Objective. To assess sexually active adolescents' knowledge, attitudes, and behaviors associated with human immunodeficiency virus (HIV) testing and to determine the factors important in their decision to obtain voluntary HIV testing.
Design. Anonymous, random, digit-dial telephone survey undertaken in 1993.
Setting. Massachusetts households.
Participants. Adolescents, 16 to 19 years of age.
Results. Of the 567 adolescents surveyed who had sexual intercourse within the past year, 127 (22%) had received HIV testing, with 54 (10%) stating that this testing was for personal reasons. A “great deal” or “some” worry about getting HIV/acquired immunodeficiency syndrome (AIDS) was expressed by 51%, and 56% felt that it was at least a little likely that they will get AIDS. Misconceptions were common about aspects of HIV testing: 35% did not believe or did not know that the HIV test results were kept in confidence, 19% thought that AIDS testers informed partners if the results were positive, and 30% did not think that the HIV test was very accurate. Although 92% (452/490) had seen a physician in the past year, only 30% (136/452) had ever discussed AIDS with a doctor. Multivariable analysis identified five factors as independently associated with voluntary adolescent HIV testing: 1) having had more than one sexual partner within the past year [odds ratio (OR): 2.9; 95% confidence interval (CI): 1.5, 5.5]; 2) believing that condoms are only somewhat effective at preventing the spread of AIDS (OR: 2.6; 95% CI: 1.4, 4.8); 3) having discussed AIDS with a doctor (OR: 2.6; 95% CI: 1.4, 4.8); 4) not having had a teacher discuss AIDS (OR: 2.2; 95% CI: 1.2, 4.2); and 5) believing that a positive test result means one has AIDS as opposed to carrying the virus (OR: 2.0; 95% CI: 1.1, 3.7). High-risk behavior of infrequent condom use and a history of a sexually transmitted disease were not significantly associated with voluntary HIV testing.
Conclusion. Among sexually active Massachusetts adolescents, voluntary HIV testing is uncommon. Teens who have had multiple sexual partners and who do not believe condoms are effective in preventing transmission were most likely to have been tested. Issues requiring clearer communication to patients include the testing process, its availability, and confidentiality. Physicians can play an influential role in the promotion of HIV testing by discussing HIV risk behaviors with patients and offering those at risk voluntary HIV counseling and testing.
In this second decade of the acquired immunodeficiency syndrome (AIDS) epidemic, the spread of human immunodeficiency virus (HIV) infection to uninfected persons continues with more frequent occurrence in adolescents as a result of sexual transmission. In the US as of June 30, 1995, 18.3% of all AIDS cases (87 294) occurred in people 20 to 29 years old.1 In many US communities, AIDS is the leading cause of death among young adults 25 to 44 years of age.2 Because the median time for development of AIDS after HIV infection is 11 years, many of these people contracted the virus in adolescence.3 Rosenberg4 plotted the calculated age at HIV infection for black, white, and Hispanic men and women, and in each case the steep rise in incidence occurred before age 20. The World Health Organization estimates that as of March 1993, half of the 14 million HIV-infected people worldwide were infected between the ages of 15 and 24 years.5 Of 269 956 Job Corps applicants 16 to 21 years of age screened between 1988 and 1992, 812 were HIV-seropositive (3 per 1000).6 D'Angelo et al7 reported the HIV seroprevalence of urban adolescents 13 through 19 years of age receiving ambulatory care and having blood drawn for other routine medical indications to be 0.37% (3.7/1000).
This phenomenon of adolescent HIV infection is not surprising considering the reported behavior of sexually active adolescents. Hingson et al8 found that 70% of sexually active adolescents in Massachusetts were either never using condoms or using them only some of the time. The Centers for Disease Control and Prevention found that more than half of adolescents in the United States reported having unprotected sexual intercourse by the age of 19 years.9
Identifying and linking HIV-infected persons to medical care are important because these measures enable early intervention with prophylactic and therapeutic treatments.10 Appropriate behavioral change to limit viral transmission, particularly regarding safe sexual practices, may be most effective if adolescents are aware of their HIV serostatus. Several studies have determined that the majority of patients with HIV present to medical care with CD4 lymphocyte counts suggestive of chronic infection for ≥5 years.11,12 Similar experience of advanced immunodysfunction has been documented in a cohort of 50 New York City youths, in which 48% had CD4 cell counts <500.13 In blinded studies, D'Angelo et al7 found that the majority of adolescents with HIV infection have never undergone HIV testing.
