Neighborhood Social Disorganization and the Acquisition of Trichomoniasis Among Young Adults in the United States

Abstract

Objectives. We examined relationships between neighborhood social disorganization and trichomoniasis among young US adults.

Methods. We employed multilevel logistic regression modeling with secondary data from wave III of the National Longitudinal Study of Adolescent Health (2001–2002). The dependent variable—trichomoniasis—was measured via urine testing. The measures for neighborhood social disorganization were derived from the 2000 US Census—racial and ethnic composition, concentrated poverty, and residential instability. The sample comprised 11 370 individuals across 4912 neighborhoods.

Results. Trichomoniasis was more likely in neighborhoods with higher concentrations of Black residents (adjusted odds ratio [AOR] = 1.16; 95% confidence interval [CI] = 1.03, 1.30). However, this association was mediated by neighborhood concentrated poverty. Furthermore, young adults who lived in neighborhoods with higher concentrations of poverty were significantly more likely to have trichomoniasis (AOR = 1.25; 95% CI = 1.07, 1.46). Neither immigrant concentration nor residential instability was significantly associated with trichomoniasis.

Conclusions. These findings strengthen the evidence that neighborhood structural conditions are associated with individual sexually transmitted infection (STI) acquisition. Research is needed to explore the mechanisms through which these conditions influence STI. In addition, STI-prevention programs that include structural interventions targeting neighborhood disadvantage are needed.

Adolescents and young adults are at increased risk for sexually transmitted infections (STIs) because of a complex interplay of biological, behavioral, and developmental factors.¹ Nearly half of all STIs diagnosed in the United States annually are among adolescents and young adults^1,2 despite national priority goals aimed at reducing infection rates.³ Trichomoniasis, a common and easily curable STI,¹ is of increasing concern because the infection facilitates HIV acquisition and transmission through mucosal inflammation of the genital tract and alterations in the innate immune response. ^4-7 The infection is caused by the protozoa, Trichomonas vaginalis, and is typically transmitted via penis-to-vagina or vulva-to-vulva contact.¹ Infected persons are often asymptomatic or experience only mild symptoms,¹ which can hinder early detection and treatment and increase the risk of STIs and HIV.

In the United States, the prevalence of trichomoniasis is difficult to ascertain because routine screening currently is not recommended nor is the reporting of positive results required.^1,4 According to urine assay data from the National Longitudinal Study of Adolescent Health (Add Health), the prevalence of trichomoniasis among the young adult population in 2001–2002 was approximately 2.3%.⁷ The study also found that women were at greater risk than were men (2.8% vs 1.7%) as were non-Hispanic Black (6.9%) and Latino (2.1%) youths compared with their non-Hispanic White peers (1.2%).⁷ In other studies of adult women, individual risk factors for trichomoniasis included poverty, lower education, douching, non-Hispanic Black race/ethnicity, and greater numbers of lifetime sexual partners.^8,9 Among clinic samples of adolescent women, research found trichomoniasis to be associated with older male sexual partners, casual sexual activity, marijuana use, and delinquency.¹⁰

However, to date, no studies have examined the role of the broader structural context in shaping trichomoniasis risk, despite theory and previous STI research suggesting that the neighborhood environment may play a role. According to social disorganization theory,^11–14 key indicators of neighborhood structural disadvantage (i.e., racial/ethnic composition, concentrated poverty, and residential instability) influence health outcomes by weakening social ties, reducing access to institutional resources, and limiting exposure to positive role models, conventional social norms, and collective efficacy. Findings from previous research examining other STIs support the hypothesis that neighborhood contexts influence STI prevalence. For example, with respect to racial and ethnic composition, studies have found that gonorrhea rates were higher in cities and neighborhoods with greater proportions of Black residents.^15,16 Furthermore, in an analysis of Chicago neighborhoods, the incidence rates of gonorrhea and chlamydia were higher for neighborhoods in which more than 60% of the residents were Black compared with those in which more than 60% of residents were Hispanic, which suggests that segregated Hispanic ethnic enclaves may be protective of STI compared with segregated Black communities.¹⁷ Researchers hypothesize that the residential segregation of Black communities has contributed to the pervasive Black-White disparities in STI through discrimination processes, which in turn has led to greater concentration of poverty, lower male-to-female gender ratios due to the disproportionate incarceration and mortality of Black men, and closed, racially segregated sexual networks that facilitate the transmission of infection.^18–20

