Abstract
Objectives. To examine state-level factors associated with late-stage HIV diagnoses in the United States.
Methods. We examined state-level factors associated with late-stage diagnoses by estimating negative binomial regression models. We used 2013 to 2016 data from the National HIV Surveillance System (late-stage diagnoses), the Behavioral Risk Factor Surveillance System (HIV testing), and the American Community Survey (sociodemographics).
Results. Among individuals 25 to 44 years old, a 5% increase in the percentage of the state population tested for HIV in the preceding 12 months was associated with a 3% decrease in late-stage diagnoses. Among both individuals 25 to 44 years of age and those aged 45 years and older, a 5% increase in the percentage of the population living in a rural area was associated with a 2% to 3% increase in late-stage diagnoses.
Conclusions. Increasing HIV testing may lower late-stage HIV diagnoses among younger individuals. Increasing HIV-related services may benefit both younger and older people in rural areas.
In 2016, 39 589 individuals received an HIV diagnosis in the United States.1 Diagnosing HIV in the early stages of infection and linking people to health care for medical treatment (i.e., antiretrovirals) improve the health and survival of those with HIV infection and reduce the risk of HIV transmission.2,3 Total reported annual diagnoses do not distinguish newly acquired infection from infection acquired years prior to diagnosis. Monitoring late-stage diagnoses (i.e., infections classified as stage 3 [AIDS] at diagnosis) provides insight into the length of time between infection and diagnosis. On average, a person develops late-stage disease 8 years after acquiring HIV infection without treatment.4 A late-stage HIV diagnosis is a missed opportunity to provide medical treatment to improve health outcomes and reduce risk of transmission through viral suppression and decreases in HIV-related risk behaviors.2,5–7
Nationally, in 2016, 21.3% of HIV diagnoses (n = 8360) were classified as late-stage infection. Although this percentage represented a decrease from 28.3% in 2010, there is room for improvement.8 In 2016, relatively high percentages of late-stage diagnoses were observed among individuals 45 to 54 years old (33.0%) and those aged 55 years or older (36.3%), male heterosexuals (33.9%), and Native Hawaiians and other Pacific Islanders (23.8%).8 Conversely, people 13 to 24 years of age (8.8%), men who have sex with men and inject drugs (17.9%), and Blacks/African Americans (20.2%) had relatively low percentages of late-stage diagnoses. Late-stage diagnoses varied across states, ranging from 13.5% in Alaska to 60% in Vermont.8
Late-stage diagnoses have been associated with urbanicity, absence of HIV testing, age, gender, race/ethnicity, transmission risk, employment status, and nativity status.9–14 People living in rural or small metropolitan statistical areas (MSAs) are more likely to be diagnosed in the late stage of HIV infection.10,11,15 For example, the percentage of late-stage diagnoses was higher in nonmetropolitan areas (29.6%) and small MSAs (26.3%) than in large MSAs (23.3%) in 2012.15 Among individuals aged 55 years or older, late-stage diagnoses were more common in nonmetropolitan areas (45.9%) than in small and large MSAs (42.2% and 38.7%, respectively). Reduced access to health care delivery sites and community-based HIV counseling and testing centers in nonmetropolitan areas and small MSAs may contribute to less testing and a higher percentage of individuals with late-stage HIV diagnoses.16
Our objective in this analysis was to examine state-level factors associated with late-stage HIV diagnoses. In particular, we aimed to identify factors associated with late-stage diagnoses that state public health programs can influence through policy change (e.g., expand Medicaid) or service provision (e.g., increase testing efforts) to improve early diagnosis of HIV infection and linkage to care as soon as possible. We hypothesized that percentages of late-stage diagnoses would be lower in states with a higher percentage of individuals tested for HIV in the preceding 12 months, states with a lower percentage of people without health insurance coverage, and states with a lower percentage of the population in rural areas.
METHODS
We created an analytic data set consisting of state-level aggregate data. We used 2013 to 2016 data (the 4 most recent years of data available) from the National HIV Surveillance System (NHSS), the Behavioral Risk Factor Surveillance System (BRFSS), and the American Community Survey (ACS). Four years of data were used to help minimize fluctuations in states with a low number of diagnoses.