Thus, HIV testing of sexually active adolescents is a critical issue. In the initial years of HIV infection, a period that is frequently clinically asymptomatic, a person can transmit the virus to a sexual partner. Although the adolescent knows that high-risk behavior for HIV infection such as unprotected sexual intercourse has occurred, during this asymptomatic period he or she is unaware of HIV serostatus unless HIV testing has been performed. The consequences on safe behaviors of an adolescent receiving a negative HIV test result are uncertain. It may serve as a potent influence to both reduce unsafe behaviors and pursue safe behaviors or, alternatively, may have the opposite effect, endorsing previous risky behaviors. In a randomized, controlled trial, college students who received HIV testing plus education questioned sexual partners about their HIV status more than subjects receiving education alone or in the control group.14 Knowledge of positive HIV serostatus has the potential to enable one to enter medical care early, avoid potential health-threatening situations (eg, exposure to toxoplasmosis and chickenpox), and obtain the greatest benefit from medical care, with both positive personal and public health implications. However, these potential benefits are not necessarily realized, as in a New York City experience in which many HIV-positive adolescents continued high-risk behaviors such as intercourse without condoms.13
In an attempt to better understand adolescent HIV testing among those at risk, we analyzed a survey of sexually active 16- through 19-year-olds in Massachusetts to address the following questions: 1) What knowledge, attitudes, and behaviors concerning HIV testing are found among sexually active adolescents? 2) What factors are significantly associated with voluntary HIV testing among sexually active adolescents? 3) What factors are not important in the decision to undergo HIV testing?
Between April 1993 and September 1993, 906 adolescents, 16 to 19 years of age, were surveyed in an anonymous random-digit-dial telephone survey in Massachusetts in a study of behavioral risks associated with HIV using methods developed by Waksberg.15 Within each household, one 16- to 19- year-old was selected using methods developed by Kish.16 Calls were made Monday through Thursday, 9 AM to 9 PM; Friday, 9 AM to 3 PM; Saturday, 10 AM to 3 PM; and Sunday, 3 PM to 9 PM. Potential respondents were told that the survey concerned AIDS transmission and their own behaviors in order to plan educational programs about AIDS. They were also told that their answers were confidential, their names and addresses were unknown, and their telephone numbers were drawn at random. The subjects were asked whether they preferred to do the interview at present or at an alternate time, allowing them the opportunity to avoid answering questions in a situation not conducive for honest responses. At the end of the interview, the interviewer gave the teenager a toll-free AIDS hotline to call if he or she had any questions about AIDS. The questionnaires were administered in English or Spanish as needed. The response rate was 77% among residential households containing 16- to 19-year-olds. Subjects were questioned about sexual behavior, substance use, attitudes toward these activities, demographics, knowledge of HIV infection, and concerns as a result of the HIV epidemic. These questions considered HIV issues from the perspective of the Health Belief Model and the Theory of Reasoned Action as well as from beliefs about testing and efficacy of treatments for HIV.17Subjects were asked whether they had ever undergone HIV testing and whether this testing had been voluntary for personal reasons. Nonvoluntary reasons for testing were blood donation, pregnancy care, and hospitalization or surgical procedure, or was a requirement for health insurance, life insurance, military service, a job other than military service, immigration, or some other activity. The option concerning voluntary testing was read as follows: “It was a voluntary decision on your part because you personally wanted to know if you were infected.” The subject could identify more than one reason.
Eligibility criteria for this study included those people surveyed who had sexual intercourse in the past year and either had not tested for HIV or had listed HIV-tested for personal reasons. The independent variables studied are listed in Table 1. Variables found to be statistically significant in bivariate analyses (two-tailed, P < .05), with the outcome of interest being voluntary HIV testing, were examined for colinearity. A correlation analysis was performed, and one of any pair of variables found to have a correlation coefficient >0.4 was eliminated from additional analysis. The remaining significant variables were clustered into three conceptually similar groups for the purpose of performing the multivariable analysis: 1) beliefs and knowledge issues, 2) previous experiences, and 3) behavioral factors. A logistic regression analysis was performed by cluster. Variables from each cluster that remained statistically significant were entered into a final logistic regression analysis.
This research study was approved by the Institutional Review Board of Boston University School of Public Health. Because the interviews were anonymous, verbal rather than written consent was obtained after the purpose of the study was explained to potential respondents.