In addition, the role of community poverty in shaping STI risk has been examined extensively and found to be positively associated with rates of chlamydia, gonorrhea, syphilis, and HIV in cross-sectional^{15,17,21–23} and longitudinal analyses.¹⁶ Other socioeconomic factors, such as unemployment^17,24 and lower educational attainment,^16,17 have also been linked to higher rates of chlamydia and gonorrhea. Research on the effects of residential instability on STI is limited, but the single study that examined these relationships found greater residential instability was associated with fewer self-reported STIs among a national sample of adolescents.²⁴ Depending on the context, perhaps residential instability could increase STI risk by disrupting social support ties and informal social control measures or reduce STI risk by dispersing closed sexual networks that facilitate infection transmission.

Although the aforementioned studies have illustrated links between neighborhood social disorganization and a variety of STIs, limitations exist. First, the majority have been ecological studies, in which the outcomes were measured as community STI rates and no adjustment was made for potential confounding relationships with individual-level data.^{15–17,21–23} Consequently, inferences can be made only about the community, and individual variation in the outcome cannot be ascertained.²⁵ Second, although 1 study examined individual STI, the measure was based on self-report,²⁴ which potentially increases bias because of underreporting as well as unrecognized or undiagnosed infection. In addition, the study only focused on STI in general, which could limit our understanding of unique relationships with specific infectious organisms. Third, data sources of previous research tend to be at local or state levels,^{15–17,21–23} which limits external validity of the findings. Therefore, the purpose of our research was to examine relationships between neighborhood social disorganization and trichomoniasis among young adults in the United States. Our research builds on previous studies in 3 significant ways: (1) we examined multiple levels of analysis, which enabled us to simultaneously examine the independent relationships between individual and neighborhood variables and individual acquisition of trichomoniasis, (2) we examined a more refined measure of STI through the use of urine screening, and (3) we examined data from a large national data set—Add Health.

METHODS

We used cross-sectional data from Add Health, wave III (2001–2002)²⁶ to examine associations between current neighborhood conditions and trichomoniasis. Add Health is a school-based, longitudinal study of students in 7th through 12th grade that utilized a multistage, stratified, and clustered sampling design to ensure a nationally representative sample of US schools with respect to region of country, urbanicity, school size, school type, and ethnicity.²⁷ Four waves of data have been collected spanning from adolescence to young adulthood. Data are available from multiple sources, including adolescents, parents, partners, schools, and communities. Wave III individual data were collected from 2001 to 2002 and the neighborhood data were derived from the 2000 US Census.

Wave III of Add Health comprises those respondents from wave I who could be located and interviewed during the data collection time frame (n = 15 170). The response rate for wave III was 77.4%.²⁷ Add Health provides sampling weights to adjust for nonresponse at waves I and III,²⁷ and when these sampling weights are utilized the 2 samples are comparable.²⁸ A total of 14 322 young adults from wave III have sampling weights, of whom 12 446 had urine assay results for trichomoniasis. We excluded participants missing data on independent variables from analysis (n = 1076). Thus, the final sample size for this study included 11 370 individuals across 4912 neighborhoods. The young adults who were excluded because of missing data on trichomoniasis or the independent variables (n = 2644) were more likely than were those included (n = 11 370) to be male, to be living with their parents, to be of minority race or ethnicity, and to have no history of vaginal sexual intercourse. In addition, they were less likely to be low income, married, employed, in school, and to report using drugs. We found no significant difference in the odds of trichomoniasis between the final sample and those who were excluded because of missing data on the independent variables (n = 1076).