The NHSS, the primary source for monitoring trends in HIV diagnoses in the United States, includes data on the number of individuals with newly diagnosed HIV infection annually and the total number of people with diagnosed HIV infection in the United States.1 We used state-level counts of the number of individuals whose infection was diagnosed at stage 3 (numerator) and the number of HIV diagnoses (denominator) to calculate the state percentage of individuals with late-stage diagnoses annually (n = 200). State average percentage is defined as a weighted mean (based on the total number of diagnoses) of the percentages of late-stage diagnoses for each state and year. Average relative percentage change is defined as the average of the relative changes for the 50 states between 2013 and 2016.
The BRFSS is a state-based annual cross-sectional telephone survey of people aged 18 years and older that collects information on health behaviors, preventive health practices, and health care access.17 BRFSS interviewers ask respondents whether they have ever been tested for HIV, excluding tests that were part of a blood donation, and the date of their most recent HIV test. During 2013 through 2016, records for respondents who met the following criteria were included in the analysis for a given year: (1) the respondents resided in the 50 states or the District of Columbia, (2) they reported a valid age (i.e., 18 years and older), (3) they responded “yes” or “no” when asked whether they had ever been tested, and (4) those who responded “yes” had at least the year of their most recent test reported.
In the case of records with a most recent HIV test date in the preceding calendar year but an unknown or missing month (17% [n = 12 518]), the month of June was assigned, and respondents with interviews conducted through June of the following year were categorized as having a test in the previous 12 months. This resulted in classifying approximately half of the tests in the prior year with an unknown month as taking place in the 12 months before the survey, as would be expected assuming that HIV testing is uniformly distributed throughout the year. The final sample included 1 644 585 records (89% of the total 1 852 070 BRFSS records). Date of most recent HIV test and interview date were used to calculate the estimated number and percentage of people aged 18 years or older who had been tested for HIV in the preceding 12 months. Data were weighted to account for the complex survey design, nonresponse, and sociodemographic factors to yield estimates representative of civilian, noninstitutionalized adults 18 to 98 years of age in the United States.
The ACS is an annual nationwide survey of approximately 3.5 million US households that collects detailed information on population and housing characteristics.18 We used the 1-year estimates for each state corresponding to the years of interest for all ACS covariates: education (percentage of the population with less than a high school education), uninsured (percentage of the population without health insurance coverage), and rural residence (percentage of the population living in a rural area). Other socioeconomic factors were considered, such as poverty (percentage of people living below the poverty level), but were not included owing to high collinearity with factors that were included (e.g., education).
Late-stage diagnosis is associated with age9,13; therefore, when data were available, we stratified late-stage diagnoses and covariates into 2 age groups (25–44 years and ≥ 45 years). The number of late-stage diagnoses was small among those younger than 25 years (25 states had 5 or fewer late-stage diagnoses), and thus this age group was not included in the analysis. Rural residence was not available by age group; therefore, the same values for rural residence were used for each age group. Among race/ethnicity groups, late-stage diagnosis is lower among Blacks/African Americans and higher among Hispanics/Latinos and Whites; as a result, we included percentages of late-stage diagnoses among Hispanics/Latinos and Whites as covariates.8 Among transmission groups, late-stage diagnosis is higher among heterosexuals, and thus percentage of late-stage diagnoses among heterosexuals was included as a covariate.
We used percentage of the population without health insurance coverage as a measure of access to HIV medical care and treatment as well as HIV testing. Also, we included year in the model to account for any change over time.
Descriptive analyses of state average percentages during 2013 to 2016 were conducted for individuals aged 25 years or older (by age group) whose HIV was diagnosed in the 50 states. Means and standard deviations are reported for late-stage diagnoses and each covariate by age group. In the District of Columbia, the percentage of individuals tested for HIV in the preceding 12 months was an outlier, at 2 times that of the state with next highest percentage. Because the District of Columbia was an outlier for our main independent variable of interest (HIV testing), it was excluded from the analysis.
To identify state-level factors associated with late-stage diagnoses, we used generalized estimating equation models (SAS version 9.3, GENMOD procedure; SAS Institute Inc., Cary, NC) with a robust Poisson distribution and stratified each model by age group.19 The number of late-stage HIV diagnoses was the dependent variable, and the log of the total number of diagnoses in the state was used as an offset to assess the association of independent variables with the percentage of late-stage diagnoses. Bivariate models were run for each of the covariates, followed by a multivariable model. All covariates were added to the multivariable model and a manual backward elimination method was used to remove the least significant covariates until only significant (P < .05) variables remained.
