Original research article
Effects of rapid versus standard HIV voluntary counselling and testing on receipt rate of HIV test results: a meta-analysis
International Journal of STD & AIDS 2015, Vol. 26(3) 196–205 ! The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0956462414533671 std.sagepub.com
Yuan Wang1, Jian Guo2 and Wenli Lu1
Abstract Rapid HIV voluntary counselling and testing (RVCT) is an alternative method of standard HIV voluntary counselling and testing (SVCT). Less is known about whether RVCT improves the receipt rate of HIV test results among clients who seek HIV counselling and testing. We aimed to evaluate effectiveness of RVCT on result receipt rate. We conducted a comprehensive search of databases containing Medline, EBSCO, Web of science, and Cochrane library to identify studies published up to August 2012. Reviewers extracted information independently. Risk of bias was evaluated with Cochrane Collaboration’s tool for assessing study quality. Five randomised controlled trials were included and analysed for the result receipt rate using a random-effects model. The pooled receipt rate of HIV test results in the RVCTwas significantly higher than in the SVCT (RR ¼ 1.74, 95% CI ¼ 1.47–2.07). Our results suggest RVCT as a favourable method to increase the receipt of HIV test results. Only two included studies assessed the modification of risk behaviour after HIV-CT in a different manner; also, the sample size was small in the current meta-analysis. In future research, it is necessary to confirm the effect of RVCT on disinhibition of post-test risk behaviour.
Keywords HIV, AIDS, testing, VCT, standard HIV counselling and testing, randomized HIV counselling and testing, rapid HIV counselling and testing, meta-analysis Date received: 6 December 2013; accepted: 26 March 2014
Introduction An estimated 232,700 HIV-infected people in the USA were unaware of their HIV status at the end of 2006,1 which impeded early diagnosis, treatment and prevention of HIV transmission. To increase the number of people who discover their HIV serostatus, the Centers for Disease Control and Prevention (CDC) recommended HIV counselling and testing (HIV-CT) as a routine strategy in all health care settings for people aged between 13 and 64.2 The standard procedure of HIV-CT (SVCT) includes pretest counselling, venipuncture, screening for HIV with enzyme immunoassay (EIA), and confirming reactive EIA results with Western blot. People once tested need a second visit to receive their results and post-test counselling, usually one or two weeks after testing. Although this strategy effectively avoids delivering false-positive results, some people do not return for their HIV test results after testing. Previous
studies reported that approximately 40% of clients in public-funded clinics and 30% of risk clients in outreach settings did not receive their test results,3,4 impairing the efficacy of HIV-CT and wasting of medical resources. For improving the effectiveness of HIVCT, the 1998 CDC testing guidelines expanded the sphere of rapid HIV counselling and testing (RVCT), which can be completed within 20 min, to various settings by providing HIV preliminary positive test results 1 Department of Health Statistics, School of Public Health, Tianjin Medical University, Tianjin, China 2 Department of AIDS control and prevention, Tianjin Binhai New Area Tanggu Center for Disease Control and Prevention, Tianjin, China
Corresponding author: Wenli Lu, Department of Health Statistics, School of Public Health, Tianjin Medical University, Meteorological Station Road No 22, Heping District, Tianjin, 300070, China. Email: [email protected]
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
197
and eliminating return visits. This enabled the tested people to obtain definitive negative and preliminary positive results in a single visit, which increased the number of clients receiving their test results.5 Awareness of HIV status influences subsequent behaviours. A meta-analysis of HIV-CT showed that risk reduction counselling could be effective among HIV-positive persons but not among HIV-negative individuals.6 The 2006 CDC testing guidelines recommend that prevention counselling should not be a necessary part of HIV screening projects and could only be offered to HIV-positive persons.2 This may increase the risk of disinhibited post-test risk behaviours among persons who receive HIV-negative test results through RVCT. Compared with RVCT, clients in SVCT can interact longer with the counsellor, obtain more information on safe behaviours and have an interval to reflect their risk behaviour before learning their HIV serostatus. A previous meta-analysis reported that RVCT could increase receipt of HIV test results in pregnant women.7 However, the efficacy of RVCT in other medical settings was not confirmed. On the other hand, the risk of disinhibited behaviour after receiving a negative HIV result by RVCT remained uncertain. Therefore, we performed a meta-analysis of RVCT on receipt of HIV test results and post-test risk behaviours among people who accepted HIV-CT in different settings. For obtaining a rigorous estimate of intervention effect, we only included randomised controlled trials (RCTs) in our study.
