In 2006, the Centers for Disease Control and Prevention (CDC) recommended expanded HIV screening in US emergency departments (EDs).1 Successful execution of the CDC recommendation requires 2 important components, minimal testing barriers (eg, low personnel and financial costs), and high test acceptance rates. The expansion of rapid HIV testing to clinical and nonclinic settings plays an important role in the implementation of the CDC guidelines by increasing test acceptance, facilitating receipt of test results, and promoting linkage to care.2–6 Yet, the uptake of rapid tests may vary based on the testing modality offered—fingerstick versus oral swab.
Early studies demonstrate high rates of oral HIV test acceptability among patients and providers due to the noninvasiveness of the test and speed of specimen collection.7–9 These favorable test attributes have been used to maximize rates of HIV testing.7,10 However, this increase in rapid oral HIV test acceptance has been also accompanied by several reported clusters of false-positive test results.11–13 A recent meta-analysis by Pai et al14 examining the accuracy of OraQuick rapid HIV-antibody-based point-of-care tests found that oral testing had a lower sensitivity than fingerstick testing (98.0% vs. 99.7%) and a lower positive predictive value in low-prevalence settings (88.6% vs. 97.7%). Such reports have raised public concern among health care providers and consumers, and numerous testing clinics have replaced oral fluid testing for fingerstick due to such occurrences.
Rapid fingerstick HIV testing has limitations as well. Patients consistently demonstrate a preference for noninvasive painless oral testing methods,7,9 which may compromise fingerstick test acceptance. Given the distinct limitations of the 2 testing methods—slightly poorer performance of oral fluid versus potentially lower test acceptance of fingerstick—it is not clear whether the frequency of rapid HIV test acceptance would differ between oral fluid and fingerstick tests. To address this question, we conducted a randomized controlled trial of routine rapid HIV screening in an urban hospital ED to directly evaluate the frequency of HIV test acceptance and test completion using oral fluid and fingerstick testing modalities.
The study was approved by the Partners Human Research Committee (2006P-000136) and was overseen by a Data Safety and Monitoring Board.
The Universal Screening for HIV Infection in the Emergency Room (USHER) Phase II trial was conducted in the ED at Brigham and Women's Hospital (BWH), a tertiary academic medical center in Boston, MA. The BWH ED is a Level 1 trauma center that treats more than 56,000 patients annually and serves a demographically diverse patient population of whom 48% are white, 25% black, and 20% Hispanic. In this ED, approximately 60% of presenting patients are women, and the median age is 44 years. Before the implementation of the USHER study, HIV testing was not performed in the BWH ED.
The USHER trial is an NIH-funded, single-center, randomized controlled trial. Details of the original USHER trial have been published elsewhere.15 Between May 5, 2009, and January 4, 2010, USHER Phase II consented eligible patients for the opportunity to be offered routine opt-in, rapid HIV screening. Enrolled subjects were randomized to fingerstick whole blood or oral fluid specimen collection. Details regarding the informed consent process have been previously reported.15 Per Massachusetts law, all subjects provided separate written informed consent for rapid HIV testing in addition to providing written informed consent for trial participation. Participants were also asked to complete a questionnaire, which gathered data on age, race, ethnicity, income, and high-risk behavior. Patients who were interested in HIV testing but refused trial participation were provided with a hard copy list—kept in the ED—of all locally available and Department of Public Health affiliated free HIV counseling, testing, and referral sites.
Patient eligibility was assessed using the ED charts and the BWH computerized patient tracking system. Eligible patients met the following criteria: (1) 18–74 years old; (2) fluent in English or Spanish; (3) not engaged in prenatal care; (4) not self-reportedly known to be HIV infected; (5) not enrolled in the USHER trial in the previous 3 months; and (6) had an Emergency Severity Index (ESI) score of 3–5 (indicating lower clinical severity)16–18 or an ESI score of 1 or 2 (potentially higher clinical severity), with signed approval from the ED attending physician indicating participant's clinical stability and clear mental status. During trial enrollment hours, HIV counselors (trained research assistants) assessed ED patients who had been registered, triaged, and escorted to their rooms to determine if these patients were eligible for USHER Phase II. Enrollment times spanned from 8 AM to 12 AM. and encompassed a minimum of 60 hours per week, including weekends. No financial incentives were provided to patients for trial participation.