RESULTS
The national average percentage of individuals with late-stage diagnoses during 2013 to 2016 increased with age, from 21.0% among those 25 to 44 years of age to 31.7% among those aged 45 years and older (Table 1). This was also true for the majority of states individually. Conversely, the state average percentage of individuals tested for HIV in the preceding 12 months decreased with age, from 15.5% among those 25 to 44 years old to 4.3% among those aged 45 years and older (Table 2). The state average percentage of the population without health insurance coverage decreased with age, from 17.2% among individuals 25 to 44 years of age to 7.2% among individuals aged 45 years and older. The state average percentage of the population with less than a high school education increased with age, whereas the percentage of annual diagnoses among Hispanics/Latinos, the percentage of annual diagnoses among Whites, and the percentage of annual diagnoses among heterosexuals decreased with age.
TABLE 1—
Average Percentages of Late-Stage HIV Diagnoses Among US Adults, by State: 2013–2016
| State | Age 25–44 Years, Mean % (SD) | Age ≥45 Years, Mean % (SD) |
| Alabama | 22.3 (2.9) | 32.5 (2.7) |
| Alaska | 18.8 (6.0) | 32.1 (13.5) |
| Arizona | 19.6 (2.7) | 32.6 (2.7) |
| Arkansas | 20.3 (3.6) | 34.9 (2.1) |
| California | 17.8 (1.1) | 27.4 (1.4) |
| Colorado | 23.6 (3.0) | 35.6 (3.4) |
| Connecticut | 25.7 (3.0) | 36.6 (1.5) |
| Delaware | 29.1 (4.7) | 40.3 (3.1) |
| Florida | 20.3 (2.0) | 29.9 (1.8) |
| Georgia | 21.9 (1.8) | 30.3 (1.8) |
| Hawaii | 21 (10.3) | 37.9 (3.7) |
| Idaho | 31 (4.7) | 43.8 (4.5) |
| Illinois | 20.7 (0.4) | 32 (0.8) |
| Indiana | 19.6 (2.9) | 33.3 (5.1) |
| Iowa | 30.9 (3.8) | 46.2 (9.0) |
| Kansas | 25.3 (3.5) | 36.3 (4.5) |
| Kentucky | 23.6 (2.9) | 36.9 (2.4) |
| Louisiana | 23.9 (2.3) | 34.7 (2.6) |
| Maine | 22.5 (5.8) | 25.8 (6.9) |
| Maryland | 21.8 (1.7) | 29.4 (1.0) |
| Massachusetts | 19.4 (0.4) | 29.1 (1.4) |
| Michigan | 23 (1.9) | 36.8 (2.2) |
| Minnesota | 25.1 (0.6) | 34.1 (2.1) |
| Mississippi | 25.6 (1.9) | 37.1 (2.7) |
| Missouri | 21.6 (3.3) | 34.2 (2.8) |
| Montana | 25 (8.9) | 41.2 (12.4) |
| Nebraska | 32 (11.8) | 40.5 (3.5) |
| Nevada | 22 (2.0) | 36.5 (4.1) |
| New Hampshire | 12.7 (6.3) | 35.1 (8.9) |
| New Jersey | 23.9 (1.0) | 31.9 (2.6) |
| New Mexico | 25.2 (3.1) | 40.5 (7.4) |
| New York | 19.2 (0.2) | 30.5 (1.5) |
| North Carolina | 21.4 (2.8) | 33.3 (1.8) |
| North Dakota | 33.9 (13.4) | 41.2 (12.8) |
| Ohio | 21.6 (2.2) | 33.7 (2.5) |
| Oklahoma | 19.6 (3.0) | 33.8 (6.4) |
| Oregon | 27.5 (2.3) | 44.3 (4.8) |
| Pennsylvania | 23.5 (0.1) | 33.9 (1.7) |
| Rhode Island | 24.2 (4.3) | 35.4 (6.8) |
| South Carolina | 24.4 (3.1) | 36.4 (2.4) |
| South Dakota | 25 (5.9) | 40.8 (12.0) |
| Tennessee | 19.1 (3.8) | 30.5 (5.2) |
| Texas | 21.4 (0.4) | 32.2 (0.2) |
| Utah | 19.2 (5.4) | 27.5 (7.6) |
| Vermont | 16 (10.2) | 36 (20.0) |
| Virginia | 19.7 (2.1) | 30.4 (0.9) |
| Washington | 21.4 (2.0) | 29.4 (2.4) |
| West Virginia | 29.9 (9.7) | 38.6 (9.4) |
| Wisconsin | 21 (4.1) | 35.8 (1.0) |
| Wyoming | 23.7 (12.9) | 40.7 (23.3) |
| National | 21.0 (3.2) | 31.7 (4.0) |
TABLE 2—
State-Level Socioeconomic and Demographic Characteristics Among US Adults, 2013–2016, by Age Group: 2013–2016
