states with a disproportionate occurrence of HIV in rural areas relative to other states (i.e., states with at least 75 reported HIV diagnoses in rural areas that accounted for ≥10% of all diagnoses in the state).† This initial geographic focus will be followed by wider implementation of the initiative within the United States. An important goal of the initiative is the timely identification of all persons with HIV infection as soon as possible after infection (5). CDC analyzed data from the Behavioral Risk Factor Surveillance System (BRFSS)§ to assess the percentage of adults tested for HIV in the United States nationwide (38.9%), in the 50 local jurisdictions (46.9%), and in the seven states (35.5%). Testing percentages varied widely by jurisdiction but were suboptimal and generally low in jurisdictions with low rates of diagnosis of HIV infection. To achieve national goals and end the HIV epidemic in the United States, strategies must be tailored to meet local needs. Novel screening approaches might be needed to reach segments of the population that have never been tested for HIV.

Since 2006, CDC has recommended universal screening for human immunodeficiency virus (HIV) infection at least once in health care settings and at least annual rescreening of persons at increased risk for infection (1,2), but data from national surveys and HIV surveillance demonstrate that these recommendations have not been fully implemented (3,4). The national Ending the HIV Epidemic initiative* is intended to reduce the number of new infections by 90% from 2020 to 2030. The initiative focuses first on 50 local jurisdictions (48 counties, the District of Columbia, and San Juan, Puerto Rico) where the majority of new diagnoses of HIV infection in 2016 and 2017 were concentrated and seven states with a disproportionate occurrence of HIV in rural areas relative to other states (i.e., states with at least 75 reported HIV diagnoses in rural areas that accounted for ≥10% of all diagnoses in the state). † This initial geographic focus will be followed by wider implementation of the initiative within the United States. An important goal of the initiative is the timely identification of all persons with HIV infection as soon as possible after infection (5). CDC analyzed data from the Behavioral Risk Factor Surveillance System (BRFSS) § to assess the percentage of adults tested for HIV in the United States nationwide (38.9%), in the 50 local jurisdictions (46.9%), and in the seven states (35.5%). Testing percentages varied widely by jurisdiction but were suboptimal and generally low in jurisdictions with low rates of diagnosis of HIV infection. To achieve national goals and end the HIV epidemic in the United States, strategies must be tailored to meet local needs. Novel screening approaches might be needed to reach segments of the population that have never been tested for HIV.
BRFSS is an annual cellular and landline telephone survey of the noninstitutionalized U.S. population aged ≥18 years. The median response rate among all participating states and territories was 47.1% (range = 30.7%-65.0%) in 2016 ¶ and 45.9% (range = 30.6%-64.1%) in 2017.** Respondents were asked whether they had ever been tested for HIV outside of blood donation; those who answered "yes" were asked for the month and year of their most recent test. Respondents were also asked whether any of the following HIV risk-related situations applied to them in the past year: injected drugs that were not prescribed, received treatment for a sexually § https://www.cdc.gov/brfss/index.html. ¶ https://www.cdc.gov/brfss/annual_da101278ta/2016/pdf/2016-sdqr.pdf. ** https://www.cdc.gov/brfss/annual_data/2017/pdf/2017-sdqr-508.pdf. transmitted disease, exchanged money or drugs for sex, had anal sex without a condom, or had four or more sex partners. Those who answered "yes" to this question were considered to have reported recent HIV risk.
Data collected in 2016 and 2017 were pooled and used to estimate the percentage and corresponding 95% confidence intervals (CIs) of ever testing for HIV and testing for HIV in the past year overall and for each of the 57 jurisdictions. Nationally and within the seven states with disproportionate rural HIV occurrence, counties were grouped as either mostly urban or mostly or completely rural according to designation by the 2010 U.S. Census. † † Rao-Scott chi-square tests were used to compare testing percentages between mostly urban and mostly or completely rural areas in the United States and in the seven states with disproportionate rural HIV occurrence. All estimates were weighted to account for the complex multistage sampling design. HIV diagnosis rates per 100,000 population among persons aged ≥13 years were calculated from HIV diagnoses reported to CDC's National HIV Surveillance System during 2016-2017 through December 2018; U.S. Census population estimates for 2016 and 2017 were used for the denominators. HIV diagnosis rates and testing percentages were examined together for each of the 50 local jurisdictions as well as urban and rural areas of the seven states to further characterize these areas with respect to their current HIV morbidity and testing coverage; Pearson's correlation coefficient was used to assess the correlation between these areas' testing percentages and HIV diagnosis rates. Although BRFSS testing percentages were calculated among those aged ≥18 years, HIV diagnosis rates were calculated among those aged ≥13 years to be consistent with methodology used to identify the jurisdictions accounting for the majority of new HIV diagnoses and because of limited availability of single-year age population estimates at the municipio (county equivalent) level in Puerto Rico. Analyses were performed using SAS (version 9.4; SAS Institute) and SUDAAN (version 11.0; RTI International).
During 2016-2017, 38.9% of adults aged ≥18 years in the United States had ever been tested for HIV (Table 1). Among 15,701 (3.2%) persons with reported recent HIV risk for whom at least annual rescreening is recommended, 64.8% were ever tested, and 29.2% were tested in the past year. Among all adults, the percentage ever tested (46.9%) was higher among residents of the 50 local jurisdictions that accounted for the majority of diagnoses of HIV infection among persons aged ≥13 years than was the percentage ever tested (35.5%) in the seven states with disproportionate rural HIV occurrence. Among persons with reported HIV risk, the percentage tested in the past year (34.3%) in the 50 local jurisdictions was also higher than that in the seven states (26.2%). Among all adults in these seven states, 32.1% of those residing in mostly rural areas and 37.2% of those residing in mostly urban areas had ever been tested. Among persons with reported HIV risk in these states, 18.4% of those residing in rural areas and 29.0% of those residing in urban areas were tested in the past year.
Testing percentages varied widely by jurisdiction (Table 2). Among the 50 local jurisdictions, the percentage of persons aged ≥18 years ever tested ranged from 36.5% in Maricopa County, Arizona, to 70.7% in the District of Columbia; the percentage tested in the past year (independent of reported recent HIV risk) ranged from 8.1% in Alameda County, California, to 31.3% in Bronx County, New York. Testing percentages were generally low in both urban and rural areas of the seven states with disproportionate rural HIV occurrence. Among the 50 local jurisdictions and seven states, the percentage of persons aged ≥18 years ever tested for HIV generally increased with increasing HIV diagnosis rate among persons aged ≥13 years (r = 0.71; p<0.01) (Figure). Most of the 50 local jurisdictions had higher testing percentages and diagnosis rates than did the seven states.

