HIV, other sexually transmitted infections (STIs), and viral hepatitis are important and challenging public health concerns. In the United States, more than 1.1 million people are living with HIV infection, 0.8 to 1.4 million have chronic hepatitis B virus (HBV) infection, and 2.7 to 3.9 million have chronic hepatitis C virus (HCV) infection.1–3 Moreover, recent estimates suggest that 20 million new STIs* occur in the United States each year.4 HIV, other STIs, and viral hepatitis are syndemics and share common routes of transmission.5 As a result, populations at highest risk for HIV transmission—men who have sex with men (MSM), injection drug users (IDUs), and certain subgroups of sexually active heterosexuals (especially those with low socioeconomic status [SES])—are also disproportionately affected by STIs or viral hepatitis.3,6–11
Targeting prevention efforts requires understanding the geographic distribution of these infections. Local prevalence estimates provide data important for planning prevention but are not always available or comparable with data from other locations. To our knowledge, no studies have examined whether and how the prevalence of HIV, other STIs, and viral hepatitis in high-risk populations varies by urbanicity (the degree to which a geographic area is urban). Prevalence of these infections may be influenced by a number of factors—including race/ethnicity, age, SES, or foreign-born status—and these factors may also vary by urbanicity. Nonetheless, regardless of any underlying demographic reasons for differences in burden, assessing whether prevalence of these infections varies by urbanicity is critical to guiding prevention, testing, and treatment programs to the geographic areas most affected.
Many types of prevention interventions such as behavioral interventions, media campaigns, and others are more efficiently delivered in focused geographic areas. Likewise, when the geographic distribution of disease is understood more fully, providers can tailor testing and counseling practices to better address local conditions, and efforts can be undertaken to ensure adequate capacity for treatment of these infections. Finally, understanding how the characteristics of these populations vary geographically can improve the design and delivery of interventions appropriate to the population.
The National Health and Nutrition Examination Study (NHANES) uses a probability sample of the US population; conducts laboratory testing for HIV, other STIs (including chlamydia, gonorrhea, syphilis, herpes simplex virus type 2 [HSV-2], and human papillomavirus [HPV]), and viral hepatitis; and is not influenced by test-seeking behavior. As a result, it is an excellent source of data to guide prevention, testing, and treatment programs. We sought to understand how prevalence of HIV, other STIs, and viral hepatitis vary among MSM, IDUs, heterosexuals, and low-SES heterosexuals by urbanicity.
MATERIALS AND METHODS
Sampling and Recruitment
The National Health and Nutrition Examination Study is a series of cross-sectional surveys designed to assess the health and nutritional status of adults and children in the United States.12 Originally conducted periodically, NHANES has operated continuously since 1999, examining a nationally representative sample of the civilian, noninstitutionalized population by surveying approximately 5000 persons from 15 counties each year, with data released in 2-year cycles. Participants are chosen using a complex, stratified, multistage probability sampling design.13 The National Health and Nutrition Examination Study consists of an interview, physical examination, and laboratory tests. The medical examination is conducted in specialized mobile examination units, and the portion of the interview regarding sexual and drug use behaviors is administered in a private room using audio computer-assisted self-interview.
We examined 6 populations: (1) the overall population; (2) men who reported ever having oral or anal sex with another man (MSM-ever); (3) men who reported having oral or anal sex with another man in the past 12 months (MSM-past 12 months); (4) persons who reported ever injecting drugs (IDUs-ever); (5) persons who did not report lifetime history of male-male sex or injection drug use and reported having oral, vaginal, or anal sex with an opposite-sex partner in the past 12 months (heterosexuals); and (6) heterosexuals who also reported low income or limited education (low-SES heterosexuals). Low income was defined as income at or below the U.S. Department of Health and Human Services (HHS) poverty guidelines,14 and limited education was defined as having a high school diploma (or equivalent) or less.
To ensure adequate sample size for smaller subpopulations (e.g., MSM and IDUs) and less common outcomes (e.g., HIV and HBV), we included data from multiple years (1999–2010), although we did not assess temporal trends. Some laboratory tests were not performed on the entire sample (i.e., they were restricted to certain age groups) or during all NHANES cycles from 1999 to 2010. For these tests, we restricted the analysis to the relevant subsample.
