The Influence of Individual, Partner, and Relationship Factors on HIV Testing in Adolescents

Abstract

Early identification of HIV by increasing testing is a national priority; however, little is known about HIV testing behaviors in high school age adolescents. We examined the association of individual, partner, and relationship factors with HIV testing using a computer-assisted survey administered from 2003 to 2006 in a community sample of 980 sexually active 14- to 17-year-olds (56% female, 55% Latino, 25% African American) living in a jurisdiction with a high AIDS burden. Twenty percent reported their first sexual encounter as having occurred when they were <13 years of age, 33% had had four or more lifetime sexual partners, 21% reported high partner HIV-risk behavior, and 428 (44%) had been tested for HIV. In our final regression model, independent associations with HIV testing included being female (OR=1.68 [1.23–2.30]), older (OR=1.41 [1.21–1.65]), and having had four or more lifetime sexual partners (OR=2.24 [1.64–3.05]). The strongest independent predictor of HIV testing was having high HIV-related partner communication (OR=3.70 [2.77–4.94]). Being in a serious committed relationship (OR=1.39 [1.02–1.87]) was also independently associated with HIV testing, whereas reporting high worry about HIV/AIDS (OR=0.53 [0.40–0.71]) was independently negatively associated with HIV testing. High HIV/AIDS knowledge, high partner HIV risk behavior, and young age at first sexual encounter were not associated with testing. These findings suggest that, for high school aged adolescents, optimal strategies to promote HIV testing should look beyond increasing HIV/AIDS knowledge and identifying individual risk behaviors to also considering the role of partners and relationships and their influence on testing behavior.

Introduction

Nearly 60% of adolescents and young adults 13–24 years of age who are living with HIV are unaware of their infection, whereas only 20% of adults with HIV are unaware that they are infected.¹ Although rates of HIV diagnoses in older age groups are stable, the rates of HIV diagnoses in adolescents increased from 2008 to 2011.^2–4 African American and Latino youth carry a disproportionate burden of HIV/AIDS, but report low rates of HIV testing.^2–4

Early identification of HIV status in adolescents is associated with earlier linkage to treatment and better health outcomes.⁵ Earlier treatment of HIV reduces viral load and decreases the risk of future transmission of the virus.⁶ In 2006, the Centers for Disease Control and Prevention (CDC) recommended routine opt-out HIV screening for all persons 13–64 years of age, and the Society for Adolescent Health and Medicine published a statement calling for HIV testing of all sexually active adolescents.^7,8 In 2011, the American Academy of Pediatrics updated their policy to recommend routine screening for all youth by 16–18 years of age, and of all sexually active youth regardless of age in communities where the prevalence of HIV infection is >0.10%.⁹

These national recommendations have not increased HIV testing rates in adolescents of high school age.² The CDC's Youth Risk Behavior Surveillance (YRBS), a nationally representative survey of high school students, showed no significant change in self-reported HIV testing rates from 2005 (11.9%) to 2011 (12.9%) and no change at all from 2007 to 2011.³ In 2011, HIV testing rates in those students who had ever been sexually active were higher in African American students (32%), than in Latino (20%) and white (20%) students.² Trend data from the CDC also show that rates of HIV testing have been increasing more in African American youth than in Latino youth.^3,10

Little is known about HIV testing behaviors in young, high school aged adolescents (14–17 years) including motivations for testing or location and timing of testing. Published studies have focused primarily on young adults (18–24 years) and individual risk factors.^11–21 Most sexually active adolescents do not identify themselves as being at risk for HIV, and have not been tested, despite having knowledge about HIV transmission and engaging in HIV-related risk behaviors such as substance use, unprotected sexual intercourse, and having multiple sexual partners.^{11–14,21,22}

Recently, increasing attention has been given to the influence of partner and relationship factors on sexual health behaviors such as condom and contraceptive use. Studies show that adolescents are more likely to use condoms with casual partners than with steady partners.^23–29 Adolescents in romantic relationships, compared with those in casual relationships, are less likely to report consistent use of condoms and other forms of contraception.^23,24 However, the influence of partner and relationship factors on HIV testing behavior has not been well studied. More than 70% of adolescents and young adults report having had at least one romantic relationship in the previous 18 months.^30,31 A study of African American adolescents and young adults found that those who reported that being in a relationship was highly important to them (had a strong relationship imperative) were more likely to engage in riskier sexual behaviors.³¹ Clearly, adolescent relationships are an important social context that may influence sexual health risk and protective behaviors.