Dependent Variable

The dependent variable—trichomoniasis—was measured via urine assay screening; thus, results were binary in nature with positive urine screens coded 1. Add Health researchers collected respondents’ first stream urine on the day of interview. The urine was then tested for trichomoniasis via polymerase chain reaction-enzyme linked immunosorbent assay.^29,30 Urine testing to detect the STI is still considered experimental, but the measure has been validated with wet mount and culture in published studies.^29,30

Independent Variables

Individual-level variables.

Individual-level variables were based on respondent self-report and selected for inclusion based on previous research and theory. Sociodemographic characteristics included age (continuous measure), gender (male = 1), married (yes = 1), race/ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic American Indian, non-Hispanic Asian, and Hispanic), foreign birth (yes = 1), economic hardship (if received food stamps, housing assistance, or Aid to Families with Dependent Children in the past year coded yes = 1), employed at least 10 hours weekly (yes = 1), enrolled in high school or college (yes = 1), current residence in parental household (yes = 1), individual residential stability (continuous variable of the number of years resided in current neighborhood), and heterosexual orientation (yes = 1). In addition, we included antibiotic use in past month (yes = 1) to adjust for possible treatment of an undiagnosed STI before urine screening. Risk factors for STI included sexual risk behaviors and substance use. Sexual risk behavior controls included a continuous measure of the number of sexual partners in the past year, and categorical measures of age at first vaginal intercourse (never had sexual intercourse = reference vs at 10–15 years, 16–17 years, and ≥18 years), exchanged money for sexual intercourse in past year (yes = 1), and had sexual intercourse with an intravenous drug user in past year (yes = 1). Substance use measures included level of binge drinking in past year (continuous measure of the number of days during the past year the respondent drank ≥ 5 drinks ranging from never to every or almost every day) and a categorical measure of drug use in the past year (yes = 1 if respondent used marijuana, cocaine, methamphetamine, or other illegal drugs).

Neighborhood-level variables.

We defined the neighborhood as a geographic unit and we measured it as the census tract of residence. Census tracts commonly serve as proxies for neighborhoods and are often the basis for geographically delimited resource allocation.^31,32 Our measures of neighborhood social disorganization were based on theory,^11–14 previous research,^33–35 and available data—racial and ethnic composition, concentrated poverty, and residential instability. We measured racial and ethnic composition via 2 variables: (1) non-Hispanic Black concentration, which was composed of 1 standardized item (proportion of Black residents living in the census tract) and (2) immigrant concentration, which was composed of 3 standardized items (proportion of Latino/Hispanic residents, proportion of linguistically isolated residents, and proportion of foreign-born residents). We conducted exploratory factor analysis and internal consistency, and results supported the inclusion of the 3 immigrant concentration items into 1 index (factor loadings > 0.65 onto 1 factor and internal consistency α = 0.95).

Concentrated poverty was a composite of 4 standardized items: proportion of households below poverty, proportion of households on public assistance, total unemployment rate, and proportion of female-headed households with children. We conducted exploratory factor analysis and internal consistency and results supported the inclusion of the 4 items into 1 index (factor loadings > 0.65 onto 1 factor and internal consistency α = 0.82). Residential instability was composed of 2 standardized items: proportion of households living in the census tract for 5 years or more and proportion of owner-occupied homes. Internal consistency was α = 0.82. Last, 2 neighborhood control variables were included: region (Northeast, Midwest, West, and South [reference]) and urbanicity (standardized item of proportion of persons living in an urbanized area).