Methods Eligibility criteria Inclusion criteria: (1) Type of studies: RCTs studying HIV-CT in health care or outreach settings published from 1990 to 2012 in English. (2) Type of participants: participants of any age who underwent HIV-CT were included. (3) Type of intervention: RVCT was defined as rapid HIV testing completed within 20 min, and receipt of HIV test result and post-test counselling completed in one visit. SVCT was defined as HIV testing completed within one or two weeks, and participants needed a second visit to receive the test result and posttest counselling. (4) Type of outcome measures: the primary outcome included receipt rate of HIV test result and incidence of risk behaviours after HIV-CT. Exclusion criteria: (1) Type of study: review (including systematic review), report of conference, etc., and studies with incomplete data of interest and duplicate publications. (2) Type of participants: those women who accepted HIV-CT in the perinatal period or while in labour.
Data source Published literatures on effects of RVCT versus SVCT on Receipt Rate and Post-test Risk Behaviours were independently searched by two reviewers in electronic databases (Medline (Pubmed, 1966-till date), EBSCO (Academic Source Complete, Research StartersEducation, Research Starters - Sociology, ERIC, Library Information Science & Technology Abstracts, 1985-till date), Web of science ((Science Citation Index Expanded (1998-till date), Conference Proceedings Citation Index - Science (2004-till date)), Cochrane library (Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, Cochrane Central Register of Controlled Trials)) up to August 2012, complemented by manual search. Finally, the authors of possible relevant studies were contacted when more information or clarification was needed. The search terms were rapid* and HIV voluntary counselling and testing* and reception*; rapid* and HIV voluntary counselling and testing* and risk behaviours*. The reference lists of relevant reports and recent relative meta-analysis were also reviewed, but no additional study was identified.
Study selection Two review authors (Yuan Wang, Jian Guo), working independently and in parallel, scanned related abstracts and obtained the full text reports of studies when the abstract indicated or suggested the study was on SVCT or RVCT. After obtaining full reports of the candidate study (either in full peer-reviewed publications or press articles), the same review authors independently assessed eligibility of the studies for inclusion in the review. In the case of multiple publications of the same study or overlapping data sets, only data from the largest or most updated results were included. Any initial disagreement was resolved by consensus after additional review of the articles with the third author (Wenli Lu). Missing information was sought and unclear information clarified by contacting the corresponding authors of the eligible reports.
Data extraction The same two review authors extracted data from the eligible studies independently using a standardized form. Information was extracted from eligible articles on: (1) Characteristics of study: first author, year of publication, period of enrollment, location, and setting. (2) Type of intervention: approach of RVCT versus SVCT. (3) Type of outcomes: number of participants allocated, tested, and received HIV test result; risk behaviours (2 sex partners, unprotected sexual
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
198
International Journal of STD & AIDS 26(3)
intercourse). Any disagreement in data collection was adjudicated by the same third reviewer.
Assessment of risk of bias Risk of bias of eligible studies was assessed by the same two review authors independently using the Cochrane Collaboration’s tool.8 Six domains were evaluated: sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective outcome report and other sources of bias. The quality of evidence was assessed with GradePro. It contained study limitations of design or performance, heterogeneity of results, directness of evidence, precision of results and publication bias.
Analytic methods Relative risk (RR) and 95% confidence interval (CI) were used to estimate effect sizes. The pooled RR and 95% CI for overall effects of RVCT versus SVCT were constructed with random-effects model on an intention to treat (ITT) basis, because the variance across studies was conservative. Only two studies assessed risk behaviours after HIV-CT with a different approach; therefore, we emphasized on describing the results within qualitative synthesis rather than meta-analysis. Heterogeneity was evaluated by the Breslow-Day test, which was suggested by Higgins (termed I2).9 We explored the potential source of heterogeneity using subgroup analysis on risk of bias, research setting and source of funding. Because studies of outpatient and inpatient departments were associated with low HIV risk, these studies were classified into the category of hospital setting. Studies performed in high HIV risk settings including bathhouses for men who have sex with men (MSM), needle exchange sites and clinics for sexually transmitted infections (STIs) were classified as non-hospital settings. Forest plot was used for presenting meta-analysis results. Funnel plot was used for assessing publication bias. The symmetry was evaluated visually and formally with Egger’s test, performed with Stata12.0.10 This study was performed according to the PRISMA guidelines for meta-analysis.11 Data analysis was performed with Review manager 5.0.
excluding 465 duplicate publications, 2045 articles were reviewed of which 2024 articles were excluded as they did not meet the inclusion criteria. After reviewing the full text of remaining articles, five articles from four studies were included in this systematic review (Figure 1).