In this study, we examine the associations of HIV testing with individual, partner and relationship factors in a sample of young, predominantly Latino adolescents living in Bronx County, New York, a jurisdiction identified by the National HIV/AIDS Strategy as having among the highest AIDS burden in the United States.³² We also explore HIV testing experiences in participants and their current or most recent partner. We hypothesize that characteristics of the partner relationship, such as partner type and partner communication, are as important, if not more important, than individual characteristics, such as gender, ethnicity, and AIDS knowledge, in predicting HIV testing behavior in a high school age sample.

Methods

Procedure

We performed a secondary analysis of screening survey data collected from two National Institutes of Health (NIH)-funded studies of adolescent sexual attitudes and risk behaviors that took place from 2003 to 2006.^33,34 For both of these studies, names and addresses of patients 14–17 years of age, seen in community practices affiliated with a large Bronx healthcare system over the previous 18 months, were obtained from computerized billing records. Parents were sent opt-out letters from the practice physicians, informing them that their child might be eligible for a paid survey about how adolescents feel about themselves and their friends, drug and alcohol use, and sexual behavior. Parents were asked to indicate if they did not want further information about the study mailed to the home. For those who did not opt out, invitations for study participation were mailed, and adolescents were invited to come to the research center to complete a confidential 60–90 min computer-assisted survey. Participants were compensated $25 for completing the survey. To be eligible, participants had to pass a fifth grade English reading level screen. Written parental permission and assent from the adolescent participant were obtained. Approvals for both studies as well as for the secondary analysis reported here were obtained separately from the Albert Einstein College of Medicine/Montefiore Medical Center Institutional Review Board (IRB).

Participants

A total of 1586 adolescents completed the computer-assisted survey, of which 67% (n=1027) reported being sexually active. We defined “sexually active” as those adolescents who provided a “yes” answer to any of three separate questions asking if they ever had vaginal, oral, or anal sex. We chose to include adolescents who engaged in any type of sexual activity that may increase their risk for HIV. Of the 1027 participants who reported ever having been sexually active, 44% were male and 56% female, 53% were Latino, 25% were African American, 19% were white, and 2% were Asian. This sample is representative of Bronx County with similar demographics reported by the 2010 United States Census for the County of 54% Latino or Hispanic, 30% African American, 11% white, and 4% Asian.³⁵ Because of the small numbers of Asian and other racial/ethnic adolescents in our sample, we excluded these participants from the present analysis. Our final sample included 980 sexually active adolescents with a mean age of 15.9±0.94 years.

Measures

The computer-assisted survey included demographic questions and an array of questions measuring adolescent sexual health and risk behavior.

HIV testing

Fourteen items created for the survey asked about HIV testing history, knowledge of partner's testing and timing of participant's most recent test and their partner's most recent test, motivations for testing and locations of testing. The primary outcome variable for this study was a “yes” or “no” response to, “Have you ever been tested for HIV?” Participants were not asked about their own HIV status.

HIV/AIDS knowledge

A 29-item HIV/AIDS knowledge measure that was previously developed for a study with minority youth with questions about transmission of HIV, protection from HIV exposure, and treatment of HIV, was used.³⁶ The score is the total number of correct responses dichotomized for analysis as high HIV/AIDS knowledge (≥80% correct responses) and low HIV/AIDS knowledge (<80% correct responses).

HIV/AIDS worry

“How worried are you that you might get HIV/AIDS?” answered on a five point scale with 1 indicating “not at all” and 5 indicating “very” worried was used.³⁷ Responses were dichotomized for this analysis as high worry (3–5) and low worry (1–2).