Discussion
In this analysis, <40% of the U.S. adult population had ever been tested for HIV. Jurisdictions with the highest rates of diagnosis of HIV infection among persons aged ≥13 years generally had higher testing percentages. The converse was also true. Ever testing for HIV was lower in rural areas of the seven states with disproportionate rural HIV occurrence, compared with that in urban areas of these states, the 50 local jurisdictions with the majority of diagnoses of HIV infection, and the United States nationally. Although past-year HIV testing was higher among persons with reported recent HIV risk than among those without such risk, the percentage tested in the past year was far below the 100% coverage recommended for this group (1,2). These findings demonstrate missed opportunities     to fully implement HIV screening recommendations in the 57 jurisdictions that will serve as the initial geographic focus of the Ending the HIV Epidemic initiative. The observed variability in both ever and past-year testing by jurisdiction highlights the need for screening strategies that are tailored to local needs. BRFSS is likely the only annual survey with a sufficient sample size to provide jurisdiction-level estimates of HIV testing to monitor long-term progress toward increasing screening coverage in the United States.
HIV screening strategies will likely need to be locally tailored and novel to reach segments of the population that have not been reached by previous efforts. Examples of novel or promising approaches to increase access to HIV testing include routinizing HIV screening in health care settings, integrating HIV screening with sexual health screenings, scaling up partner notification and other strategies (using social network strategy § § or mobile applications) that offer screening of the social and sexual networks of persons seeking HIV screening, promoting pharmacist-led screening ¶ ¶ as well as screening in other alternative clinical settings such as urgent care, and mass § § https://effectiveinterventions.cdc.gov/en/care-medication-adherence/group-4/ social-network-strategy-for-hiv-testing-recruitment. ¶ ¶ https://effectiveinterventions.cdc.gov/en/hiv-testing/group-1/ hiv-testing-in-retail-pharmacies. distribution of HIV self-tests*** (6-10). Further efforts will be needed to identify which approaches are most effective in increasing access to HIV testing in various settings and jurisdictions with different baseline needs. Early diagnosis and effective treatment that suppresses HIV replication not only minimize immune system damage and reduce individual morbidity and mortality but also reduce the risk for transmission to others. † † † Delayed diagnosis limits these benefits. *** https://www.cdc.gov/hiv/testing/hometests.html. † † † https://www.cdc.gov/hiv/risk/art/index.html.

FIGURE. Percentage of adults aged ≥18 years ever tested for human immunodeficiency virus (HIV) infection and HIV diagnosis rate* among adults and adolescents aged ≥13 years -Behavioral Risk Factor Surveillance System and National HIV Surveillance System (NHSS), 50 local jurisdictions accounting for the majority of new HIV diagnoses and seven states with disproportionate occurrence of HIV in rural areas
HIV screening is a critical entry point to a range of HIV prevention and treatment options. For persons at ongoing risk for HIV infection exposure, annual screening also offers the opportunity to discuss options to reduce risk, including HIV preexposure prophylaxis. § § § The findings in this report are subject to at least six limitations. First, because the proportion of respondents reporting recent HIV risk was small, testing percentages for this group could not be reported separately in the 57 jurisdictions. Second, self-reported data might be subject to social desirability and recall biases, which might have led to over-or underestimation of testing. Third, BRFSS response rates were low; however, the response rates are comparable with those of other national landline and cellular telephone surveys, and survey weights were designed to ensure generalizable findings. Fourth, the measure of HIV-related risk did not include every behavior that might increase risk for HIV infection, such as unprotected sex with a partner who is known to have HIV or whose HIV status is unknown. Fifth, the assessment of HIV diagnosis rates and HIV testing percentages relied on disparate age ranges (≥13 years and ≥18 years, respectively). Finally, this analysis included data from surveys conducted during 2016-2017 and HIV diagnoses that occurred during the same period. These are the most current data available for these measures but represent a delayed crosssection of the current state of HIV testing and diagnoses for 2019. To monitor progress toward national goals, closer to realtime reporting of select HIV testing activities might be needed.
HIV screening remains suboptimal for persons residing in the 57 jurisdictions that will constitute the initial geographic focus of the Ending the HIV Epidemic initiative. These data provide a baseline from which to measure changes in screening in these jurisdictions and other parts of the United States over time. To achieve national goals and end the HIV epidemic in the United States, innovative and novel screening approaches might be needed to reach segments of the population that have never been tested for HIV.