Evaluation of Youth Preferences for Rapid and Innovative Human Immunodeficiency Virus Antibody Tests FREE

ALTHOUGH the Centers for Disease Control and Prevention (CDC) estimates that 50% of those newly infected with human immunodeficiency virus (HIV) in the United States are younger than 25 years,¹ youth remain underdiagnosed for HIV infection and are reluctant to seek HIV counseling and testing services.^{2 - 4} These services may ensure that at-risk adolescents enter care early and receive the maximum benefits from antiretroviral therapy, early opportunistic infection prophylaxis, and perinatal transmission prevention interventions.^{2 - 4} Improving early identification of HIV-infected youth remains a high priority for the public health system.⁵

The CDC recommends a client-centered approach to HIV counseling and testing services in which HIV antibody testing is given during an initial pretest counseling session followed by the delivery of results in a posttest counseling session.⁶ This process takes approximately 2 weeks and requires the client to schedule and make 2 health care appointments. Adolescents aged 13 to 19 years accepted receiving HIV testing at a much lower rate than adult counterparts (fewer than half) with a posttest counseling return rate of 60%.² A report on metropolitan adolescents has indicated a much lower posttest return rate (34%).⁷ To reduce this problem of low acceptance of receiving HIV testing and posttest counseling return rates, existing research has focused on predictors, motivators, and deterrents of HIV counseling and testing for adolescents.^{3 - 4 ,8 - 11}

To date, there has been no investigation to our knowledge of adolescents' preferences for HIV antibody testing techniques, particularly rapid testing using alternative fluids. Currently, only 2 HIV screening strategies are available in the United States: enzyme-linked immunosorbent assays (ELISAs) and 1 rapid test, the Single-Use Diagnostic System (Abbott Laboratories, North Chicago, Ill). The ELISA can be performed with serum, saliva, or urine samples, but in each, at least 3 hours are required to confirm the results. The rapid test can only be performed using serum or plasma but offers results within 15 minutes.¹² However, numerous novel and rapid assays are available internationally, some of which can be performed using oral fluid or fingerstick blood^{13 - 14} (Table 1). In this study, we investigated the acceptability to adolescents of a variety of different collection methods and innovative testing devices that varied in invasiveness and result response time. This study evaluated youth preferences for Food and Drug Administration (FDA)-approved and investigational HIV antibody testing methods and whether preferences would change after subjects learned of the result response times of each test.

Subjects aged 12 to 24 years were recruited for health education sessions in 24 different locations throughout the state of Maryland, although most of the participants were from Baltimore. All adolescents received pretest counseling, a standardized health education intervention, and posttest counseling, and they signed a consent form before participation. The study was approved by the Institutional Review Board of the University of Maryland School of Medicine, Baltimore. Most adolescents participated in health education groups as part of the HIV education activities of the state-approved HIV counseling and testing ambulatory unit. These health education groups ranged in size from 5 to 10 adolescents per counselor and were conducted in community- and clinic-based settings. The group sessions were held from January to July of 1999 in a variety of community settings, including public schools, youth recreation centers, and sports programs. The study purpose, details of the informed consent, and confidentiality information were explained to all subjects prior to voluntary participation in the health education sessions.

During each group or individual health education session, a standardized health education exercise was conducted by counselors with training in health education who offered to try new HIV testing methods. A table was shown to each participant that contained realistic printed drawings of 6 different HIV testing devices, and above the drawings were displayed the actual respective devices. Table 1 lists the 6 testing systems that were offered to the participants. Three of these use oral fluid (1 rapid), 1 uses urine, and 2 rapid tests use whole blood obtained by a fingerstick method. All oral fluid test methods use mucosal transudate, which is collected from cheek and gum tissue that contains high concentrations of serum-derived immunoglobulin G to detect the presence of HIV type 1 (HIV-1). Throughout this article, the term "saliva" will be used because it is the common usage for HIV tests that use oral fluid.