Partner communication

Communication with sexual partners was measured using three previously used scales.³⁸ The first scale, Openness of Communication, contained three items asking about partner general communication (i.e., “Do you and your partner have things to talk about?”; “Does your partner listen when you need someone to talk to?” and “Can you tell your partner how you feel when you are upset?”). The second scale, Comfort with Discussing Sex, contained three items asking about ease of sex-related partner communication (i.e., “Do you find that it is too uncomfortable to talk to your partner about your sex life?”; “Are you willing to talk about your sex life with your partner?” and “Is it easy to talk to your partner about how you like to be touched?”). These two three-item scales each had three response choices, never (1), rarely (2), or always/almost always (3), which were summed with a possible total score of 3–9. The third scale, HIV/AIDS-related Partner Communication, was a five item, “yes” or “no” response to questions that asked: “Have you and your partner ever talked about the AIDS virus?”; “Have you and your partner ever talked about getting tested for the AIDS virus?”; Have you ever asked your partner to be tested for the AIDS virus?”; “Have you ever told your partner that you were worried about getting the AIDS virus from him/her?” and “Have you ever asked your partner to change his/her behavior in any way so that he/she would not get the AIDS virus?”. The HIV/AIDS-specific Partner Communication scale score was the total number of “yes” responses. Scores of all three scales were dichotomized at their medians as “high” (score≥median) or “low” (score<median).

Self and partner HIV/AIDS risk behavior

Self and partner HIV/AIDS risk behavior was assessed using items from previous studies on adolescent HIV/AIDS risk.^39,40 For individual risk behavior, we used age at first sexual intercourse, which we dichotomized as age ≥13 years or <13 years, and number of lifetime sexual partners, which we dichotomized as ≥4 partners or <4 partners. These are similar to the dichotomies used by the CDC's YRBS.⁴¹ To assess the known increased HIV/AIDS risk in males who have sex with males (MSM), male participants were asked “Have you ever had any sexual contact with a male?” To assess HIV/AIDS risk from intravenous drug use, participants were asked a series of yes or no questions “Did you ever inject cocaine, ecstasy, steroids, heroin, crack, amphetamines, tranquilizers or hallucinogens?”

We used five items, each with a possible “yes,” “no,” or “don't know” response to measure risk history of the participant's current or most recent previous sexual partner. The items included having a partner who they knew or suspected had traded sex for money, drugs, food or shelter; had HIV/AIDS; or was an intravenous drug user; and whether they had ever had a partner who they knew or suspected had been a partner of an HIV/AIDS-infected person or a partner of an intravenous drug user. For our analyses, participants were categorized as “high” in partner HIV/AIDS-risk behavior if they answered “yes” to any of these five questions and “low” in partner HIV/AIDS-risk behavior if they answered “no” or “don't know” to all five items.

Relationship type

Relationship types were measured as in a “committed/serious relationship,” “in a relationship but not really serious,” “starting in a new relationship,” “dating/seeing,” or a “one-night stand.” For this analysis, we categorized adolescents according to whether or not they reported being in a serious, committed relationship versus any other type of relationship in the prior 6 months.

Relationship uncertainty

To assess relationship uncertainty regarding current sexual partners, we asked the degree of agreement with the statement, “My partner might be having sex with someone else.” Participant's responses on a four point Likert-type scale were dichotomized as agree (strongly agree and agree) versus disagree (disagree and strongly disagree).

Analysis

We used descriptive statistics to report the demographics and risk behaviors of the participants, their rates of HIV testing, and the characteristics and information regarding their most recent HIV testing experiences. Bivariate comparisons of the outcome variable, self-report of HIV testing, by gender, age, ethnic group, and by other independent variables addressing individual, partner, and relationship factors were performed using χ² tests, with post-hoc pairwise comparisons as needed. Variables that were significant at the bivariate level were entered into a stepwise logistic regression model to identify those that were independently associated with HIV testing. Given the cross-sectional study design, model results were interpreted in terms of the probability or odds of self-reported HIV testing being associated with individual, partner, and relationship factors. A p value≤0.05 was used to define statistical significance for all analyses. All analyses were conducted using SPSS Version 19 (Chicago IL).

Results

Characteristics of the sample

Of the 980 sexually active adolescents in our sample, more than half were female (56%) and Latino (55%) (see Table 1). Most participants had Medicaid as their health insurance, and 33% reported receiving welfare or public assistance benefits. Twenty percent of the participants had had their first sexual intercourse before the age of 13, one third reported having had four or more lifetime sexual partners, and one third had a history of unprotected sex (sex without a condom) in the past 6 months. One in five participants reported having a partner with high HIV/AIDS risk behavior.

Table 1.