| Characteristic | Aged 25–44 Years, Mean % (SD) | Aged ≥ 45 Years, Mean % (SD) |
| Tested for HIV infection in the preceding 12 mo | 15.5 (5.2) | 4.3 (1.8) |
| Population with less than a high school education | 9.8 (2.9) | 12.3 (3.4) |
| Population without health insurance coverage | 17.2 (6.2) | 7.2 (3.0) |
| Population living in a rural areaa | 26.4 (14.5) | 26.4 (14.5) |
| Annual HIV diagnosesb | ||
| Hispanic/Latino | 27.1 (15.1) | 22.6 (12.7) |
| White | 26.0 (10.5) | 20.4 (6.8) |
| Heterosexual contact | 23.2 (7.8) | 20.8 (7.3) |
Note. The sample size was n = 200.
2010 only.
Weighted means and standard deviations by number of diagnoses in the state (per age group).
Factors associated with state percentage of late-stage diagnoses in the bivariate models varied by age group. As shown in Table 3, among those 25 to 44 years of age, higher state percentages of people tested for HIV in the preceding 12 months were associated with lower state percentages of people with late-stage diagnoses. Rural residence was significantly associated with state percentage of people with late-stage diagnoses among those 25 to 44 years old and those aged 45 years and older, with higher percentages of the population in rural areas associated with higher state percentages of individuals with late-stage diagnoses.
TABLE 3—
Bivariate and Final Multivariable Models of State-Level Socioeconomic and Demographic Characteristics and Late-Stage HIV Diagnoses Among US Adults, by Age Group, 2013–2016 (n = 200)
| Age 25–44 Years |
Age ≥45 Years |
|||
| Independent Variables by Increments of 5% | Bivariate Model, RR (95% CI) | Final Multivariable Model ARR (95% CI) | Bivariate Model, RR (95% CI) | Final Multivariable Model ARR (95% CI) |
| Year | 0.96 (0.94, 0.99) | 0.96 (0.94, 0.98) | 0.96 (0.94, 0.98) | 0.96 (0.95, 0.98) |
| Tested for HIV infection in the preceding 12 mo | 0.97 (0.94, 0.99) | 0.97 (0.95, 0.98) | 1.01 (0.94, 1.09) | . . . |
| Population with less than a high school education | 0.96 (0.87, 1.06) | . . . | 0.94 (0.89, 0.99) | 0.95 (0.91, 0.995) |
| Population without health insurance coverage | 1.01 (0.98, 1.04) | . . . | 1.00 (0.97, 1.04) | . . . |
| Population living in a rural areaa | 1.02 (1.004, 1.03) | 1.02 (1.01, 1.03) | 1.04 (1.02, 1.06) | 1.03 (1.02, 1.04) |
| Annual HIV diagnoses | ||||
| Hispanic/Latino | 1.00 (0.99, 1.01) | . . . | 1.01 (0.99, 1.03) | . . . |
| White | 1.00 (0.98, 1.01) | . . . | 1.00 (0.98, 1.03) | . . . |
| Heterosexual contact | 1.02 (1.001, 1.04) | 1.03 (1.01, 1.05) | 1.00 (0.98, 1.02) | . . . |
Note. ARR = adjusted rate ratio; CI = confidence interval; RR = rate ratio. Ellipses indicate variables not included in the final model.
2010 only.
Higher state percentages of annual diagnoses among individuals with diagnoses attributed to heterosexual contact were associated with higher state percentages of people with late-stage diagnoses in the 25- to 44-year age group. Higher state percentages of the population with less than a high school education were associated with lower state percentages of late-stage diagnoses among those aged 45 years and older. Although the factors associated with late-stage diagnoses varied among age groups, the directionality of the effect was the same for most factors.