Study characteristics Three studies recruited male and female participants,12–14 and one study recruited only female participants.15 The study settings included public STI clinics, clinic hospital, needle exchange and MSM bathhouse. The duration of research ranged from 4–10 months (Table 1). In the report of Speilberg_2005, there were two studies performed in different settings: MSM bathhouse and needle exchange, so that this report was classified into two studies: Spielberg_2005A (Needle exchange), Spielberg_2005B (MSM bathhouse). 1. Participants: the eligible studies contained 4 412 participants. All of the participants were above 14 years old, were proficient in English and were unaware of HIV status before study entry. 2. Intervention and control: all studies were conducted between 1999 and 2007 in the USA. The following procedures were contained in RVCT arm: singlesession counselling, streamlined counselling and standard counselling; HIV test with Single Use Diagnostic System for HIV–1 test (SUDS), Orasure or Oraquick. In SVCT arm, clients would need a second visit for their HIV test results after their first visit including brief two-session counselling and standard counselling; and HIV test with EIA or Orasure. Although some studies (Metcalf_2005, Spielberg_2005A, Spielberg_2005B) used oral fluid test (Orasure) in SVCT arm, their clients needed a return visit for their HIV test results. 3. Outcomes: the outcomes evaluated were receipt rate of HIV test result and incidence of risk behaviours after HIV-CT. All of the five RCTs compared the result receipt rate. Only two RCTs assessed risk behaviour after HIV-CT, but with a different approach.
Risk of bias within studies Results Study selection A total of four studies containing five RCTs were identified for inclusion in our study. The search from electronic databases provided 2 510 articles. After
Three RCTs were considered as having adequate sequence generation (Metcalf_2005, Spielberg_ 2005(A,B)), indicating a low risk. In the other two RCTs (Wurcel_2005, Anaya_2008), the process of sequence generation was mentioned but not clearly described, and their risks were judged to be high.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
199
Figure 1. Flowchart of included studies.
Only one RCT (Metcalf_2005) used allocation concealment which blinded assignment to laboratory staff and data analysts, but not to participants and study staff. None of the remaining RCTs reported implementing concealment or blindness. Because receipt of HIV test result was unlikely to be influenced by awareness of allocation, the risks of bias about concealment and blindness among all RCTs were judged to be low. Two studies (Metcalf_2005, Spielberg_2005B) reported close proportions of incomplete outcome between RVCT and SVCT, indicating a low risk. Other RCTs had a disproportional loss of those participating in HIV-CT (Spielberg_2005A, Anaya_2008, Wurcel_2005,) which was not discussed and were considered as high risk. All RCTs were free of selective reporting, and the risks were judged to be low. One RCT (Metcalf_2005) was free of other bias because the characteristics of RVCT and SVCT were adequately reported and the study design was appropriate, indicating a low risk. The sample of Wurcel_2005 and Anaya_2008 was relatively small, which perhaps did not adequately power the outcomes. In addition, the RCTs of Spielberg_2005A and Spielberg_2005B did not report the baseline characteristics between the two groups. Thus, other sources of bias in these four RCTs (Spielberg_2005A, Spielberg_2005B, Wurcel_2005, Anaya_2008) were judged to be high. Overall, the studies of Metcalf_2005 and Spielberg_2005B were considered to have
low risk of bias. The risks of bias for Wurcel_2005, Spielberg_2005A, and Anaya_2008 were judged to be high.