Analysis

We conducted descriptive analyses with SAS, version 9.2 (SAS Institute, Cary, NC) to better understand the characteristics and prevalence of trichomoniasis infection among our sample. Subsequently, we conducted multilevel logistic regression modeling using HLM, version 6.08 (Scientific Software International, Lincolnwood, IL) to examine the contribution of neighborhood social disorganization to trichomoniasis among young adults, adjusting for individual and neighborhood control variables. Because the prevalence of trichomoniasis in the sample was low (2%), odds ratios approximate relative risk ratios.³⁶ However, we also analyzed the data by using the Poisson distribution and findings were consistent with those from the multilevel logistic regression analyses presented here. We examined a series of 3 random intercept models in which the neighborhood racial and ethnic composition variables and the social disorganization variables of concentrated poverty and residential instability were included sequentially. We focused on the fixed effects of the multilevel model to examine direct associations of the neighborhood variables on the odds of trichomoniasis infection, adjusting for individual- and neighborhood-level control variables, as well as the potential mediating effects of concentrated poverty and residential instability on the associations between neighborhood racial and ethnic composition and the odds of trichomoniasis infection. Continuous variables were grand mean centered. We examined multicollinearity before multilevel analyses; tolerance and variance inflation factors were within range. The findings presented are from unweighted analyses because Add Health sampling weights account only for the clustering of schools and not neighborhoods. Thus, their inclusion could lead to erroneous findings (oral communication, Kim Chantala, MS, Add Health User's Conference, July 2008). However, we conducted sensitivity analyses using the weights and found no differences in statistical significance, although the strength of the associations was greater for weighted versus unweighted analyses.

RESULTS

Descriptive statistics for the sample are presented in Table 1, including unweighted and weighted proportions and means. The unweighted and weighted findings were in accordance with each other except for race and ethnicity because of oversampling and weighted adjustments. Unweighted findings indicated that the prevalence of trichomoniasis was 2.4% among the sample. Approximately 47% of the sample was male and the mean age was 22 years. Approximately 56% self-identified as White, 20% as Black, 16% as Hispanic, 7% as Asian, and 1% as American Indian, and 8% were foreign-born. In respect to socioeconomic position, 7% reported economic hardship in the past year, 70% were employed at least part time, and 38% were enrolled in high school or college. In addition, 18% of the young adults were married, 89% reported heterosexual orientation, and 40% lived with their parents. Respondents lived an average of 5.7 years at their current residence. The majority (88%) reported a history of vaginal intercourse and the average number of vaginal sexual partners in the previous year was 1.5. Nearly 34% reported illicit drug use in the past year and the mean binge drinking score was 1.2 or approximately 1 or 2 binge drinking episodes in the past year.

TABLE 1.

Characteristics of the Sample of Young Adults Aged 18 to 27 Years: Add Health, 2001–2002

	No.	Unweighted,% or Mean ±SD	Weighted,% (SE) or Mean ±SE
Trichomoniasis urine screen
Positive	277	2.4	2.1 (0.21)
Negative	11 093	97.6	97.9 (0.21)
Gender
Male	5318	46.7	50.4 (0.72)
Female	6052	53.2	49.6 (0.72)
Race/ethnicity
Hispanic	1859	16.4	11.6 (1.7)
Black	2286	20.1	15.0 (2.0)
Asian	802	7.0	3.6 (0.74)
American Indian	100	0.9	0.7 (0.34)
White (Ref)	6323	55.6	69.0 (2.9)
Foreign-born
Yes	920	8.1	5.8 (0.84)
No	10 450	91.9	94.2 (0.84)
Economic hardship
Yes	824	7.3	7.0 (0.51)
No	10 546	92.7	93.0 (0.51)
Employed
Yes	7995	70.3	70.6 (0.99)
No	3375	29.7	29.4 (0.99)
Enrolled in school
Yes	4312	37.9	36.8 (1.5)
No	7058	62.1	63.2 (1.5)
Married
Yes	2044	18.0	18.0 (0.99)
No	9326	82.0	82.0 (0.99)
Lives with parents
Yes	4507	39.6	38.9 (1.3)
No	6863	60.4	61.1 (1.3)
Heterosexual orientation
Yes	10 174	89.5	89.7 (0.45)
No	1196	10.5	10.3 (0.45)
Antibiotic use past mo
Yes	1535	13.5	13.8 (0.49)
No	9835	86.5	86.2 (0.49)
Age at first vaginal intercourse
18–25 y	3151	27.7	26.2 (0.97)
16–17 y	3524	31.0	31.3 (0.62)
10–15 y	3298	29.0	30.1 (1.0)
Has not had vaginal intercourse	1397	12.3	12.4 (0.57)
Sexual intercourse with intravenous drug user in past y
Yes	74	0.7	0.7 (0.12)
No	11 296	99.3	99.3 (0.12)
Exchanged sexual intercourse for money in past y
Yes	253	2.2	2.0 (0.22)
No	11 117	97.8	98.0 (0.22)
Drug use in past y
Yes	3837	33.8	36.1 (0.95)
No	7533	66.2	63.9 (0.95)
Age, y	11 370	22.0 ±1.7	21.8 ±0.12
No. of vaginal sexual partners in past y	11 370	1.5 ±1.9	1.5 ±0.03
Individual residential stability	11 370	5.7 ±8.3	5.4 ±0.21
Level of binge drinking in past ya	11 370	1.2 ±1.6	1.4 ±0.04