Of the 906 adolescents surveyed, 567 (63%) stated that they had sexual intercourse within the past year. Of these, 563 (99%) answered the question about previous HIV testing. A total of 127 (22%) had undergone previous testing, with 54 (10%) stating that this testing was for personal reasons. The 73 subjects who did not test for personal reasons were excluded from this analysis; thus, 490 subjects met eligibility criteria. The 73 excluded subjects were similar to the 490 in the study sample in terms of gender and race; however, those excluded were significantly older. In a comparison of the 73 excluded subjects with the 436 nontesters, we found that the nonvoluntary testers were very similar to the nontesters. In fact, after controlling for age, none of the 26 other variables was significantly different between the two groups. Comparing the 73 excluded subjects with the 54 who tested for personal reasons, we found that in multivariable analysis, the nonvoluntary testers differed from the voluntary testers in two ways: the former group felt that condoms were effective at preventing the spread of AIDS (P = .004) and did not usually drink while having sex in the past 6 months (P = .02). Consistent with the objectives of the study, these findings support our original approach to analyze only those that tested for personal reasons.
Explanations for HIV testing, other than for personal reasons, offered by the 73 adolescents included blood donation (n = 23), pregnancy care (n = 11), hospital procedure (n = 15), health insurance (n = 3), life insurance (n = 1), job requirement (n = 5), military requirement (n = 8), and immigration requirement (n = 2), or were because of other activity (n = 10). It was possible to provide more than one reason for testing; for example, 9 of 54 subjects in the voluntary testers group stated that pregnancy care was a reason for HIV testing.
The population under study is described in Table2. Subjects were distributed similarly by age and gender. Ethnic minorities were oversampled for Massachusetts. The vast majority reported sexual activity with the opposite gender only. Nearly half reported having had more than one sexual partner in the past year, and 22% (107/488) had not used a condom in the past 6 months. Injection drug use was identified by <1% (1/488).
Attitudes about and understanding of HIV, AIDS, and medical care among this group are described in Table 3. A “great deal” or “some” worry about getting HIV/AIDS was expressed by more than half of those surveyed, and 56% felt that it was at least a little likely that they will get AIDS. One quarter believed that if “tested for the AIDS virus today,” it was at least a little likely to “show you are infected.” An understanding of HIV transmission was the norm, with the exception of the risk of giving blood. Less than half acknowledged having heard a message in school about unsafe sex. Almost half had heard of AZT (zidovudine), whereas only 12% had known anyone with AIDS.
Misconceptions were much more common about aspects of HIV testing. More than one third of subjects did not believe or did not know that the HIV test result was kept in confidence. A total of 19% thought that AIDS testers informed partners if the results were positive, and 30% (148/490) did not think that the HIV test was very accurate. Of those not tested who said they would want to get tested, 30% (78/260) did not know where to go for HIV testing. Although 92% (452/490) had seen a physician in the past year, only 30% (136/452) had ever discussed AIDS with a doctor. Of those doctor–patient discussions, 85% (116/136) were physician-initiated.
Bivariate analysis found 18 factors significantly associated (P ≤ .05) with voluntary HIV testing among this sample of adolescents. These factors were divided into three clusters as noted in Table 4: beliefs and knowledge issues, previous experiences, and behavioral factors.
We had complete data sets on 94% (461/490) of the subjects interviewed; thus, this group comprised our sample for multivariable analyses. Among these subjects, 409 (89%) had no previous HIV testing, and 52 (11%) had tested for personal reasons. Tests for colinearity found the following pairs of highly correlated variables: effectiveness of condoms in preventing AIDS and effectiveness of condoms in preventing sexually transmitted diseases (STDs); effectiveness of condoms in preventing AIDS and effectiveness of condoms in preventing pregnancy; and in the past 6 months, when subjects had sex, how often had they been using drugs and had they ever sold drugs to others? The last variable of each of these three pairs was excluded from subsequent analyses.
Logistic regression analyses of the clusters of variables found five beliefs and knowledge issues and three each of the previous experiences and behavioral factors that were independently and significantly associated with HIV testing for personal reasons. These factors are noted by asterisks in Table 4. As shown in Table5, the final regression analysis identified five factors as independently associated with adolescent HIV testing in our sample: 1) having had more than one sexual partner within the past year; 2) believing that condoms are only somewhat effective at preventing the spread of AIDS; 3) having discussed AIDS with a doctor; 4) not having had a teacher discuss AIDS; and 5) believing that a positive test result means one has AIDS as opposed to carrying the virus.