Characteristics of Participants

	%
Gender
Female	56
Male	44
Race/ethnicity
Latino	55
African American	25
White	19
Age
14 and 15 years	30
16 and 17 years	70
Resides with both parents	25
Medicaid insurance	62
Mostly or only Spanish spoken at home	10
Welfare or public assistance	33
First sex <13 years of age	20
Four or more lifetime sex partners	33
Unprotected sex in past month	32
High partner HIV risk behavior	21

n=980, mean age=15.9±0.94 years.

HIV testing rate and experiences

Most participants (56%) had not been tested for HIV. Of the 428 participants who had been tested, 66% reported knowing that their current or most recent partner had been tested for HIV/AIDS, 42% reported going with their current or most recent partner when he/she was tested, and a total of 44% reported that they had been tested within 60 days of their partner's most recent test (see Table 2). Almost 30% reported having been tested on the same day as their partner. The most common motivations for testing reported by participants included wanting to know their status (71%), and that their healthcare provider suggested the HIV test (19%). The most common locations of HIV/AIDS testing included a healthcare clinic (54%), private doctor's office (17%), and hospital (14%). Very few participants (7%) reported being tested at an HIV-testing site.

Table 2.

Participant's HIV Testing Experiences

Question	%
Knowledge of partner's testing status
“Yes”	66
“No”	9
“I Don't Know”	25
Did you go with your current or most recent partner when he/she tested for HIV?
“Yes”	42
Timing of participant's most recent HIV test in relation to their partner's most recent test
Same day	28
Within 30 days	6
Within 60 days	10
Motivations for HIV testing
“I wanted to know my status”	71
“I wanted to be tested before I had sex with my current partner”	4
“My health care provider (doctor) suggested it”	19
Location of HIV testing
“Health care clinic”	54
“Private doctor's office”	17
“Hospital”	14
“HIV-testing site”	7

n=428; of 980 participants, 428 were tested for HIV.

Individual factors associated with HIV testing

We found that a higher proportion of girls and older adolescents (ages 16–17) reported having been tested for HIV than boys and younger adolescents (ages 14–15); there was no difference in HIV testing rate by race/ethnicity (see Table 3). We also found no difference in HIV testing by level of HIV/AIDS knowledge (high vs. low). However, we found a positive association of HIV testing with having had four or more lifetime sexual partners compared with having had fewer partners. In addition, males who reported sexual contact with other males were significantly more likely to have been HIV tested than males without same sex contact. No participants in our sample reported injection drug use. We found a negative association between HIV/AIDS worry and HIV testing. Participants who reported high worry about getting HIV/AIDS were significantly less likely than those with less worry to have been tested.

Table 3.

Factors Associated with HIV Testing

	HIV tested %	p Value
Individual factors
All participants (n=980)	44
Gender
Female (n=552)	51
Male (n=428)	35	<0.001
Age
14 and 15 years (n=297)	32
16 and 17 years (n=683)	49	<0.001
Race/ethnicity
Latino (n=542)	45
Black (n=249)	43
White (n=189)	42	ns
HIV/AIDS knowledge
High (n=258)	47
Low (n=722)	42	ns
Early age at first sex (<13 years old)
Yes (n=194)	48
No (n=786)	43	ns
No. of lifetime sex partners
Less than 4 (n=658)	39
Four or more (n=322)	53	<0.001
Males with sexual contact with males
Yes (n=27)	56
No (n=401)	33	<0.05
HIV/AIDS worry
High (n=388)	36
Low (n=592)	49	<0.001
Partner or relationship factors
HIV-related partner communication
High (n=494)	60
Low (n=486)	27	<0.001
Openness of communication
High (n=448)	48
Low (n=509)	41	<0.05
Comfort discussing sex with partner
High (n=456)	48
Low (n=501)	41	<0.01
Relationship type
Serious committed (n=371)	54
Not serious committed (n=609)	37	<0.001
Partner HIV-risk behavior
Yes (n=361)	47
No (n=619)	42	ns
Relationship uncertainty
Yes (n=155)	43
No (n=802)	45	ns

Partner and relationship factors associated with HIV testing

We found significantly higher rates of HIV testing among participants who scored high on each of the three scales measuring partner communication: Openness of Communication, Comfort with Discussing Sex, and HIV-specific Partner Communication (see Table 3). Relationships that were reported as “serious and committed” were also more likely to be associated with HIV testing than any other relationship type. In contrast, participants who reported high partner HIV/AIDS risk behavior and those who were concerned that their partners were having sex with someone else were not more likely to report HIV testing than those without such concerns.