The following is a brief description of the 6 tests used. (1) The SalivaStrip (Saliva Diagnostics System [SDS], Vancouver, Wash) is a rapid chromatographic strip test that uses a saliva sample collected by the Sampler (SDS) saliva collection device (a pad on a stick). Although this test can be performed within minutes, it requires that the sample be centrifuged prior to testing (therefore, it is not considered to be a rapid test in our study). (2) The SalivaCard (Trinity Biotech, Dublin, Ireland) is a rapid flow-through device that uses saliva collected via Orapette (Trinity Biotech) (a rayon ball). The test can be performed in less than 15 minutes and produces a blue dot as a reactive result. (3) Oral Fluid Vironostika HIV-1 ELISA (Organon Teknika, Durham, NC) is designed specifically for use with the OraSure (Epitope, Beaverton, Ore) (a pad on a stick) saliva collection device. Although the sample is collected rapidly, the test is a typical ELISA method requiring approximately 3 hours for results; however, it is FDA approved. (4) The Sentinel HIV-1 Urine EIA (Calypte BioMedical, Berkley, Calif) is an FDA-approved ELISA designed specifically for urine samples. (5a) The Uni-Gold HIV (Trinity Biotech) is a rapid chromatographic strip test that can be performed in minutes and uses whole blood collection directly from a fingerstick; and (5b) the HemaStrip HIV (SDS) is a fingerstick whole blood testing device. It is a chromatographic strip test in that the fingerstick blood diffuses vertically into the device and then reacts with buffer and immobilized antigens within minutes. For the purpose of this study, both fingerstick methods were included in 1 preference option since the procedure for collection of the whole blood was the same (ie, drop of blood by lancet). However, the placement of the blood into the test varies. The HemaStrip HIV device is applied to the drop of blood on the fingertip, while a blood drop from the finger is applied by plastic pipette to the Uni-Gold HIV. All collections and testing were performed as recommended by the manufacturers. Although the testing devices for this study were supplied by the different manufacturers, this study was not funded by the manufacturers.

The health educators demonstrated how each collection and testing device is used and asked participants which ones they would like to try. Participants were offered testing by the 6 different tests and were assisted by the health educators and counselors when necessary. Participants were not required to be tested by all methods if they declined. They were also informed that many of these tests were investigational and nonapproved by the FDA. Therefore, they would not receive the test results. If the participant wanted to obtain the HIV test result with posttest counseling, the FDA-licensed OraSure HIV test was performed.

After the participants were tested by the collection and testing methods, they were asked to rank each according to how much they preferred each one relative to the other methods. Participants were instructed to write in the table, next to each drawing of the devices, a number from 1 to 5 (1, most preferred; 5, least preferred). They were instructed to give each method a different ranking value. However, if a participant chose not to use any particular testing device, then they were instructed to omit the ranking for that method.

Following testing, the health educators collected all of the testing devices and proceeded to show each participant another graphic containing drawings of the test devices, except that it also listed the time required to get results from each method. The health educators explained how the results are derived for each procedure and pointed out that the SalivaCard with Orapette collection method and the fingerstick method using either the Uni-Gold HIV or HemaStrip devices produce results in approximately 10 minutes compared with the other 3 test methods, which require considerably longer times (a few days to 2 weeks). After participants asked questions, they again ranked the test methods they had used in order of preference as during the first set of rankings.

Participants were also requested to complete a brief self-administered questionnaire as part of their health education session that included information on age, sex, ethnic group status, and whether they had ever been tested for HIV prior to this study. The ethnic group categories included African American (n = 237), white (n = 27), Hispanic (n = 7), Native American (n = 4), and other (n = 2). Results are presented with a collapsed variable that compares minority subjects with white subjects. The questionnaire also contained a checklist for risk exposure to HIV that listed having sex with a man, with a woman, with a person with HIV or acquired immunodeficiency syndrome, using drugs, having sex for drug money, or having been a victim of sexual assault. A dichotomous variable was developed so that responses to any of these items were coded in 1 category compared with a category for no risk factors indicated.

Analyses were conducted with independent sample t tests and χ² analysis for sex comparisons. We conducted the same set of paired-sample t tests comparing mean rankings for each test method before and after the educational intervention when controlling for the demographic and other characteristics listed in Table 2. For the reanalyses of the paired-sample t tests by test methods, we divided the sample by sex, ethnic group, age (12-14 years vs 15-24 years), setting (community vs clinic-based), whether the participants were ever tested for HIV, and whether they ever experienced HIV risk behaviors. An illustration of 95% confidence intervals for these comparisons was constructed to demonstrate graphically the relative shifts and positions of the preference rankings before and after the intervention.

The paired-sample t tests showed high levels of power for all the test method comparisons, including the urine and fingerstick HIV testing methods. We conducted a separate power analysis for each test. Effect sizes ranged from a low of f = 0.36 for the OraSure rank difference, to a high of f = 0.65 for the SalivaCard rank difference. The power for the SalivaStrip, SalivaCard, and fingerstick preference-change tests were nearly 100%, the OraSure test power was 91%, and the Sentinel HIV-1 Urine EIA test power was 82%.

Multiple analysis of variance and ordinary least squares models using the 5 rating differences as the dependent variables with the demographics, HIV status tested previously, and experienced HIV risk behaviors present or not present were conducted. No analyses results were significant; therefore, no specified individual characteristic could predict a change in preference for the 5 testing methods.