The factors significantly associated with state percentage of late-stage diagnoses in the final multivariable models varied in the 2 age groups (Table 3). Among those 25 to 44 years of age, the state percentage of individuals tested for HIV in the preceding 12 months and the state percentage of individuals living in a rural area were significantly associated with late-stage diagnoses. A 5% increase in the state percentage of people tested for HIV in the preceding 12 months was associated with a 3% decrease in late-stage diagnoses among those 25 to 44 years old (adjusted rate ratio [ARR] = 0.97; 95% confidence interval [CI] = 0.95, 0.98). A 5% increase in the state percentage of the population in rural areas was associated with a 2% increase in late-stage diagnoses (ARR = 1.02; 95% CI = 1.01, 1.03). A 5% increase in the state percentage of annual diagnoses attributed to heterosexual contact was associated with a 3% increase in late-stage diagnoses (ARR = 1.03; 95% CI = 1.01, 1.05). Lastly, year was also significant; late-stage diagnoses decreased by 4% each year (ARR = 0.96; 95% CI = 0.94, 0.98).
Among individuals aged 45 years and older, a 5% increase in the state percentage of the population in a rural area was associated with a 3% increase in late-stage diagnoses (ARR = 1.03; 95% CI = 1.01, 1.03). A 5% increase in the state percentage of the population with less than a high school education was associated with a 5% decrease in late-stage diagnoses (ARR = 0.95; 95% CI = 0.91, 1.00; P = .029). Again, year was also significant; late-stage diagnoses decreased by 4% each year (ARR = 0.96; 95% CI = 0.95, 0.98). No other variables remained significant for this age group.
DISCUSSION
Our findings support part of our hypothesis: states with a lower percentage of the population in rural areas were more likely to have higher rates of late-stage diagnoses. However, our hypothesis that states with higher testing rates were more likely to have lower rates of late-stage diagnoses was correct only in the young age group, and no association was found between late-stage diagnoses and insurance coverage for either age group. After taking sociodemographic factors into account, we found a strong state-level association for lower late-stage diagnoses with higher percentages of HIV testing in the preceding 12 months among individuals 25 to 44 years of age, and a higher percentage of the population in rural areas was associated with higher late-stage diagnoses for both age groups (25–44 years and ≥ 45 years).
In 2006, the Centers for Disease Control and Prevention (CDC) released the “Revised Recommendations for HIV Testing of Adults, Adolescents, and Pregnant Women in Health-Care Settings,” which recommended screening for HIV (i.e., routine testing at least once) among all individuals 13 to 64 years of age, as well as yearly targeted rescreening for populations at high risk for HIV infection (e.g., men who have sex with men, people who inject drugs).20 In addition, the CDC’s Expanded HIV Testing Initiative (ETI) provided increased funding for HIV testing in jurisdictions with the highest burden of disease from 2007 to 2017 to target disproportionately affected populations.21
A previous study of HIV diagnoses in New York City revealed an association between an increase in HIV testing related to the ETI and decreased late-stage HIV diagnoses among men.22 The same association was not found among women; the authors suggested that campaigns that coincided with the ETI may have been better suited to reaching undiagnosed men. An association between the ETI and increased testing was found at the national level but varied according to race/ethnicity.23 The ETI targeted those who are disproportionately affected by HIV, and much of the testing was completed in the younger age group (i.e., 25–44 years of age).21,24 HIV testing was lower in the oldest age group than in the youngest age group (4% state average among people aged 45 years and older vs 16% among people 25–44 years of age), and there was less variability among states; therefore, the level of HIV testing might have been too low to detect the importance of testing in the oldest age group.