Note. Add Health = National Longitudinal Study of Adolescent Health. Sample size was n = 11 370 young adults across 4912 neighborhoods.

^aContinuous measure of the number of days during the past year the respondent drank ≥ 5 drinks ranging from never to every or almost every day.

Findings from unweighted multivariate analyses on the associations between neighborhood social disorganization and trichomoniasis are presented in Table 2. In model 1, we examined associations between neighborhood racial and ethnic composition and trichomoniasis and found that young adults who lived in neighborhoods with higher proportions of Black residents were more likely to have trichomoniasis after we adjusted for individual and neighborhood-level control variables (adjusted odds ratio [AOR] = 1.16; 95% confidence interval [CI] = 1.03, 1.30). No significant association was found between immigrant concentration and trichomoniasis. In model 2, we introduced concentrated poverty into the regression equation and found that young adults who lived in neighborhoods with higher proportions of concentrated poverty had significantly higher odds of having trichomoniasis (AOR = 1.19; 95% CI = 1.02, 1.38). In addition, consistent with our hypothesis, concentrated poverty mediated the relationship between neighborhood concentration of Black residents and trichomoniasis (AOR = 1.05; 95% CI = 0.91, 1.21). These findings suggest that the neighborhood racial disparity in trichomoniasis was attributable to a greater concentration of poverty in these neighborhoods. In model 3, we entered residential instability into the regression equation, but we found no significant associations with trichomoniasis (AOR = 0.87; 95% CI = 0.73, 1.04). However, concentrated poverty remained statistically significant; for every unit increase in concentrated poverty, the odds of infection increased by 25% (AOR = 1.25; 95% CI = 1.07, 1.46).

TABLE 2.

Multilevel Logistic Regression Findings of the Associations Between Neighborhood Social Disorganization and Trichomoniasis Among Young Adults Aged 18 to 27 Years: Add Health, 2001–2002