To determine urbanicity of residence, we used the 2006 Centers for Disease Control and Prevention National Center for Health Statistics Urban-Rural Classification Scheme for Counties, which classifies all US counties and county equivalents into 1 of 6 urbanization levels—4 (large central, large fringe, medium, and small metro) for metropolitan counties (those counties included in metropolitan statistical areas as delineated by the Office of Management and Budget) and 2 (micropolitan and noncore) for nonmetropolitan counties.15 Large metropolitan (metro) areas (those with population ≥1 million) are divided into large central metro counties, which contain the urban core of a city, and large fringe metro counties, which are suburban. The 63 large central metro counties comprise 2% of the counties in the United States but contain 30% of the US population. Owing to the small sample size and low prevalence of some infections, there was a risk of disclosure of confidential information that necessitated combining some of the urbanization levels. For analyses of the overall sample, we collapsed medium and small metropolitan counties into one category (“medium/small metro”) and micropolitan and noncore counties into another category (“nonmetro”), resulting in 4 categories (Fig. 1). For subgroup analyses, we further collapsed medium/small metro and nonmetro into one category (“medium/small metro/nonmetro”). Sampling strategies and data release agreements do not allow for presentation of data for specific jurisdictions (e.g., cities and states).
We present data on sex, race/ethnicity, education, household income, SES, health insurance, sexual identity, number of partners, and condom use, stratified by urbanicity. Sexual identity was only asked in 2001 to 2010. Because the questions on condom use changed in 2005, we used data on condom use from 2005 to 2010 only.
We also present data on prevalence of HIV; HSV-2; HPV 6, 11, 16, and 18; chlamydia; HBV; and HCV. Full documentation of tests conducted is available at www.cdc.gov/nchs/nhanes.htm. Per NHANES protocol, not all tests were conducted on all populations. HIV antibody testing (and confirmation by Western blot testing) was conducted among persons aged 18 to 49 years (1999–2008) or 18 to 59 years (2009–2010) years. Herpes simplex virus type 2 antibody testing was conducted among persons aged 14 to 49 years (1999–2010). Serum HPV antibody testing was conducted for 4 vaccine-preventable HPV types among persons aged 14 to 59 years (2003–2004 only), and HPV types were collapsed into high-risk, or oncogenic, types (HPV 16 and 18) and low-risk, or nononcogenic, types (HPV 6 and 11) for analysis. Urine chlamydia DNA amplification testing was conducted among persons aged 14 to 39 years (1999–2010). Hepatitis B surface antigen was conducted among persons 6 years or older who had a positive hepatitis B core antibody test result (1999–2010), and HCV antibody testing (and confirmation by recombinant immunoblot assay) was conducted among persons 6 years or older (1999–2010). In this analysis, we included all available data for adults aged 18 to 59 years in all or part of 1999 to 2010.
Owing to small sample sizes (cell sizes <5 in multiple strata), we were not able to assess the prevalence of gonorrhea and syphilis in this analysis. In addition, we were not able to present data on HPV, chlamydia, or HBV for MSM or IDUs or data on HIV for IDUs.
We calculated weighted percentages and confidence intervals for the sample, overall and stratified by urbanicity. Data were analyzed using the statistical software packages SAS (SAS Institute Inc, 2002) and SUDAAN (Release 10.0, 2008). SUDAAN uses sample weights and calculates variance estimates that account for complex survey design. All estimates were weighted to account for selection probabilities and nonresponse using the medical examination weights. Weights were constructed for these analyses as described by National Center for Health Statistics documentation and, depending on how many survey cycles, were combined to produce a given estimate.16 All standard error (SE) estimates were calculated using the Taylor series (linearization) method within SUDAAN. Confidence intervals were calculated using a logit transformation. The degrees of freedom for variance estimation were based on subtracting the number of strata from the number of primary sampling units. P values for comparisons by urbanicity were based on a test for independence using an F statistic derived from a Wald χ2 for categorical variables; values less than 0.05 were considered significant. Estimates with relative SE greater than 30% are indicated and should be interpreted with caution.