Results of meta-analyses and qualitative analyses All of five RCTs contained in the four studies reported that the receipt rate of HIV test results in the RVCT was significantly higher than in the SVCT. The pooled effect also showed that RVCT increased the receipt rate significantly compared with SVCT (N ¼ 4 412, RR ¼ 1.74, 95% CI ¼ 1.47–2.07) (Figure 2) Two RCTs (Metcalf_2005, Anaya_2008) assessed the risk behaviours after HIV-CT. However, the behaviours were measured with a different approach in the two studies; therefore, we emphasized on describing the results within qualitative synthesis rather than metaanalysis. The first RCT, Meltcalf_2005, compared the incidence of risk behaviours three months after HIV-CT (N ¼ 2 495). The rate of follow-up was 76.4% in RVCT and 75.0% in SVCT. There was no significant differences on the behaviours of 2 sex partners (RVCT: 33.7%, SVCT: 30.0%; RR ¼ 1.11, 95% CI ¼ 0.99– 1.25), any unprotected sex (RVCT: 64.2%, SVCT: 62.5%; RR ¼ 1.03, 95% CI ¼ 0.97–1.09), any unprotected sex with non-primary partner (RVCT: 18.7%, SVCT: 15.8%; RR ¼ 1.18, 95% CI ¼ 0.99–1.41), any unprotected sex when drunk or high (RVCT: 23.7%,
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
EIA, Orasure
EIA
SUDS
Oraquick
Oraquick, Orasure Standard counseling
MSM bathhouse
Inpatients, outpatients of hospital Hospital, outpatient clinic
Seattle
Massachusetts
Southern California –
EIA
EIA, Orasure SUDS
Standard counseling; pretest written material Standard counseling; pretest written material Standard counseling Needle exchange
1999–2000 (122 days) 1999–2000 (122 days) 2003.9–2004.6
SVCT: 22.2%; RR ¼ 1.07, 95% CI ¼ 0.92–1.23), sex with new partner on day first met (RVCT: 6.7%, SVCT: 7.4%; RR ¼ 0.90, 95% CI ¼ 0.67–1.20), and one-time sex partner (RVCT: 20.1%, SVCT: 18.7%; RR ¼ 1.07, 95% CI ¼ 0.91–1.26). In addition, the Meltcalf_2005 study assessed the incidence rate of STI and did not find any significant difference between the two groups at 12 months after HIV-CT.12 The second RCT, Anaya_2008, evaluated the post-test reduction of sexual risk behaviours, and 81.0% of clients in RVCT and 73.6% of clients in SVCT completed the four weeks follow-up (N ¼ 191). There was no significant differences on reduction of risk behaviours between the two groups (RVCT: 48.5%, SVCT: 46.3%; RR ¼ 1.05, 95% CI ¼ 0.77–1.43).15
Risk of bias across studies Significant evidence of heterogeneity was found (I2 ¼ 93%, P < 0.001). To explore the heterogeneity, we conducted subgroup analysis according to risk of bias, setting and funding of research. 1. Risk of bias: The low-risk RCTs (Metcalf_2005, Spielberg_2005B) and the high-risk RCTs (Spielberg_2005A, Wurcel_2005, Anaya_2008) were associated with a higher RR (low: N ¼ 3 734, RR ¼ 1.37, 95% CI ¼ 1.33–1.41; high: N ¼ 678, RR ¼ 2.35, 95% CI ¼ 1.43–3.85). The I2 of the low and high group was 0% and 90%, respectively (Figure 2) 2. Setting: the non-hospital setting RCTs of Metcalf_2005, Spielberg_2005A, and Spielberg_ 2005B (N ¼ 4 058, RR ¼ 1.40, 95% CI ¼ 1.31–1.51) had a narrower interval than the hospital setting of Wurcel_2005 and Anaya_2008 (N ¼ 354, RR ¼ 2.85, 95% CI ¼ 1.86–4.37). The I2 of the non-hospital setting and the hospital setting decreased to 68% and 70%, respectively (Figure 3). 3. Funding: The Melcalf_2005 was funded by CDC (N ¼ 3 297, RR ¼ 1.38, 95% CI ¼ 1.33–1.42). The I2 was 97% for the remaining four RCTs funded by other organizations (N ¼ 1 115, RR ¼ 2.03, 95% CI ¼ 1.25–3.29) (Figure 4).
Anaya_2008
Wurcel_2005
Spielberg_2005B
Spielberg_2005A
Meltcalf_2005
Sensitivity analysis
Study ID
Approach of counselling
EIA, Orasure SUDS Brief two-session counseling
Single-session counseling Single-session counseling Single-session counseling Standard counseling Streamlined counseling STI clinics
Denver, Long beach, Newark Seattle 1999.2–2000.12
Setting Location Period of enrollment
Table 1. Summary of RCTs evaluating the receipt rate of HIV test results between RVCT and SVCT.
Approach of testing
SVCT RVCT SVCT
International Journal of STD & AIDS 26(3)
RVCT
200
A fixed-effects model was conducted for testing the robustness of our results. Compared with the random-effects model, the fixed-effects model provided less conservative results and identical patterns of results.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
201
Figure 2. Meta-analysis result of receipt rate stratified by risk of bias.