Of the 278 total participants, approximately half were female and half were male, with an average age of 15 years (Table 2). Most subjects were minorities, predominantly African American, and participated in this study at a community setting. None of these characteristics was significantly different by sex. Slightly more than one quarter of the sample reported having ever been tested for HIV (also not statistically significant by sex) and this was similar to findings from a self-report study of subjects aged 15 to 17 years.¹⁵ Males were more likely than females to report experience with any of the 6 HIV risk behaviors (P<.001), primarily owing to relatively more males reporting having sex with females than females reporting having sex with males. Participants attended health education sessions in community (69.4%) and clinical (25.2%) settings.

Results from the paired-sample t tests given in Table 3 clearly indicate that the second set of preference rankings (combination of collection method and result response time) shifted significantly from the first set of preference rankings (collection method only). After participants learned of the time to receive results, the preferences for the 2 rapid result test methods improved as choices for HIV test method, whereas the other 3 test methods were ranked lower in preference than they were during the first set of rankings. Table 4 indicates that the shifts toward more favorable preference rankings for the SalivaCard and fingerstick methods were substantial, resulting from the fact that it did not require days to weeks to receive test results.

Figure 1 shows sets of confidence rankings for each HIV test method before and after the educational intervention introducing test response time. When confidence intervals do not overlap, they are significantly different from one another. The most highly preferred HIV test method was the SalivaCard (with Orapette) when participants ranked the procedures the second time, displacing the SalivaStrip (with Sampler) as the most preferred test. However, all 3 of the saliva test methods were preferred as much as or more than the urine and fingerstick test methods in both ranking conditions. Even though preference for the fingerstick method improved significantly after subjects learned of the rapid result response time, the less invasive yet slightly longer response time of the SalivaStrip method remained more highly preferred in general. Since the urine (47.8%) and fingerstick (40.3%) HIV testing methods were used relatively less frequently, there are correspondingly fewer rankings for these procedures.

Paired-sample t tests comparing mean rankings for each test method before and after the educational intervention when controlling for the demographic and other characteristics given in Table 2 were significant (P<.001). For the reanalyses, the paired-sample t tests were analyzed by test methods when the sample was divided by sex, ethnic group, age (12-14 years vs 15-24 years), whether or not the participants were ever tested for HIV, whether they ever experienced HIV risk behaviors, and whether the health education setting was conducted in clinical or community settings. The only exception was that white subjects did not significantly change their preference ranking for the urine test. Essentially, the improvement in preference for 2 rapid result response test methods did not seem to be a function of major demographic characteristics, setting, HIV test history, or HIV risk behavior experience.

This study clearly provides support for the use of noninvasive and rapid HIV antibody testing methods for adolescents. Adolescents have shown distinct preferences for innovative HIV antibody testing technologies (rapid saliva and rapid fingerstick testing) regardless of age, sex, ethnic group status, setting (community vs clinic-based), whether they have been previously tested for HIV, or self-reported HIV risk experience. Despite the belief that adolescents would not undergo fingerstick testing owing to fear of pain, 60% of teenagers accepted the fingerstick testing method with only a few describing the experience as painful. The fact that HIV antibody test results could be received in 10 minutes was a powerful indicator in determining adolescent preferences (Figure 1). Preference ranking of the rapid-result testing methods (1 saliva, 2 fingerstick) significantly improved after adolescents learned about the result response time (10 minutes), whereas the other 3 test methods (OraSure, SalivaStrip, and Sentinel HIV-1 Urine EIA) were ranked lower in preference (Table 3). Furthermore, 43.4% of youth rated the SalivaCard and 35.6% of youth rated the fingerstick method more favorably after they learned about the rapid result times of these test methods (Table 4).

Although the preference for the fingerstick methods greatly improved after the adolescents knew the availability of rapid test results, adolescents generally preferred oral testing methods, both before and after the result response time education intervention (Figure 1). We have shown that the most preferred HIV antibody testing method for subjects age 12 to 24 years was a rapid test (SalivaCard) that uses oral fluid as a testing medium (Table 3 and Figure 1). When given the option, adolescents would clearly prefer both a noninvasive and an HIV antibody test with a rapid result response time.

As with all screening tests for HIV, repeatedly reactive results must be confirmed using a more specific supplemental assay such as the Western blot or indirect fluorescence assay. As with venipuncture specimens, saliva and urine samples can be used for screening and confirmation; there are currently FDA-licensed Western blot tests that can be used with saliva and urine matrices. Although confirmation using these specimens does require sending these to a laboratory, the total time for confirmed results is less than with venipuncture specimens, since repeatedly reactive screening test results are obtained in less than 1 hour with rapid tests compared with 1 to 5 days when specimens are sent to a laboratory for initial ELISA testing. Furthermore, for most patients who are tested, results are nonreactive, thereby completing the testing process. For reactive results using a fingerstick specimen, another fingerstick specimen can be obtained on filter paper and sent to a laboratory for testing using a specific ELISA. Although the time for this testing is identical to that of the ELISA and Western blot for venipuncture specimens, the collection is much simpler, does not require a phlebotomist, and is less costly. Finally, our group has reported on the use of a prototype rapid confirmatory HIV assay that has produced an excellent correlation with Western blot results¹⁶ and which may become available in the near future. This would allow for the screening and confirmation of HIV infection in less than one-half hour.