Predictors of HIV testing

Several factors were independently associated with HIV testing in a stepwise logistic regression model (see Table 4). Demographics including female gender (OR=1.68) and older age (OR=1.41) were independently associated with testing. However, the strongest independent predictor of HIV testing was having a high HIV-related partner communication (OR=3.7). One other relationship factor, being in a “serious/ committed” relationship, was also independently related to HIV testing (OR=1.39). Two additional individual factors were also independent predictors. Having had four or more lifetime sexual partners (OR=2.33) was positively associated with HIV testing, whereas high worry about getting HIV/AIDS was negatively associated with HIV testing (OR=0.53).

Table 4.

Logistic Regression Model

	Odds ratio (confidence interval)	p Value
Gender
Female	1.68 (1.23–2.30)	0.001
Age
Older age	1.41 (1.21–1.65)	<0.001
No. of lifetime sex partners
Four or more partners	2.24 (1.64–3.05)	<0.001
HIV/AIDS worry
High	0.53 (0.40–0.71)	<0.001
HIV-related partner communication
High	3.70 (2.77–4.94)	<0.001
Relationship type
Serious committed	1.39 (1.02–1.87)	<0.05

Discussion

Despite the frequency of high-risk sexual behavior reported by adolescents in this sample, the majority had not been tested for HIV. Our testing rate of 44% is, however, twice the 22% rate reported by sexually active high school students in the 2011 YRBS, a nationally representative sample.⁴² There are only a few community studies describing rates of HIV testing in adolescents and young adults.^11–20 Most of these studies intentionally sample sexually active African American young adults in high-risk urban settings. Rates of testing in these studies range from 34% to 63%, with higher rates reported in populations of youth who identify as gay, lesbian, or bisexual, live in correctional facilities or report high-risk sexual activity. The testing rate we found in our community-dwelling Bronx sample is within this previously reported range, consistent with the high-risk behaviors reported by our participants.

Consistent with prior studies, we found that a higher proportion of girls and older adolescents reported HIV testing than did boys and younger adolescents.^13,18,21 The increased testing rate found in females may be related to increased utilization of healthcare services by women in general, and specifically for sexually transmitted infection (STI) and pregnancy concerns than by males.¹⁸ Unlike prior studies, we found no ethnic differences in HIV testing among Latino, African American, and white participants, despite having sufficient power to detect 5–10% differences among groups. Most other studies find a higher rate of both HIV/AIDS and HIV testing in African Americans.^{4,11,12,16,17,21} However, the Bronx is a unique urban environment with a predominance of Latino residents, and the highest rates of HIV/AIDS diagnoses in Latinos living in New York City.⁴³ The CDC reports that the rate of new diagnoses of HIV infection among Latinos in the Northeast census region was more than twice as high as in other regions.⁴⁴ Therefore, it is not surprising that Latino teens were tested as frequently as African American teens in our sample. White adolescents comprised 20% of our sample; however, we found no differences in testing rates compared with Latino or African American participants. Perhaps similarities in social factors, including risk behaviors, access to care, and socioeconomic status in the multiethnic Bronx community influenced HIV testing rates more than did race alone. Our multiethnic sample of adolescents living in a county with a heavy HIV/AIDS burden is a priority population in which to explore HIV testing behaviors to better inform intervention efforts.

To the best of our knowledge, we are the first to report the role of partners in adolescents' HIV testing experiences. A majority, 66%, of our high school aged participants who had been tested were aware of their partner's testing status. Forty-two percent went with their current or most recent partner when they were tested for HIV, and almost 30% were tested on the same day as their partner. These findings highlight opportunities for healthcare providers to include a discussion of the partner's testing status when counseling adolescents about HIV testing.