Figure 3. Meta-analysis result of receipt rate stratified by research setting.
Publication bias To assess the presence of publication bias, a funnel plot was drawn (Figure 5), and the Egger test was used to
investigate funnel plot asymmetry. No evidence of publication bias was observed in the receipt rate of HIV test result studies (Egger test: interception ¼ 0.548, P ¼ 0.514).
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
202
International Journal of STD & AIDS 26(3)
Figure 4. Meta-analysis result of receipt rate stratified by funding source of research.
Figure 5. Funnel plot for assessing publication bias.
Discussion Summary of main results A total of four studies involving five RCTs were eligible for inclusion in the review. The overall effect of the
meta-analysis suggested that the result receipt rate was significantly higher in the RVCT than in the SVCT. On the other hand, the evidence from the two studies comparing the risk behaviours after HIV-CT did not show significant difference between RVCT and SVCT.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
203
Therefore, it was not robust enough to determine the effect of RVCT on post-test risk behaviours.
Implications for practice An effective strategy for HIV prevention is HIV-CT. The main purposes of HIV-CT were to prompt people to know their HIV serostatus, to provide medical care to the infected clients, to provide counselling to reduce clients’ risk behaviour for avoiding infection and to prevent infected clients from spreading HIV.16 The transmission rate of people who were unaware of their positive status was higher than people who were aware of their infections.17 In the USA, approximately 14,000 new infections per year could be attributed to unawareness of HIV infection.18,19 In low-income countries such as sub-Saharan Africa, this proportion ranged from 30% to 70%.20 Failure to return for HIV test results was a possible reason for unawareness of infection. As an alternative strategy of SVCT, RVCT can effectively elevate awareness of HIV serostatus. Because most persons test negative, RVCT allows clients to receive counselling and HIV test results in one visit. This saves time and cost of a return visit and increases the proportion of receiving HIV test result among the tested clients, especially for the less reluctant ones. Rigorously designed RCTs were associated with adequate sequence generation, complete outcome data and comparable baseline, indicating reliable conclusions. This led to low risk of bias and homogeneity of the pooled effect (I2 ¼ 0%), suggesting high quality of evidence. The heterogeneity in studies with high risk of bias (I2 ¼ 90%) may be because of the relative small sample sizes. Therefore, studies with larger sample sizes are needed in future research. On the other hand, heterogeneity may indicate that the pooled effect was more vulnerable to variations in study location, setting and population. Early diagnosis of HIV infection helped clients access timely medical care and reduce risk behaviours, which avoided further transmission.21 In the area with high prevalence but low receipt rate, the counsellor or physician could use RVCT to increase the number of clients who become aware of their HIV serostatus, which reduces the difficulty of locating infected clients. The subgroup analysis of research setting also showed that RVCT increased the receipt in both hospital and non-hospital settings. However, the effect of RVCT on increase of receipt rate was lower in non-hospital settings than in hospital settings, indicating that RVCT may be more efficient in hospital settings. The potential reason could be that compared with patients in hospital, persons in high-risk settings, such as bathhouses, needle exchange sites and STI clinics were more likely
to return for test results because of their risky behaviours and fear of being infected. The result receipt rate of SVCT was clearly lower in hospital settings than non-hospital settings in our studies, reducing the RR between the two settings. On the other hand, the utilization of different modalities of rapid HIV testing (venipuncture and oral fluid) may be another explanation for the significant difference of pooling effects between hospital settings and non-hospital settings. Because oral fluid testing was more acceptable than blood testing, the unbalanced distribution of testing approaches may also affect the effectiveness of RVCT in the two settings. The CDC was responsible for widespread dissemination of HIV counselling and testing. The Meltcalf_2005 study was a multicentre study in STI clinics; funded by CDC it accounted for the high proportion of participants in the meta-analysis. The network of CDC provides support for recruiting large number of participants in three locations including Denver, Long beach and Newark. Although investigators may tend to favour interventions or products of the sponsor, no statistical difference was found between pooling effects of the CDC-sponsored study and nonCDC sponsored study. However, the relative smaller sample size of studies funded by non-CDC organizations might limit the generalizability of the findings. Costs of the HIV-CT procedure, prevalence and receipt rates after testing determine the entire costs of the HIV-CT approach. Compared with SVCT, the process of RVCT was more expensive because of the cost of the test kit and further counselling for the infected clients.22 However, considering the acceptance rate of testing and the receipt rate of results, the cost of informing an infected client was lower with RVCT than with SVCT in STI clinics and emergency departments.22 A costeffectiveness study also recommended that RVCT could be used as a routine HIV-CT approach for adults in areas where the prevalence of undiagnosed HIV infection is higher than 0.2%.23 Therefore, when planning HIV-CT policy, RVCT is an ideal alternative strategy that provides opportunity for people with potential for infection to learn their HIV serostatus. Our study was insufficient to confirm the difference in post-test risk behaviours between RVCT and SVCT. Previous meta-analysis focusing on the effect of HIVCT on sexual risk behaviour revealed that HIV-CT could reduce sexual risk behaviours among the HIVpositive clients but not among the HIV-negative ones.21 Some investigators argued that according to the Information-Motivation-Behaviour Skills model, perception of risk was the precondition of modification of behaviour.24 Anxiety of being infected was a critical motivation for people who seek HIV test.25,26 After learning the negative test result, the anxiety of being
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
204
International Journal of STD & AIDS 26(3)
infected was eliminated for clients in both RVCT and SVCT, which might lead to similar effect on modification of risk behaviours.