After providing informed consent for trial participation, subjects were randomized to 1 of the 2 test modality arms as follows: (1) fingerstick whole blood HIV testing, or (2) oral fluid HIV testing. The fingerstick consent form stated the manufacturer-reported accuracy of the test and discussed the required blood collection methods,19 whereas the consent form for the rapid oral test stated the risk of false-positive results associated with this test as identified in USHER Phase I (e.g. “3 of 4 patients with a ‘reactive’ test do not have HIV infection”).13 Because HIV test acceptance has been found to vary with sex and age groups,20 USHER Phase II participants were randomized using computer-generated block randomization within 4 strata (ie, men <40 years old; men ≥40 years old; women <40 years old; and women ≥40 years old). Neither subjects nor counselors were blinded to the assigned arms.
Counselors were trained by the Massachusetts Department of Public Health21 in both methods of specimen collection for the OraQuickADVANCE Rapid HIV 1/2 Antibody Test (OraSure Technologies, Inc. Bethlehem, PA).
Primary and Secondary Outcome Measures and Statistical Methods
The primary outcome measure was HIV test acceptance rate defined by the proportion of participants whom accepted HIV testing among those randomized within each trial arm (fingerstick or oral fluid). The secondary outcome measure was frequency of HIV test completion as defined by the proportion of participants who completed HIV testing among those randomized within each trial arm.
Data from the original cohort of the USHER trial (Phase I) were used to inform the sample size estimation for this study.15 Sample size was chosen to detect a 10% difference in acceptance rates between 2 testing modality arms (90% power, 0.05 level of significance) and was estimated at 992 subjects tested, 496 per arm.
We used the intention-to-treat principle in which data were analyzed according to the arm to which they were randomly allocated, irrespective of whether they actually received the assigned test. To illustrate the balance between arms achieved by randomization, we present baseline demographic information stratified by study arm. Means and standard deviations are provided for continuous variables (age), whereas frequencies are presented for categorical variables (gender, race/ethnicity, primary language, and education). The difference in the acceptance rates was estimated along with 95% confidence intervals and tested using the χ2 test. All analyses were performed using SAS statistical software Version 9.2 (Cary, NC).
From May 5, 2009, through January 4, 2010, 5612 patients were screened for USHER Phase II trial eligibility, and 2012 (36%) were eligible for enrollment. The most frequently documented reason for ineligibility was age (n = 1767; 49% of all ineligible). Among the 2012 eligible patients approached, 1651 (82%) agreed to study participation (Fig. 1). The 361 eligible patients who refused USHER Phase II trial enrollment were older than study participants (40 vs. 34 years of age; P < 0.0001) yet similar in other demographic features and ESI scores.
Among those 1651 patients who agreed to enrollment, 830 were randomized to the fingerstick arm and 821 to the oral fluid arm. Trial arms were balanced in their demographic distribution (Table 1); mean age was 33 years (SD: 13), 65% were female, 24% were white, 24% African American, and 38% were Hispanic.
Test Acceptance Rates
Among subjects randomized to rapid HIV testing, the test acceptance did not differ meaningfully between arms, 67% (553 of 830) in the fingerstick arm compared with 69% (565 of 821) in the oral fluid arm (P = 0.34). Frequencies of test acceptance did not differ by race, gender, or education. The proportion of HIV tests completed—the proportion of subjects who were tested among those who were randomized—was 66% (549 of 830) in the fingerstick arm and 69% (563 of 821) in the oral fluid arm (P = 0.29) (Table 2). More than 99% of those who accepted an HIV test received the test in both arms.