Several variables considered as potentially important in the decision to undergo HIV testing were not found as significant factors in this sample. Age and gender were not associated with HIV testing. Use of alcohol or marijuana had no association with the outcome of interest. Even those who felt that undergoing AIDS testing was likely to show that they were infected were not more likely to test. High-risk behavior of infrequent condom use and a history of an STD were not significantly associated. An a priori hypothesis that knowing someone with AIDS would be a factor associated with HIV testing was also not found in bivariate analyses (P = .15).
The results of this study indicate that clinicians are an influential source of HIV information. In studies in which clinician behavior is examined, it has been documented repeatedly that clinicians who address harmful patient health behaviors effect positive changes in these behaviors.18,19 Such evidence has been reported with regard to adolescent condom use and adult alcohol abuse. This survey indicates that physicians can play an important role in promoting HIV testing in at-risk adolescents. Adolescents who had discussed AIDS with physicians were 2.6 times more likely to have been HIV tested. This was one of the strongest predictors of HIV testing. Because 93% of all respondents had a clinician visit in the past year, the opportunity for clinicians to impact patient behavior is significant. Our finding is consistent with the work of Goodman et al,20 in which adolescent at-risk girls who had repeated discussions with a health care provider about HIV testing were more likely to receive HIV counseling and testing.
Capitalizing on this opportunity is a challenge. Active participation by physicians in providing preventive services for adolescents is a formidable and as yet unfulfilled responsibility.21Regarding HIV issues, practitioners may not be as vigilant about asking questions as they are about other psychosocial risks. In our study, substance use was discussed with doctors in 48% of those surveyed, whereas HIV was brought up only 28% of the time, with 85% of these discussions initiated by the physician. In a busy office practice, the time at which a discussion of substance use occurs may be an opportune time to incorporate HIV information. In fact, physician inquiry about HIV-related issues including testing was desired in a survey of adolescents, even though only 27% reported ever discussing HIV with a physician.22
Physician counseling is particularly important given the shortcomings of school-based education in this area. Although Massachusetts mandates HIV education, only 43% had safer sexual practice discussions in school. Counterintuitively, one of the most significant predictors of testing in this survey was whether the adolescent had not had any classroom instruction. This is in contrast to the 1988 AIDS supplement to the National Health Interview Survey in adults, in which one of the factors significant for predicting voluntary testing was increased knowledge of HIV/AIDS.23 In addition, Strehlow and Kampmann24 in Germany found that adolescents who were well informed about AIDS, depressed, anxious, and sexually active were more likely to undergo an HIV test. School-based education regarding AIDS risk reduction may require more than a simple message, which was possibly the exposure for adolescent respondents in this survey.25
This survey supports the observation that adolescents do not seek HIV testing. Although most respondents in this survey believed that they would contract AIDS at some point in their lifetime and one quarter of the respondents suspected that if they were tested at the time of the interview, there was a likelihood they would be HIV-positive, only 10% had pursued voluntary HIV testing. Is it something in the adolescents' beliefs and knowledge about HIV, their understanding of the testing process, and/or their access to the test that contribute to such low frequency of testing despite their high concern?
The results suggest that some at greatest perceived risk but not necessarily greatest actual risk are more likely to be tested. For example, one factor associated with voluntary HIV testing was the belief that condoms are only somewhat effective at preventing the spread of AIDS. The association with HIV testing of this perceived risk factor is in contradistinction to the lack of association of actual risk factors such as infrequent condom use and a history of an STD.
Although in this survey, for the most part, adolescents appear to be relatively well informed about HIV, several prominent misconceptions may indicate that adolescents have only a superficial understanding of HIV. Although the majority of the respondents understand that this virus is not transmitted through casual contact such as airborne spread or use of eating utensils, 59% are not aware that one cannot contract HIV by giving blood. This suggests that adolescents have made the connection between HIV and blood products without conceptually understanding the true mechanism of infectivity.
The distinctions between HIV-positivity and having AIDS may also be lost on adolescents. Almost one third of the respondents understood HIV-positivity to be equivalent to having AIDS. The majority were unaware that there were drugs available to delay the onset of symptoms. This may reflect a knowledge gap, or it may indicate a pervasive fatalistic belief in the inevitability of the disease progression toward death. This belief may help to explain why those with more of some risk factors (no condom use, previous STD) tested no more frequently than those with no risk factors, despite a high rate of concern that they may be HIV-positive. The Health Belief Model posits that where the risk becomes too high, individuals will escape the situation by avoiding action on the issue.26 In this construct, HIV-positivity being associated with no cure, there may be no compelling reason for the present-oriented adolescent to undergo testing.