We found that the majority of our participants reported testing in a healthcare clinic and only a few, 7%, reported testing at an HIV testing site. These locations of testing are similar to those reported by young adults (18–24 years) in the nationally representative 2009 Behavioral Risk Factor Surveillance System study: 70% tested at healthcare clinics and 3.1% at HIV testing sites.⁴⁵ As an age group, rates of preventive services delivery are generally low in adolescents, making it less likely for these youth to access routine screening.⁴⁶ Only ∼20% of our study participants reported that their motivation for testing was that a doctor recommended it, despite 54% having been tested in a healthcare setting, suggesting that our participants may have been tested in an emergency department or other healthcare site or that a healthcare worker other than a doctor recommended the testing. More high-risk minority youth who are found to be HIV positive report lacking a primary care provider.⁴⁷ Moreover, many primary care providers miss opportunities to provide HIV and STI preventive services.⁴⁸ Kaiser Family Foundation's National Survey of Teens in 2000 found that two thirds of adolescents ages 15–17 were not sure about where to get tested for HIV.²² Taken together, our study findings and these prior studies highlight the importance of increased screening at preventive healthcare visits as well as increasing the awareness of alternate testing venues that are available to adolescents living in high-risk communities.^49,50 For example, Calderon et al. in a study conducted, like ours, in Bronx County, New York, showed that testing in community pharmacies was feasible for adults, and may prove equally helpful for adolescents.⁴⁹

We found that the most common motivation for getting an HIV test was “I wanted to know my status” (71%), whereas only 19% reported getting HIV tested because their healthcare provider suggested it, and only 4% reported wanting to be tested prior to initiating sexual activity with their partner. This finding is similar to motivations for HIV testing described by high-risk minority young adults (18–25 years), with 47% reporting “peace of mind,” 15% reporting that testing was “part of a physical exam.” Only 2% reported “testing prior to sex with partner.” ⁵¹ The low rate of mention of partners is surprising, given the proportion of adolescents in our study who either went with their partner to their test or got tested in close proximity to their partner's test.

Recently, increased attention has been paid to the role of partners and relationships in protective and sexual risk behaviors in adolescents and young adults.^{23–28,51,52} Our study examined the roles of individual, partner, and relationship factors with HIV testing in an effort to inform high-impact interventions in minority adolescents at risk for HIV.

Among individual risk factors that we examined in our sample of high-school-aged adolescents, a high number of lifetime sexual partners, and males reporting sexual contact with other males, were positively associated with HIV testing, whereas high knowledge about HIV/AIDS and young age at first sexual intercourse were not associated with HIV testing. Prior studies of adolescents and young adults also find that a history of having had four or more lifetime partners and being MSM are associated with testing.^13,19,21 Our findings show that high school aged adolescents with the most risky behavior are more likely to get HIV tested. Nevertheless, more than half of our participants were not motivated to be tested, confirming the impression that many adolescents do not consider themselves at risk even though their behavior is risky.^11–14,53 A study on acceptance of an HIV test offer showed that even after administration of a self-risk checklist assessment tool, only ∼ 50% of adolescents at risk accepted an offer of an HIV test, with African American adolescents accepting more often than Latino adolescents.⁵⁴ HIV/AIDS knowledge and awareness of an adolescent's own risk behavior may be necessary, but not sufficient, to change health risk or protective behavior.

Among the partner and relationship factors we examined, consistent with our hypothesis, high levels of partner communication and a “serious committed” relationship type were positively associated with HIV testing. As shown previously in two large community samples of adolescents, we also found that partner HIV-risk behavior and relationship uncertainty were not associated with HIV testing, suggesting that adolescents do not recognize these as risk factors for HIV or that partner's risk behaviors were not motivators for them to obtain risk-reduction measures such as HIV testing.^11,52 Greater efforts must be made to increase testing for at-risk adolescents who have high-risk partners. It is possible that for many young adolescents, they are at risk only because of past or concurrent risk behaviors of their partners.

In our final regression model, we found that both partner communication about HIV/AIDS and a “serious committed” relationship type were independently associated with HIV testing, with partner communication being the strongest predictor. Participants who had high HIV/AIDS-related partner communication were 3.7 times more likely to have been tested than were those with low communication. Only one prior study of 900 African American adolescents (13–18 years of age) evaluated the association of HIV testing with HIV-specific communication (number of times partner reported talking about HIV or STI testing), and found an independent association, but only in males.¹³ A strength of our study is that we examined HIV/AIDS-related communication, comfort with discussing sex, and openness of communication with partners separately; although all were positively associated with HIV testing at the bivariate level, only HIV/AIDS-specific communication was independently associated with testing. Fostering open and comfortable communication about HIV/AIDS between adolescent sex partners could be important for HIV testing interventions in the community. Our findings suggest that encouraging HIV/AIDS-specific communication between adolescents and their partners may enhance HIV testing rates.