Fixed Effects	Model 1, AOR (95% CI)a	Model 2, AOR (95% CI)a	Model 3, AOR (95% CI)a
Individual level
Age in y	1.08 (0.99, 1.17)	1.08 (0.99, 1.17)	1.08 (0.99, 1.17)
Male gender	0.65** (0.49, 0.87)	0.65** (0.49, 0.87)	0.65** (0.49, 0.87)
Race/ethnicity
Hispanic	1.72* (1.07, 2.79)	1.75* (1.08, 2.83)	1.75* (1.07, 2.84)
Black	3.45*** (2.41, 4.94)	3.47*** (2.43, 4.96)	3.54*** (2.47, 5.08)
Asian	1.84 (0.96, 3.55)	1.89 (0.98, 3.65)	1.93 (0.99, 3.75)
American Indian	1.53 (0.33, 7.16)	1.53 (0.33, 7.16)	1.52 (0.32, 7.14)
White (Ref)	1.00	1.00	1.00
Foreign-born	0.97 (0.54, 1.74)	0.98 (0.55, 1.75)	0.97 (0.54, 1.75)
Economic hardship	1.72** (1.23, 2.40)	1.66** (1.19, 2.32)	1.65** (1.18, 2.31)
Employed	0.66** (0.52, 0.84)	0.67** (0.52, 0.85)	0.67** (0.53, 0.86)
Enrolled in school	0.65** (0.48, 0.88)	0.66** (0.49, 0.89)	0.67* (0.50, 0.91)
Individual residential stability	0.99 (0.97, 1.01)	0.99 (0.97, 1.01)	0.99 (0.97, 1.01)
Married	0.60* (0.41, 0.89)	0.60* (0.41, 0.89)	0.60* (0.40, 0.89)
Lives with parents	0.87 (0.62, 1.22)	0.89 (0.64, 1.26)	0.86 (0.61, 1.21)
Heterosexual orientation	1.07 (0.72, 1.60)	1.06 (0.71, 1.59)	1.06 (0.71, 1.59)
Antibiotic use in past mo	1.11 (0.79, 1.55)	1.10 (0.78, 1.54)	1.10 (0.78, 1.54)
Age at first vaginal intercourse
18–25 y	1.24 (0.74, 2.11)	1.24 (0.73, 2.10)	1.23 (0.73, 2.08)
16–17 y	1.04 (0.62, 1.76)	1.04 (0.62, 1.77)	1.03 (0.61, 1.75)
10–15 y	1.46 (0.87, 2.46)	1.45 (0.86, 2.44)	1.43 (0.85, 2.40)
Has not had vaginal intercourse (Ref)	1.00	1.00	1.00
No. of vaginal sexual partners in past y	0.99 (0.94, 1.05)	0.99 (0.94, 1.05)	0.99 (0.94, 1.05)
Sexual intercourse with intravenous drug user in past y	0.39 (0.06, 2.56)	0.39 (0.06, 2.56)	0.41 (0.06, 2.78)
Exchanged sexual intercourse for money in past y	1.34 (0.74, 2.41)	1.32 (0.73, 2.39)	1.33 (0.74, 2.40)
Drug use in past y	1.18 (0.90, 1.56)	1.18 (0.89, 1.55)	1.18 (0.89, 1.55)
Level of binge drinking in past yb	1.00 (0.91, 1.10)	1.00 (0.91, 1.10)	1.00 (0.91, 1.10)
Neighborhood level
Neighborhood controls
West region	0.65* (0.44, 0.97)	0.64* (0.43, 0.97)	0.63* (0.41, 0.95)
Midwest region	0.99 (0.71, 1.39)	0.94 (0.68, 1.31)	0.94 (0.67, 1.31)
Northeast region	1.01 (0.66, 1.56)	0.96 (0.62, 1.49)	0.93 (0.60, 1.45)
South region (Ref)	1.00	1.00	1.00
Urbanicity	0.96 (0.79, 1.16)	0.98 (0.87, 1.11)	1.02 (0.89, 1.18)
Neighborhood disorganization
Black concentration	1.16* (1.03, 1.30)	1.05 (0.91, 1.21)	1.03 (0.89, 1.18)
Immigrant concentration	0.96 (0.79, 1.16)	0.89 (0.73, 1.09)	0.90 (0.74, 1.11)
Concentrated poverty		1.19* (1.02, 1.38)	1.25** (1.07, 1.46)
Residential instability			0.87 (0.73, 1.04)
Intercept	0.02*** (0.01, 0.04)	0.01*** (0.01, 0.04)	0.02*** (0.01, 0.04)

Note. Add Health = National Longitudinal Study of Adolescent Health; AOR = adjusted odds ratio; CI = confidence interval. Sample size was n = 11 370 young adults across 4912 neighborhoods.