Quality of the evidence No evidence of publication bias was observed in the receipt rate of HIV test results; however, it should be noted that the low number of RCTs limits the potential for evaluating substantial publication bias. The quality of evidence presented was discussed on the basis of the GRADE approach. The receipt of test result from the included studies was graded as moderate. However, the risk behaviours after RVCT can not be evaluated based on present studies, indicating a very low grade. The chief objective of HIV-CT is to help people learn whether they are infected or not. Awareness of HIV serostatus determines the client’s future behaviours. Some people in the SVCT fail to return for the test results, but RVCT can solve this problem and increase the number of people becoming aware of their HIV serostatus. At present, although the effect of RVCT on the post-test risk behaviours is uncertain, we evaluate RVCT as a strong contender for increasing receipt of HIV test results.
Limitations There are several limitations in our study. First, as with other systematic reviews, the study population and setting of research was inconsistent across studies. This led to a high level of heterogeneity in our meta-analysis, which could be because of the research setting in the subgroup analysis. Second, the four studies other than the Metcalf_2005 study provided a limited sample size. It should be noted that a meta-analysis of several small studies does not predict the results of a single large study. The inequality may affect extrapolation of study conclusion. In future research, RCTs with larger sample sizes are needed to increase representativeness of different circumstances. Third, because of the limited financial and database resources, our search process did not cover databases such as Embase. The included studies were restricted to reports published in English owing to limitations in translating other languages. Finally, all of the studies included in our meta-analysis were performed in the USA. It limited the application of the result in other countries with different socioeconomic background and HIV prevalence.
Conclusions In our study, we recommend RVCT as a favourable method to increase the receipt of HIV test results.
Only two included studies assessed the modification of risk behaviour after HIV-CT in a different manner; also, the sample size was small in the current meta-analysis. In future research, it is necessary to confirm the effect of RVCT on disinhibition of post-test risk behaviour. Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by The Ministry of Education Humanities and social science research (Project No. 13YJCZH188). It was designed at the school of Public Health in Tianjin Medical University and will carry out at Tianjin Binhai New Area Tanggu Center for Disease Control and Prevention.