A study in Florida showed that late-stage diagnoses were more common in rural areas even after control for age, gender, race/ethnicity, HIV transmission mode, country of birth, and diagnosis year.11 That study also revealed that late-stage diagnoses in rural areas were more likely among older individuals and men, whereas late-stage diagnoses in urban areas were more likely among Hispanics/Latinos, Blacks/African Americans, and individuals with HIV infection attributed to heterosexual contact. A retrospective study in South Carolina showed that men, Blacks/African Americans, and people aged 50 years or older were more likely to receive an AIDS classification within 1 year of HIV diagnosis.9 Infection attributable to heterosexual transmission and having never previously been tested for HIV are also associated with higher percentages of late-stage diagnoses.10,14,25–27
A national study at the MSA level showed a decrease in late-stage diagnoses over time (2003–2012) in large and medium-sized MSAs. Although nonmetropolitan areas also saw a decrease in late-stage diagnoses, the decline was smaller than in MSAs.15 These findings are consistent with our results and suggest that residence in an urban or rural area affects late-stage diagnoses differently by age and race/ethnicity, and there is a need for different services and interventions to increase HIV testing.11 Routine opt-out testing may help diagnose HIV among individuals missed by targeted testing (i.e., risk for HIV is unrecognized by both the patient and the provider).
Significant findings in our study vary across age groups, suggesting that other factors may be influencing the interplay of HIV testing, education, and rural residence. The other demographic and socioeconomic factors significant in some but not all age groups may modulate these associations. Our finding that, in the 25- to 44-year age group, higher percentages of diagnoses among heterosexuals are associated with more late-stage diagnoses is consistent with previous studies.12,25,26 However, previous researchers found the inverse of our result indicating that, among people aged 45 years and older, decreases in late-stage diagnoses are associated with higher percentages of individuals with less than a high school education.10,28
Limitations
This study is subject to limitations. Our ecological analysis addressed only associations between state-level late-stage diagnoses and actionable characteristics and did not infer causality. Misclassification of late-stage diagnoses may occur when the date of the first HIV diagnosis is inaccurate or reporting of CD4 counts below 200 cells per milliliter or opportunistic illness is incomplete. Small numbers of HIV diagnoses (greater fluctuations in percentages of late-stage diagnoses) in some states were also a limitation; on average over the 4 years, there were 3 states (Montana, Vermont, and Wyoming) with fewer than 25 diagnoses and another 6 states (Alaska, Idaho, Maine, New Hampshire, North Dakota, and South Dakota) with fewer than 50 diagnoses.
The average relative percentage change among states was 26% between 2013 and 2016, and states with very small numbers accounted for much of the change between years. Sensitivity analyses were conducted in which states with fewer than 50 HIV diagnoses were removed, and the direction and statistical significance of results from models were unchanged. We were not able to stratify late-stage diagnoses by race/ethnicity because of small numbers. Late-stage diagnoses were high among Hispanics/Latinos at the national level, but total numbers of diagnoses varied greatly at the state level.8 The percentage of Hispanics/Latinos with late-stage diagnoses ranged from 0% to 100% among states; many states had a small number of annual diagnoses among Hispanics/Latinos, resulting in the large range. In addition, testing rates and HIV incidence and transmission affect late-stage diagnoses, and these variables change over time; it may be useful to take these issues into consideration when assessing late-stage diagnoses at the local level.29
Conclusions
Among younger individuals (25–44 years of age) with diagnosed HIV, states with higher percentages tested for HIV in the preceding 12 months and lower percentages of the population living in a rural area had lower percentages with late-stage diagnoses. These findings suggest increasing HIV testing efforts and expanding access to HIV-related services may help states reduce delayed HIV diagnoses among younger residents, improve health outcomes, and decrease further transmission of HIV through early diagnosis and entry into HIV medical care. States have done well to focus testing on people 25 to 44 years of age because they account for a large portion of estimated undiagnosed infections, and diagnosing early leads to better health outcomes; however, individuals in this age group still have a large portion (an estimated 29%) of undiagnosed infections, and HIV testing efforts targeting them should continue if not increase.3,30 More routine opt-out and targeted HIV testing among younger age groups will also help reduce late-stage diagnosis at a later age.
People aged 45 years and older accounted for 40% of all late-stage diagnoses in 2015.8 Progression from HIV to AIDS is faster among those aged 40 years and older, making testing in this group also important.31 Among older individuals with HIV, living in a state with higher percentages of the population in rural areas was associated with higher percentages of late-stage diagnoses. This association may have been due to a lack of access to HIV testing or stigma associated with HIV, which is more common in rural communities.11
Some factors associated with late-stage diagnoses in our study are more actionable by state public health departments and policymakers. Enhanced HIV testing can be implemented relatively more quickly, including in rural areas. Routine opt-out testing in health care settings in combination with more focused testing in specific high-priority subgroups will help find those whose disease is undiagnosed and late stage. These efforts can be supported by existing HIV campaigns that encourage individuals to undergo testing (https://www.cdc.gov/actagainstaids/campaigns/doingit/index.html) as well as those targeting health care providers to increase routine testing in health care settings (https://www.cdc.gov/actagainstaids/campaigns/hssc/index.html).