Because of risk for disclosure of confidential information, estimates based on underlying cell sizes less than 5 are not presented. For estimates of disease prevalence suppressed for this reason, we replaced the estimate and confidence limits with an indication of the upper confidence limit (e.g., <11%).
Sample characteristics, displayed in Table 1, varied significantly by urbanicity for all sociodemographic variables analyzed, except sex. Compared with persons in other levels of urbanicity, persons in large central metro areas were more likely to be racial/ethnic minorities, of younger ages, have no health insurance, identify as something other than heterosexual or straight, have more than 2 sex partners during the past 12 months, and have higher levels of consistent condom use during the past 12 months. The highest education levels and household incomes occurred in persons from large fringe metro areas.
In the overall sample, prevalence of HIV, HSV-2, and HBV varied significantly by urbanicity; for all 3 infections, prevalence was highest in large central metro areas. Prevalence of HIV and HBV was low (<1% even in large central metro areas). On the other hand, HSV-2 was common in all levels of urbanicity (15.6%–21.0%). Prevalence of low-risk HPV, high-risk HPV, chlamydia, and HCV did not vary by urbanicity in the overall sample; of these 3 infections, low-risk and high-risk HPV were the most common.
Men Who Have Sex With Men
Most sociodemographic characteristics (i.e., race/ethnicity, age, education, income, SES, and health insurance) of MSM-ever and MSM-past 12 months (Table 2) demonstrated patterns of variation by urbanicity that were similar to patterns of variation seen in the overall sample, although for several characteristics, differences among MSM did not achieve statistical significance. Higher proportions of MSM-ever and MSM-past 12 months in large central and large fringe metro areas identified as homosexual or gay than in other areas.
Among MSM-ever and MSM-past 12 months, prevalence of HIV varied significantly by urbanicity, with prevalence highest in large fringe metro areas (13.4% among MSM-ever and 16.9% among MSM-past 12 months), slightly lower in large central metro areas (10.1% among MSM-ever and 14.5% among MSM-past 12 months), and lowest in medium/small metro/nonmetro areas (upper bound of confidence limit 7% for MSM-ever and 11% for MSM-past 12 months; Table 2). For both groups, prevalence of HSV-2 and HCV did not vary significantly by urbanicity.
Injection Drug Users
In large central metro areas, approximately half of IDUs-ever were male, whereas approximately three-quarters of IDUs-ever in large fringe metro areas and medium/small metro/nonmetro areas were male (Table 3). In large central metro areas, IDUs-ever were less likely to be white; other sociodemographic characteristics of IDUs-ever did not vary by urbanicity. Likewise, prevalence of HSV-2 and HCV did not vary significantly by urbanicity.
Among heterosexuals, patterns of variation in sociodemographic characteristics by urbanicity were similar to patterns of variation among the overall sample (Table 4). Prevalence of HSV-2, chlamydia, and HBV varied significantly by urbanicity, with the highest prevalence of all 3 infections in large central metro areas. Prevalence of HIV, low-risk HPV, high-risk HPV, and HCV did not vary significantly by urbanicity among heterosexuals.
Among low-SES heterosexuals, prevalence of HIV varied by urbanicity, with higher prevalence in large central and large fringe metro areas (see Table, Supplemental Digital Content 1, http://links.lww.com/OLQ/A82). Prevalence of HSV-2 and HBV was highest in large central metro areas, whereas prevalence of low-risk HPV, high-risk HPV, chlamydia, and HCV did not vary significantly by urbanicity for low-SES heterosexuals.
These nationally representative, population-based data demonstrate important geographic differences in the prevalence of HIV, HBV, and HSV-2. HIV and HBV consistently had a higher prevalence in large central metro areas, and HIV also had a high prevalence in large fringe metro areas. Both infections were uncommon in the overall sample (<1% even in areas with highest prevalence) and relative differences by urbanicity were large. These data suggest that for infections such as HIV and HBV, it is important to strengthen programs that provide screening and prevention as well as linkage to and retention in care and treatment in large metro areas.