Among the 1111 study participants who had a valid rapid HIV test result, 5 tests were reactive. Three of these subjects consented to confirmatory testing. Two new cases of HIV infection were identified (1 fingerstick and 1 oral)—a yield of new case identification of 0.2% (95% CI: 0.0% to 0.6%). One fingerstick test was a false positive. No harm was reported in this trial.
In a randomized controlled trial of routine ED-based HIV screening, we found no meaningful difference in the acceptability of an HIV test when comparing fingerstick versus oral fluid rapid HIV test collection modalities.
Our randomized trial corroborates the 2009 study result of White et al.22 White et al22 allocated fingerstick versus oral HIV testing based upon the day of the week and showed that testing modality had minimal effect on testing rates. Reasons for declining screening in that study were generally similar for both screening modalities and seldom related to testing method—the most common being “having recently been tested for HIV” (50%) and “lack of perceived HIV risk” (31%).22
This study is subject to several limitations. Because the USHER Phase II trial was a single-site study with findings that are applicable to rapid HIV screening using fingerstick or oral collection modality, some of our results may not be generalizable to other settings or test kits. Failure to enroll overnight, in addition to the lengthy consent process required to conduct an Institutional Review Board–approved randomized trial in Massachusetts (1 for trial, 1 for testing per Massachusetts state law, and 1 for confirmation of reactive results, if necessary) may have led to selection bias. We did not collect preference data and were therefore not able to fully characterize trade-offs considered in the decision for oral testing given its ease of administration versus the decision for fingerstick testing given its reported superior test characteristics. Furthermore, the frequency of test offers and test acceptance may be lower in EDs that do not utilize ancillary testing personal, as we used in our trial. Finally, several important factors may also influence acceptance of HIV testing and were not measured in our study. These include system-level factors such as location convenience, confidentiality, consent processes, cost, counseling opportunities, and results disclosure.2,10
In the development of HIV screening protocols in the ED, urgent care, and primary care settings, it is critical to tailor optimal screening approaches to enhance test acceptability by decreasing testing barriers, particularly among those less approachable for the test offer and/or those less willing to accept testing. Many EDs have recently started to favor HIV screening using specimens collected for clinical purposes.23 However, this streamlined testing process fails to account for patients who refuse or do not require phlebotomy or who do not present to an ED setting, making our findings still very relevant.
In an era when public health efforts are emphasizing prevention, routine HIV screening, and early entry to care, the use of testing methods that can expand acceptability of HIV testing is essential. Our study is among the first to demonstrate the application of 2 rapid HIV test modalities in a randomized trial where the frequencies of test offer and acceptability are compared within the context of routine voluntary counseling and screening in an ED. We find that test modality was not an important factor in test offer and acceptance rates among patients. Given the option, preference should therefore be given to fingerstick rapid HIV testing because of its slightly superior test characteristics compared with that of the oral rapid HIV test. In low prevalence settings, rapid oral testing should remain an operational solution for a subset of patients who refuse whole blood HIV testing. In addition to test characteristics, however, we believe that system factors such as ease of staff use, necessary Clinical Laboratory Improvement Amendments waivers, laboratory capacity, and HIV prevalence in the testing setting should also be heavily considered.
The authors would like to acknowledge the co-investigators and staff of the USHER Phase II Trial including: Carrie Braverman; Kenneth A. Freedberg, MD, MSc; Susan Larrabee; A. David Paltiel, PhD; Mariesa Ricks; Paul Sax, MD; and Ron Walls, MD. The authors would like to thank the faculty and staff of the BWH ED for their participation in the USHER Phase II Trial and for their dedication to the identification of undiagnosed HIV infection. Dr. Walensky affirms that she has listed everyone who contributed significantly to the work in this Acknowledgment section.
1. Branson BM, Handsfield HH, Lampe MA, et al.. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep. 2006;55:1–17; quiz CE11–14.
2. Peralta L, Deeds BG, Hipszer S, et al.. Barriers and facilitators to adolescent HIV testing. AIDS Patient Care STDS. 2007;21:400–408.