Diagnosing HIV infection early through increased testing efforts is a key strategy in ending the HIV epidemic in the United States. Scaling up testing programs within a state to correspond to the state’s prevalence and accessibility to testing centers is needed to make HIV testing simple, accessible, and routine.
ACKNOWLEDGMENTS
Publication of this article was made possible by the contributions of the state and territorial health departments and the HIV surveillance programs that provided surveillance data to the Centers for Disease Control and Prevention.
Note. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
CONFLICTS OF INTEREST
The authors declare no conflicts of interest.
HUMAN PARTICIPANT PROTECTION
No protocol approval was necessary for this study because no human participants were involved.
Footnotes
See also Adams and Hunt, p. 1486.
REFERENCES
- 1.HIV Surveillance Report 2017. Atlanta, GA: Centers for Disease Control and Prevention; 2018. [Google Scholar]
- 2.Cohen MS, Chen YQ, McCauley M et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365(6):493–505. doi: 10.1056/NEJMoa1105243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hutchinson AB, Farnham PG, Duffy N et al. Return on public health investment: CDC’s Expanded HIV Testing Initiative. J Acquir Immune Defic Syndr. 2012;59(3):281–286. doi: 10.1097/QAI.0b013e31823e5bee. [DOI] [PubMed] [Google Scholar]
- 4.Lodi S, Phillips A, Touloumi G et al. Time from human immunodeficiency virus seroconversion to reaching CD4+ cell count thresholds <200, <350, and <500 cells/mm(3): assessment of need following changes in treatment guidelines. Clin Infect Dis. 2011;53(8):817–825. doi: 10.1093/cid/cir494. [DOI] [PubMed] [Google Scholar]
- 5.Marks G, Crepaz N, Janssen R. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS. 2006;20(10):1447–1450. doi: 10.1097/01.aids.0000233579.79714.8d. [DOI] [PubMed] [Google Scholar]
- 6.Marks G, Crepaz N, Senterfitt JW, Janssen R. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States: implications for HIV prevention programs. J Acquir Immune Defic Syndr. 2005;39(4):446–453. doi: 10.1097/01.qai.0000151079.33935.79. [DOI] [PubMed] [Google Scholar]
- 7.Skarbinski J, Rosenberg E, Paz-Bailey G et al. Human immunodeficiency virus transmission at each step of the care continuum in the United States. JAMA Intern Med. 2015;175(4):588. doi: 10.1001/jamainternmed.2014.8180. [DOI] [PubMed] [Google Scholar]
- 8.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. 2016. Available at: https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-supplemental-report-vol-23-4.pdf. Accessed August 30, 2019.
- 9.Duffus WA, Weis K, Kettinger L, Stephens T, Albrecht H, Gibson JJ. Risk-based HIV testing in South Carolina health care settings failed to identify the majority of infected individuals. AIDS Patient Care STDS. 2009;23(5):339–345. doi: 10.1089/apc.2008.0193. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sheehan DM, Trepka MJ, Fennie KP et al. Individual and neighborhood determinants of late HIV diagnosis among Latinos, Florida, 2007–2011. J Immigr Minor Health. 2017;19(4):825–834. doi: 10.1007/s10903-016-0422-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Trepka MJ, Fennie KP, Sheehan DM, Lutfi K, Maddox L, Lieb S. Late HIV diagnosis: differences by rural/urban residence, Florida, 2007–2011. AIDS Patient Care STDS. 2014;28(4):188–197. doi: 10.1089/apc.2013.0362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Saganic L, Carr J, Solorio R, Courogen M, Jaenicke T, Duerr A. Comparing measures of late HIV diagnosis in Washington State. AIDS Res Treat. 2012;2012:182672. doi: 10.1155/2012/182672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mugavero MJ, Castellano C, Edelman D, Hicks C. Late diagnosis of HIV infection: the role of age and sex. Am J Med. 2007;120(4):370–373. doi: 10.1016/j.amjmed.2006.05.050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Noble G, Okpo E, Tonna I, Fielding S. Factors associated with late HIV diagnosis in North-East Scotland: a six-year retrospective study. Public Health. 2016;139:36–43. doi: 10.1016/j.puhe.2016.06.007. [DOI] [PubMed] [Google Scholar]
- 15.Hall HI, Tang T, Espinoza L. Late diagnosis of HIV infection in metropolitan areas of the United States and Puerto Rico. AIDS Behav. 2016;20(5):967–972. doi: 10.1007/s10461-015-1241-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.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(1):681. doi: 10.1186/1471-2458-11-681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System. Available at: https://www.cdc.gov/brfss/about/index.htm. Accessed August 30, 2019.