In contrast, HSV-2, although most prevalent in large central metro areas, was common among all populations and all levels of urbanicity. Therefore, targeted geographic interventions for HSV-2 would not be as beneficial from a public health perspective. The same holds for infections such as HPV, chlamydia, and HCV, which showed no evidence of concentration in large metro areas. Still, for all of these infections, including those with substantial geographic variation, it is important to ensure that all persons have access to basic information about transmission and prevention. Moreover, because large metro areas contain substantial proportions of the US population, they are an important area for prevention even when geographic differences in prevalence are absent.
Knowing how the characteristics of these populations vary by urbanicity also helps design better interventions. For example, that MSM who do not reside in large metropolitan areas are less likely to identify as gay suggests that efforts to reach MSM in these areas must be inclusive of men, regardless of sexual identity.
In addition to helping guide overall allocation and design of prevention efforts, these data are also useful for researchers and health departments interested in comparing their local prevalence with data from other areas of comparable urbanicity. For example, the National HIV Behavioral Surveillance System (NHBS), which conducts surveys and HIV testing among populations at high risk for HIV infection (MSM, IDUs, and heterosexuals at increased risk), operates exclusively in large urban areas. Before this analysis, few sources of data from nationally representative surveys of these risk groups have been available to compare with NHBS prevalence estimates. The HIV prevalence among MSM interviewed in NHBS, 19%, is fairly similar to and falls well within the confidence limits of the HIV prevalence estimate for MSM-past 12 months in large metro areas from this analysis.17 On the contrary, HIV prevalence among low-SES heterosexuals in this nationally representative sample is 0.4%, even in large central metro areas, much lower than the 2% prevalence among an NHBS sample with similar inclusion criteria recruited using respondent-driven sampling.18 This, combined with differing sample characteristics,18 suggests that NHBS sampling methods are reaching a specific subpopulation of minority, extremely low-income heterosexuals at particularly elevated risk of HIV infection. Likewise, groups that conduct serosurveys of other STIs and viral hepatitis in areas of a particular urbanicity level will be able to compare their estimates to those found in this analysis, allowing triangulation of data and improved interpretation of findings.
Several of our findings may be explained by population characteristics that vary by urbanicity. For example, HIV prevalence among MSM is higher in older age groups,17 and MSM in large fringe metro areas were older than MSM in large central metro areas. Therefore, age differences may explain variation in HIV prevalence among MSM in large central versus large fringe metro areas. Likewise, HBV prevalence may be higher in large central metro areas because prevalence is higher among foreign-born persons, who are more likely to live in large central metro areas.19,20 Nonetheless, regardless of the underlying reasons for differences in prevalence, these findings argue for focusing prevention efforts in these areas. Additional exploration of the reasons for these differences, including demographic, social, and network factors, could prove useful in designing specific interventions in these areas.21,22
Owing to small sample sizes of some populations, we combined data from multiple cycles and could not assess trends. However, we have no reason to believe that differences by urbanicity would have changed significantly during this 12-year period. At the time of this analysis, data on HPV were only available for 2003 to 2004. The prevalence of vaccine type HPV has decreased in adolescents since HPV vaccine became available in 2006,23 but it is unclear whether adult populations have experienced reductions in prevalence since that time. Participants may have underreported stigmatized behaviors such as male-male sex and injection drug use; this may have affected the composition of our risk groups. However, use of audio computer-assisted self-interview may have minimized this bias.
Finally, NHANES is designed to produce samples that are representative by age, sex, and race/ethnicity, but not SES. This fact, combined with possible clustering of infections with a low prevalence in the general population, could lead to biased prevalence estimates. Moreover, NHANES is designed to be representative at the national level but does not necessarily produce estimates that are representative by urbanicity level, as only 15 counties are sampled each year. Also, we were not able to assess or account for possible regional differences in prevalence by urbanicity.
Several populations are at increased risk for both HIV and STI transmission and acquisition, including MSM, IDUs, and certain subgroups of heterosexuals. This analysis demonstrates the importance of considering the impact of place on health of these populations. Our findings help to understand the geographic variation in the prevalence of HIV and other STIs among MSM, IDUs, and heterosexuals and contribute to effective planning for prevention programs, especially at the national level.
1. Armstrong GL, Wasley A, Simard EP, et al. The prevalence of hepatitis C virus infection in the United States, 1999 through 2002. Ann Intern Med 2006; 144: 705–714.