^aUnweighted analysis.

^bContinuous measure of the number of days during the past year the respondent drank ≥ 5 drinks ranging from never to every or almost every day.

*P < .05; **P < .01; ***P < .001.

We also found several significant individual-level associations (model 3). Specifically, in relation to STI disparities, male respondents were less likely to have trichomoniasis than were female respondents (AOR = 0.65; 95% CI = 0.49, 0.87). Non-Hispanic Black (AOR = 3.54; 95% CI = 2.47, 5.08) and Hispanic young adults (AOR = 1.75; 95% CI = 1.07, 2.84) had approximately 3.5 and 1.7 times the odds of trichomoniasis, respectively, compared with non-Hispanic White young adults. Socioeconomic disparities were also found as young adults who reported economic hardship were more likely to have trichomoniasis (AOR = 1.65; 95% CI = 1.18, 2.31) compared with their more socioeconomically advantaged peers. Being enrolled in school (AOR = 0.67; 95% CI = 0.50, 0.91) and being employed (AOR = 0.67; 95% CI = 0.53, 0.86) were each independently protective against the STI. None of the STI risk factors were significantly associated with trichomoniasis.

DISCUSSION

We found that key indicators of neighborhood social disorganization were significantly related to the acquisition of trichomoniasis among a sample of young adults, above and beyond individual and neighborhood control variables. Specifically, we found that young adults who lived in neighborhoods with a higher concentration of Black residents were more likely to have trichomoniasis compared with those who lived in neighborhoods with lower concentrations of Black residents. However, concentrated poverty mediated this relationship, and, once adjusted, the relationship between neighborhood concentration of Black residents and trichomoniasis was not significant. These findings suggest that the neighborhood racial disparity in young adults’ acquisition of trichomoniasis is attributable to higher levels of concentrated poverty within segregated Black neighborhoods. Our findings are consistent with theory on the pathways through which racial residential segregation contributes to STI outcomes.^18–20 However, previous ecological analyses found that, even after adjustment for community levels of poverty, gonorrhea rates were significantly higher in communities with greater concentrations of Black residents.^15,16 Although not shown, we found similar results, but the association between neighborhood concentration of Black residents and trichomoniasis was not statistically significant once individual-level factors were introduced into the model. Consequently, the disparate findings between our study and previous research may be attributable to our use of multilevel modeling, which enables adjustment of individual-level factors that may potentially mediate or confound relationships between neighborhoods and individual outcomes.²⁵

Consistent with findings in earlier research,^{15–17,21–23} we found significant associations between neighborhood concentrated poverty and young adults’ acquisition of trichomoniasis. These findings illustrate the deleterious relationships between neighborhood poverty and trichomoniasis among young adults, and future research is needed to investigate the pathways through which neighborhood poverty shapes STI outcomes, including racial, ethnic, and socioeconomic disparities. For example, previous studies support the role of social processes, such as social cohesion, collective efficacy, and high-risk behavioral norms in mediating or moderating relationships between neighborhood structural conditions and sexual risk behavior^33–35,37 as well as community STI rates.³⁸ Thus, future studies should explore these potential explanations. In addition, researchers should consider the mediating or moderating role of neighborhood sexual network structure as neighborhood social disorganization may contribute to more closed and assortative sexual networks that increase STI risk compared with more advantaged neighborhoods.³⁹ Lastly, physical disorder also may play a role because previous ecological research found that gonorrhea rates were higher in neighborhoods characterized by boarded-up housing, graffiti, accumulating garbage, abandoned vehicles, and poor physical conditions of public high schools.²¹ Furthermore, the association was independent of neighborhood poverty. The findings support Wilson and Kelling's “broken windows” hypothesis that neighborhood physical disorder may signal neglect, diminished social control, and greater tolerance for high-risk behavior to neighborhood residents and outsiders.⁴⁰ Future multilevel research should explore these potential relationships with a variety of STI organisms.