- 18.US Census Bureau. American Community Survey and Puerto Rico Community Survey 2014 definitions. Available at: https://www2.census.gov/programs-surveys/acs/tech_docs/subject_definitions/2014_ACSSubjectDefinitions.pdf. Accessed August 30, 2019.
- 19.Gardner W, Mulvey E, Shaw EC. Regression analyses of counts and rates: Poisson, overdispersed Poisson, and negative binomial models. Psychol Bull. 1995;118(3):392–404. doi: 10.1037/0033-2909.118.3.392. [DOI] [PubMed] [Google Scholar]
- 20.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(RR-14):1–17. [PubMed] [Google Scholar]
- 21.Centers for Disease Control and Prevention. Expanded Testing Initiative. Available at: https://www.cdc.gov/hiv/policies/eti.html. Accessed August 30, 2019.
- 22.Ransome Y, Terzian A, Addison D et al. Expanded HIV testing coverage is associated with decreases in late HIV diagnoses. AIDS. 2015;29(11):1369–1378. doi: 10.1097/QAD.0000000000000684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gaines T, Caldwell J, Ford C, Mulatu M, Godette D. Relationship between a Centers for Disease Control and Prevention expanded HIV testing initiative and past-year testing by race/ethnicity: a multilevel analysis of the Behavioral Risk Factor Surveillance System. AIDS Care. 2016;28(5):554–560. doi: 10.1080/09540121.2015.1131968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Expanded HIV Testing for Disproportionately Affected Populations: Monitoring & Evaluation Plan PS 10-10138. Atlanta, GA: Centers for Disease Control and Prevention; 2010. [Google Scholar]
- 25.Manavi K, McMillan A, Ogilvie M, Scott G. Heterosexual men and women with HIV test positive at a later stage of infection than homo- or bisexual men. Int J STD AIDS. 2004;15(12):811–814. doi: 10.1258/0956462042563585. [DOI] [PubMed] [Google Scholar]
- 26.Schwarcz S, Hsu L, Dilley J, Loeb L, Nelson K, Boyd S. Late diagnosis of HIV infection: trends, prevalence, and characteristics of persons whose HIV diagnosis occurred within 12 months of developing AIDS. J Acquir Immune Defic Syndr. 2006;43(4):491–494. doi: 10.1097/01.qai.0000243114.37035.de. [DOI] [PubMed] [Google Scholar]
- 27.Kivelä PS, Krol A, Salminen MO, Ristola MA. Determinants of late HIV diagnosis among different transmission groups in Finland from 1985 to 2005. HIV Med. 2010;11(6):360–367. doi: 10.1111/j.1468-1293.2009.00783.x. [DOI] [PubMed] [Google Scholar]
- 28.Mascolini M. Less education linked to later HIV diagnosis, higher mortality in 13-Cohort European Study. Paper presented at: 7th IAS Conference on HIV Pathogenesis, Treatment and Prevention, July 2013, Kuala Lumpur, Malaysia.
- 29.Hall HI, Song R, Tang T et al. HIV trends in the United States: diagnoses and estimated incidence. JMIR Public Health Surveill. 2017;3(1):e8. doi: 10.2196/publichealth.7051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Centers for Disease Control and Prevention. Estimated HIV incidence and prevalence in the United States, 2010–2015. Available at: https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-supplemental-report-vol-23-1.pdf. Accessed August 30, 2019.
- 31.Collaborative Group on AIDS Incubation and HIV Survival. Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: a collaborative re-analysis. Lancet. 2000;355(9210):1131–1137. [PubMed] [Google Scholar]