The most striking misconceptions involved the testing process itself. The widespread belief of the respondents that test results would not be kept confidential, especially with regards to their sexual partners, may be a powerful deterrent to testing. In an adolescent whose world revolves around peer gossip, the fear of disclosure may be overwhelming. Cheng et al27 found that a majority of high school students had concerns they wished to keep confidential, and 25% reported that they would forgo health care in some situations if their parents might find out.
Confidential versus anonymous testing may not be a clear concept to adolescents as well. Studies of adult testing behavior indicate HIV testing trends vary with state/local confidentiality guidelines. As the number of anonymous sites increases, the level of testing increased as well. When the local policy supported reporting of seropositive cases, the level of testing decreased significantly.28 The relatively high number of adolescents who stated that they were tested at a blood bank may be the result of subjecting oneself to an HIV test in an indirect manner. This may possibly be perceived as a means of testing with greater associated confidentiality. Another concern is the fact that 30% of sexually active adolescents surveyed did not think that the HIV test was very accurate. This misperception may be a barrier to voluntary testing.
Consistent with the health belief model, our results suggest that beliefs about susceptibility (having had more than one sex partner within the past year), severity (believing that a positive test result means one has AIDS as opposed to carrying the virus), and lack of effective prevention measures (believing that condoms are only somewhat effective at preventing the spread of AIDS) are associated with sexually active adolescents seeking HIV testing for personal reasons. It remains to be determined whether efforts to increase cues to action, such as discussing AIDS with a doctor, can effectively modify behavior and increase voluntary testing.
Based on the survey methods used, the results of this survey should be generally applicable to nonhomeless adolescents in the United States. This study does not address whether participants were offered testing as part of their clinician visit or whether they had refused testing and for what reasons. This survey did not address the role of anonymous, confidential, or mandatory testing for adolescents as a means of promoting HIV detection and care and public health.29 We also did not assess the potential negative consequences of undergoing HIV testing for certain adolescent populations. Possible consequences such as increased drug or alcohol use in a population already abusing substances necessitate careful and comprehensive pretest and posttest counseling and follow-up when undergoing HIV antibody testing.30 Other consequences of HIV testing and guidelines for testing have been reviewed elsewhere.31
In summary, in Massachusetts among sexually active adolescents, voluntary HIV testing is uncommon. Issues requiring clearer communication to patients include the testing process and its availability, accuracy, and confidentiality. Underscoring the patient's personal risk such as the number of sexual partners may be useful but, clearly, even those who perceive themselves with present risk of infection avoid testing. Informing adolescents of advances in the treatment of this chronic disease and diminishing the fatalistic associations with HIV infection may increase the percentage of at-risk adolescents who seek HIV testing. School health programs should use evaluations of the effectiveness of HIV education programs. Finally, physicians can play an influential role in the promotion of HIV testing by discussing HIV risk behaviors with patients and offering those at risk voluntary HIV counseling and testing.
This work was supported by the Robert Wood Johnson Foundation.
- Received May 30, 1996.
- Accepted December 11, 1996.
Reprint requests to (J.H.S.) Section of General Internal Medicine, Research Unit, Boston Medical Center, 91 E Concord St, Suite 200, Boston, MA 02118.
Dr Samet is a Robert Wood Johnson Foundation Generalist Physician Faculty Scholar.
- AIDS =
- acquired immunodeficiency syndrome •
- HIV =
- human immunodeficiency virus •
- AZT =
- zidovudine •
- STD =
- sexually transmitted disease
- ↵Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report. Atlanta, GA: Centers for Disease Control and Prevention; 1995.
- Rosenberg PS
- D'Angelo LJ,
- Getson PR,
- Luban NLC,
- Gayle HD
- Hingson R,
- Strunin L,
- Berlin B
- Futterman D,
- Hein K,
- Reuben N,
- Dell R,
- Shaffer N
- Waksberg J
- ↵Kish L. Survey Sampling. New York, NY: Wiley; 1965
- ↵Glanz K, Lewis FM, Rimer BK. Health Behavior and Health Education, Theory Research and Practice. San Francisco, CA: Jossey-Bass; 1990
- Goodman E,
- Tipton AC,
- Hecht L,
- Chesney MA
- Igra V,
- Millstein SG
- Rawitscher LA,
- Saitz R,
- Friedman LS
- ↵Becker MH. The Health Belief Model and Personal Health Behavior. Thorofare, NJ: Slack; 1994
- Copyright © 1997 American Academy of Pediatrics