In contrast to theory^11–14 and previous research,²⁴ neither immigrant concentration nor residential instability was associated with the acquisition of trichomoniasis in our study. First, with respect to immigrant concentration, the majority of our sample lived in neighborhoods that contained few immigrant residents. This limitation may have reduced the power to detect significant relationships that may occur in more segregated neighborhoods. Second, residential instability was found to be negatively associated with adolescent self-reported STI in a previous study that also used Add Health data.²⁴ Methodological differences between our studies as well as developmental differences in our samples may account for the disparate findings. Specifically, our study used multilevel analyses, urine assay data of a specific STI, and neighborhood data collected in 2000 whereas previous research used contextual analyses, self-report of STI, and neighborhood data collected in 1990. Developmentally, young adults also may be more likely to live in neighborhoods with greater residential mobility, such as college campuses and apartment communities. This developmental transition is normative; thus, residentially unstable environments may have less of an impact. Future research on the role of residential instability on health outcomes over the life course is needed to elicit potential developmental differences.

In addition to our neighborhood findings, we also found disparities at the individual level for gender, race/ethnicity, and socioeconomic position. These findings are consistent with previous research and are hypothesized to be attributable to a complex interplay of individual, sexual network, community, institutional, and policy-related factors.^1,2,7–10 Future research should examine these potential mediators and also explore the extent to which relationships between neighborhood social disorganization and STI are moderated by individual sociodemographic factors as neighborhoods may influence the acquisition and transmission of STI differently for individuals with different levels of risk. In contrast to previous research,^8–10 STI risk behaviors were not statistically associated with trichomoniasis in our study. In bivariate analysis (not shown), we did find that trichomoniasis was significantly associated with a greater likelihood of exchanging sexual intercourse for money, younger age at first intercourse (vs no sexual intercourse), and binge drinking, but these relationships were not statistically significant when we included sociodemographic variables in multivariate analyses. These disparate findings may be attributable to measurement error or potentially to differences in the pathogenesis of specific STIs.

Several limitations to our study warrant further discussion. First, this study was cross-sectional; thus, causal inferences cannot be made. Second, Add Health questions related to anal sexual intercourse were asked only in a section on detailed relationships. However, 1852 young adults in our sample did not complete this section (2.8% of these tested positive for trichomoniasis). Consequently, we were unable to adjust for anal sexual intercourse as a risk factor, and we used sexual orientation as a crude proxy. The prevalence of trichomoniasis for those who did not complete the relationship data file was consistent with the overall sample, which suggests that the trichomoniasis risk for those with relationship data and those missing may be similar. In addition, because trichomoniasis is not typically transmitted via anal sexual intercourse,¹ bias from the omitted variable may be minimal. Third, Add Health's school-based design limited the sample to young adults attending school at wave 1 (1995). Thus, the sample does not include high-risk youths in the community at wave I who dropped out of school. The wave III sample does include young adults who participated in wave I, but dropped out of school after their wave I interview. Last, wave III Add Health data do not contain STI prevalence information at any geographic level; thus, we could not capture infectious disease risk at the community level.

Despite these limitations, our study offers evidence that neighborhood poverty is associated with young adults’ acquisition of trichomoniasis, above and beyond known individual risk and protective factors. Furthermore, we found that the neighborhood racial disparity in trichomoniasis was attributable to higher concentrations of poverty in these areas. The Centers for Disease Control and Prevention's Strategic Plan for 2008–2013 calls for a reduction in STI disparities and enhanced efforts to address the social and economic determinants of STI, including the incorporation of structural interventions into their STI prevention efforts.⁴¹ Consequently, further research is needed to better understand the pathways through which neighborhood poverty contributes to STI and STI disparities, including studies on how to effectively create structural change aimed at eliminating neighborhood poverty.