Currently, the Single-Use Diagnostic System is the only rapid HIV antibody test approved by the FDA.¹² Although a negative result can be given in less than 1 hour during a counseling and testing session, an invasive venipuncture is required, and the sample must be centrifuged before testing. Although the test indices of rapid tests were questionable in the late 1980s when they were developed, manufacturers have addressed the issues, and these tests are now considered to have the same sensitivities, specificities, and predicative values as ELISAs.^{12 - 14 ,17 - 18} Even the analytical sensitivity of rapid assays, as assessed by seroconversion panels, to detect early infection has proven to be comparable to other licensed methods.¹⁹ Furthermore, the ability for rapid assays to detect viral variants, such as HIV-1 group O and HIV-2, is excellent.²⁰ Other publications using the newer lateral flow rapid tests have also indicated excellent test indices, including the use of salvia, urine, and whole blood via fingerstick.^{14 ,21 - 23}

In our study, the purpose was not to compare results between the methods but to offer clients a variety of tests that use sample media other than blood collected via venipuncture to determine their preferences. Although each of the 3 saliva collection systems was simple, results indicate that not all were equally desirable by participants. Some participants commented to the health educators that some testing methods were uncomfortable, salty, "felt dry," or had a "funny taste." Regardless, it was obvious that the ability to perform HIV antibody testing on samples other than via venipuncture blood collection (ie, saliva, urine, fingerstick) was desirable to participants. At present, at least 1 of the fingerstick rapid HIV antibody testing devices used in this study is pending approval by the FDA (Table 1). A urine test is currently available in many clinical settings, but it is not rapid, and our adolescents preferred this collection method least of all. This seems to be owing to the need to visit a restroom and carry back the sample to the health education area. Also, some adolescent girls mentioned they felt uncomfortable giving a urine specimen because they were menstruating. Finally, as OraSure HIV antibody testing gains popularity at youth-specific HIV counseling and testing sites because of its simple saliva collection method, in our study, adolescents preferred it the least compared with the other 2 saliva testing devices (Table 3). Two reasons included the salty taste and the longer response time for results. Additionally, the fingerstick rapid tests were also preferred over the FDA-approved OraSure test.

Our interpretation is tempered by several limitations. First, there are fewer rankings for the urine (47.8%) and fingerstick (40.3%) HIV testing methods. Each participant was instructed to omit a ranking value if they did not use the device. A participant may have refused to take the test owing to their comfort level or to time constraints when multiple collection and testing systems were requested. Second, most of our adolescent population is representative of one metropolitan city. Our study may not address whether adolescents in different geographic locations or from a differing socioeconomic status had different HIV antibody testing preferences. Although this may be viewed as a limitation, our sample predominantly includes minority subjects, the population that continues to be disproportionately affected by HIV and AIDS.

As further information becomes available that describes the impact of HIV testing devices and sampling media on the acceptance of HIV counseling and testing services, public health professionals need to advocate the development, approval, and use of innovative HIV antibody testing technologies. Our study is the first to show which sample collection methods are preferred and the importance of rapidly obtained results. The goal of this ongoing research is to increase the number of adolescents accepting HIV counseling and testing services while increasing opportunities for HIV prevention, early identification, and linkage to care. Recent studies by the CDC¹⁷ have shown the importance of rapid HIV testing, such that rapid results can have a major impact on counseling to change behavior, eliminating the need for return visits, and for providing antiretroviral treatment in a clinically relevant time frame.

Accepted for publication March 6, 2001.

We thank Rebecca Saville, MS, Daniel Edelman, MS, and Fassil Ketema, MS, for their assistance in the laboratory and data management. We also thank Mary Beth Raven, Pat Caldwell, Sue Miller, PhD, Celia Ryder, CPNP, and Tay Croxton for assistance in data collection and organization.

Presented in part at the XIII International AIDS Conference, July 9-14, 2000, Durban, South Africa.

Corresponding author and reprints: Ligia Peralta, MD, University of Maryland School of Medicine, Department of Pediatrics, Division of Adolescent Medicine, 655 W Lombard St, Suite 311, Baltimore, MD 21201 (e-mail: lperalta@peds.umaryland.edu).