3. Greenwald JL, Burstein GR, Pincus J, et al.. A rapid review of rapid HIV antibody tests. Curr Infect Dis Rep. 2006;8:125–131.
4. San Antonio-Gaddy M, Richardson-Moore A, Burstein GR, et al.. Rapid HIV antibody testing in the New York State Anonymous HIV Counseling and Testing Program: experience from the field. J Acquir Immune Defic Syndr. 2006;43:446–450.
5. Spielberg F, Branson BM, Goldbaum GM, et al.. Choosing HIV counseling and testing strategies for outreach settings: a randomized trial. J Acquir Immune Defic Syndr. 2005;38:348–355.
6. Hutchinson AB, Corbie-Smith G, Thomas SB, et al.. Understanding the patient's perspective on rapid and routine HIV testing in an inner-city urgent care center. AIDS Educ Prev. 2004;16:101–114.
7. Peralta L, Constantine N, Griffin Deeds B, et al.. Evaluation of youth preferences for rapid and innovative human immunodeficiency virus antibody tests. Arch Pediatr Adolesc Med. 2001;155:838–843.
8. Liang TS, Erbelding E, Jacob CA, et al.. Rapid HIV testing of clients of a mobile STD/HIV clinic. AIDS Patient Care STDS. 2005;19:253–257.
9. Kowalczyk Mullins TL, Braverman PK, Dorn LD, et al.. Adolescent preferences for human immunodeficiency virus testing methods and impact of rapid tests on receipt of results. J Adolesc Health. 2010;46:162–168.
10. Spielberg F, Kurth A, Gorbach PM, et al.. Moving from apprehension to action: HIV counseling and testing preferences in three at-risk populations. AIDS Educ Prev. 2001;13:524–540.
11. Delaney KP, Branson BM, Uniyal A, et al.. Performance of an oral fluid rapid HIV-1/2 test: experience from four CDC studies. AIDS. 2006;20:1655–1660.
12. False-positive oral fluid rapid HIV tests–New York City, 2005–2008. MMWR Morb Mortal Wkly Rep. 2008;57:660–665.
13. Walensky RP, Arbelaez C, Reichmann WM, et al.. Revising expectations from rapid HIV tests in the emergency department. Ann Intern Med. 2008;149:153–160.
14. Pai NP, Balram B, Shivkumar S, et al.. Head-to-head comparison of accuracy of a rapid point-of-care HIV test with oral versus whole-blood specimens: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12:373–380.
15. Walensky RP, Reichmann WM, Arbelaez C, et al.. Counselor- versus provider-based HIV screening in the emergency department: results from the Universal Screening for HIV Infection in the Emergency Room (USHER) randomized controlled trial. Ann Emerg Med. 2011;58:S126–S132. e124.
16. Wuerz RC, Travers D, Gilboy N, et al.. Implementation and refinement of the emergency severity index. Acad Emerg Med. 2001;8:170–176.
17. Wuerz RC, Milne LW, Eitel DR, et al.. Reliability and validity of a new five-level triage instrument. Acad Emerg Med. 2000;7:236–242.
18. Eitel DR, Travers DA, Rosenau AM, et al.. The emergency severity index triage algorithm version 2 is reliable and valid. Acad Emerg Med. 2003;10:1070–1080.
20. Liddicoat RV, Losina E, Kang M, et al.. Refusing HIV testing in an urgent care setting: results from the "Think HIV" program. AIDS Patient Care STDS. 2006;20:84–92.
21. JRI Health HCoSM, Inc, AIDS Housing Corporation, AdCare Educational Institute. HIV/AIDS provider training calendar September–December 2008.
22. White DA, Scribner AN, Huang JV. A comparison of patient acceptance of fingerstick whole blood and oral fluid rapid HIV screening in an emergency department. J Acquir Immune Defic Syndr. 2009;52:75–78.
23. Hoxhaj S, Davila JA, Modi P, et al.. Using nonrapid HIV technology for routine, opt-out HIV screening in a high-volume urban emergency department. Ann Emerg Med. 2011;58:S79–S84.