2. Centers for Disease Control and Prevention. Monitoring selected national HIV prevention and care objectives by using HIV surveillance data—United States and 6 U.S. dependent areas—2010. HIV Surveillance Supplemental Report 2012; 17(No. 3, part A).
3. Weinbaum CM, Williams I, Mast EE, et al. Recommendations for identification and public health management of persons with chronic hepatitis B virus infection. MMWR Recomm Rep 2008; 57: 1–20.
4. Satterwhite CL, Torrone E, Meites E, et al. Sexually transmitted infections among US women and men: Prevalence and incidence estimates, 2008. Sex Transm Dis 2013; 40: 187–193.
5. Ivankovich MB, Fenton KA, Douglas JM Jr. Considerations for national public health leadership in advancing sexual health. Public Health Rep 2013; 128 (suppl 1): 102–110.
6. Workowski KA, Berman S. Sexually transmitted diseases treatment guidelines, 2010. MMWR Recomm Rep 2010; 59: 1–110.
7. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2011. 2012.
8. Centers for Disease Control and Prevention. HIV Surveillance Report, 2011. February 2013 2013;23.
9. Dinenno EA, Oster AM, Sionean C, et al. Piloting a system for behavioral surveillance among heterosexuals at increased risk of HIV in the United States. Open AIDS J 2012; 6: 169–176.
10. Centers for Disease Control and Prevention. Characteristics associated with HIV infection among heterosexuals in urban areas with high AIDS prevalence—24 cities, United States, 2006–2007. MMWR Morb Mortal Wkly Rep 2011; 60: 1045–1049.
11. Smith BD, Morgan RL, Beckett GA, et al. Recommendations for the identification of chronic hepatitis C virus infection among persons born during 1945–1965. MMWR Recomm Rep 2012; 61: 1–32.
12. Curtin LR, Mohadjer LK, Dohrmann SM, et al The National Health and Nutrition Examination Survey: Sample Design, 1999–2006. Vital Health Stat 2 2012(155):1–39.
13. Zipf G, Chiappa M, Porter KS, et al. National Health and Nutrition Examination Survey: Plan and operations, 1999–2010. National Center for Health Statistics. Vital Health Stat 2013; 1.
14. Department of Health and Human Services. Annual update of the HHS poverty guidelines. Fed Regist 2008; 73.
15. Ingram DD, Franco SJ. NCHS urban-rural classification scheme for counties. Vital Health Stat 2 2012(154):1–65.
16. National Center for Health Statistics. Analytic and Reporting Guidelines: The National Health and Nutrition Examination Survey. 2005.
17. Centers for Disease Control and Prevention. Prevalence and awareness of HIV infection among men who have sex with men—21 cities, United States, 2008. MMWR Morb Mortal Wkly Rep 2010; 59: 1201–1207.
18. Centers for Disease Control and Prevention. HIV infection among heterosexuals at increased risk—United States, 2010. MMWR Morb Mortal Wkly Rep 2013; 62: 183–188.
19. Mitchell T, Armstrong GL, Hu DJ, et al. The increasing burden of imported chronic hepatitis B—United States, 1974–2008. PLoS One 2011; 6: e27717. doi: 27710.21371/journal.pone.0027717. Epub 0022011 Dec 0027717.
20. Kandel WA. The U.S. Foreign-Born Population: Trends and Selected Characteristics. Congressional research Service, 2011.
21. Rothenberg R. Maintenance of endemicity in urban environments: A hypothesis linking risk, network structure and geography. Sex Transm Infect 2007; 83: 10–15.
22. Rothenberg R. The relevance of social epidemiology in HIV/AIDS and drug abuse research. Am J Prev Med 2007; 32 (suppl 6): S147–S153.
23. Markowitz LE, Hariri S, Lin C, et al. Reduction in human papillomavirus (HPV) prevalence among young women following HPV vaccine introduction in the United States, National Health and Nutrition Examination Surveys, 2003–2010. J Infect Dis 2013; 208: 385–393.
*Includes chlamydia, gonorrhea, syphilis, herpes, HPV, hepatitis B, HIV, and trichomoniasis.