Adding to the limited current literature, we found that being in a serious committed relationship was also independently associated with HIV testing in our high school aged sample. Two studies of older adolescents and young adults from minority ethnic groups found that rates of HIV testing were higher in youth in serious relationships versus those in casual relationships, in those who reported having had one sexual partner in the past 3 months, and in those who reported having lived with two or more serious partners.^51,54 Only one study of Ohio high school aged adolescent couples reported on relationship characteristics and found that HIV testing was associated with romantic love, but not with sexual non-exclusivity or partner's risk behaviors.⁵² It makes sense that youth in serious relationships who have greater opportunity for communication about HIV/AIDS may report higher testing rates. Youth who have less committed, less serious partners or non-exclusive partners, however, may be at higher risk for HIV, and, therefore, it is discouraging that they are less likely to be tested. Intervention studies are needed to target these at risk adolescents.

Our finding of an independent negative association between worry and HIV testing in our regression model was not surprising. Fear is commonly reported by adult patients in urban clinics as a reason for refusing HIV testing.⁵⁵ Adolescents, however, often do not perceive their own risk, despite their risk behaviors. In addition, our adolescent participants who had been HIV tested were less worried about getting HIV/AIDS. This suggests that future studies are needed to prospectively examine the impact of a negative HIV test on sexual behavior. A 6 month longitudinal study of 636 high school aged African American adolescents found that there was no effect of a negative STI test on participant's future number of partners or unprotected sex.⁵⁶ A qualitative study of urban minority young adults (18–25 years of age) found that these youth use STI/HIV screening behavior to justify discontinuing condom use with a serious partner.⁵⁷ This study also found that youth do not maintain monogamy even in “long-term, serious relationships.”⁵⁷ Adolescent relationships are dynamic, of variable duration, and often not monogamous.^28–31,58 Younger adolescents value romantic relationships for their friendship, physical attraction, and emotional intensity, whereas older adolescents value commitment and caring support in their relationships.²⁸ HIV testing may, therefore, play different roles in younger than in older adolescents' relationships.

Our results must be interpreted in light of certain limitations. First, our cross-sectional study design limits causal inferences. The temporal relationship between HIV testing and related risk behaviors also cannot be determined. Second, our primary outcome variable was self-report of HIV testing, and is, therefore, subject to recall bias and potential misreporting of sensitive information. The sexual health risk behaviors included in our survey are similar to those in the YRBS, which also relies on self-reported data. A reliability study of the YRBS questionnaire indicated that students generally report consistent health risk behavior. Additionally, reliability of self-reported HIV/STI measures among community samples of minority adolescents has also previously been shown.⁵⁹ Finally, a strength of our study is that we used a computer-assisted survey tool, which has been shown to increase disclosure of sensitive information in adolescent study populations.⁶⁰

Although our community sample of high-school aged adolescents may not be widely generalizable, our participants live in a county with a high HIV/AIDS burden, and are predominantly Latino, therefore representing a high priority population in which to study HIV testing behaviors in order to inform intervention programs.

Conclusions

Our results add to prior work showing that knowledge about HIV/AIDS and HIV/AIDS risk status are not associated with higher rates of HIV testing in adolescents. Therefore, focusing on increasing HIV/AIDS knowledge or identifying an individual adolescent's risk behavior may not be optimal intervention strategies to increase HIV testing rates. Instead, our findings suggest that for the adolescent age group, explorations of partner and relationship characteristics may be more promising pathways for potential intervention. Partner communication about HIV testing status may be an especially important modifiable factor, as we found that it had the strongest independent association with HIV testing in our adolescent sample. For adolescents, promotion of HIV testing through routine screening, as is widely recommended, should take into consideration the role of partners and relationships.

Acknowledgments

Funding for this study was provided by the National Institute of Mental Health (R01MH070299) and National Institute on Drug Abuse (R01DA019095) awarded to Dr. Bauman. We are grateful for the assistance of the survey staff who recruited the participants and helped collect the data. We also want to thank the adolescents who contributed their experiences to the study.

Author Disclosure Statement

No competing financial interests exist.