References 1. Centers for Disease Control and Prevention (CDC). HIV prevalence estimates—United States, 2006. MMWR Morb Mortal Wkly Rep 2008; 57: 1073–1076. 2. 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. 3. Valdiserri RO, Moore M, Gerber AR, et al. A study of clients returning for counseling after HIV testing: implications for improving rates of return. Public Health Rep 1993; 108: 12–18. 4. Sullivan PS, Lansky A, Drake A, et al. Failure to return for HIV test results among persons at high risk for HIV infection: results from a multistate interview project. J Acquir Immune Defic Syndr 2004; 35: 511–518. 5. Centers for Disease Control and Prevention (CDC). Update: HIV counseling and testing using rapid tests – United States, 1995. MMWR Morb Mortal Wkly Rep 1998; 47: 211–215. 6. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89: 1397–1405. 7. Minnier KS, van der Walt SJ and Klopper HC. A systematic review of counselling for HIV testing of pregnant women. J Clin Nurs 2009; 18: 1827–1841. 8. Higgins JPT and Green S. Cochrane handbook for systematic reviews of interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, www.cochranehandbook.org (2011). 9. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. 10. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
205
11. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med 2009; 151: W65–W94. 12. Metcalf CA, Douglas JM Jr, Malotte CK, et al. Relative efficacy of prevention counseling with rapid and standard HIV testing: a randomized, controlled trial (RESPECT2). Sex Transm Dis 2005; 32: 130–138. 13. 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. 14. Wurcel A, Zaman T, Zhen S, et al. Acceptance of HIV antibody testing among inpatients and outpatients at a public health hospital: a study of rapid versus standard testing. AIDS Patient Care STDS 2005; 19: 499–505. 15. Anaya HD, Hoang T, Golden JF, et al. Improving HIV screening and receipt of results by nurse-initiated streamlined counseling and rapid testing. J Gen Intern Med 2008; 23: 800–807. 16. Centers for Disease Control and Prevention (CDC). Technical guidance on HIV counseling. Center for Disease Control and Prevention. MMWR Recomm Rep 1993; 42: 11–17. 17. Marks G, Crepaz N and Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 2006; 20: 1447–1450. 18. Hall HI, Song R, Rhodes P, et al. Estimation of HIV incidence in the United States. JAMA 2008; 300: 520–529.
19. Campsmith ML, Rhodes PH, Hall HI, et al. Undiagnosed HIV prevalence among adults and adolescents in the United States at the end of 2006. J Acquir Immune Defic Syndr 2010; 53: 619–624. 20. Department WHOHA. Progress report 2011: Global HIV/AIDS response epidemic update and health sector progress towards universal access WHO, UNICEF, UNAIDS. 2011. 21. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89: 1397–1405. 22. Farnham PG, Hutchinson AB, Sansom SL, et al. Comparing the costs of HIV screening strategies and technologies in health-care settings. Public Health Rep 2008; 123: 51–62. 23. Paltiel AD, Walensky RP, Schackman BR, et al. Expanded HIV screening in the United States: effect on clinical outcomes, HIV transmission, and costs. Ann Intern Med 2006; 145: 797–806. 24. Fisher JD and Fisher WA. Changing AIDS-risk behavior. Psychol Bull 1992; 111: 455–474. 25. Takahashi S, Yamazaki Y and Kawata C. Client anxiety from awareness of personal HIV risk and relation to taking HIV test at public health centers in Japan. Nippon Koshu Eisei Zasshi 1999; 46: 275–288. 26. Earl A and Albarracin D. Nature, decay, and spiraling of the effects of fear-inducing arguments and HIV counseling and testing: a meta-analysis of the short- and longterm outcomes of HIV-prevention interventions. Health Psychol 2007; 26: 496–506.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Effects of rapid versus standard HIV voluntary counselling and testing on receipt rate of HIV test results: a meta-analysis
International Journal of STD & AIDS 2015, Vol. 26(3) 196–205 ! The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0956462414533671 std.sagepub.com
Yuan Wang1, Jian Guo2 and Wenli Lu1
Abstract Rapid HIV voluntary counselling and testing (RVCT) is an alternative method of standard HIV voluntary counselling and testing (SVCT). Less is known about whether RVCT improves the receipt rate of HIV test results among clients who seek HIV counselling and testing. We aimed to evaluate effectiveness of RVCT on result receipt rate. We conducted a comprehensive search of databases containing Medline, EBSCO, Web of science, and Cochrane library to identify studies published up to August 2012. Reviewers extracted information independently. Risk of bias was evaluated with Cochrane Collaboration’s tool for assessing study quality. Five randomised controlled trials were included and analysed for the result receipt rate using a random-effects model. The pooled receipt rate of HIV test results in the RVCTwas significantly higher than in the SVCT (RR ¼ 1.74, 95% CI ¼ 1.47–2.07). Our results suggest RVCT as a favourable method to increase the receipt of HIV test results. Only two included studies assessed the modification of risk behaviour after HIV-CT in a different manner; also, the sample size was small in the current meta-analysis. In future research, it is necessary to confirm the effect of RVCT on disinhibition of post-test risk behaviour.
Keywords HIV, AIDS, testing, VCT, standard HIV counselling and testing, randomized HIV counselling and testing, rapid HIV counselling and testing, meta-analysis Date received: 6 December 2013; accepted: 26 March 2014
Introduction An estimated 232,700 HIV-infected people in the USA were unaware of their HIV status at the end of 2006,1 which impeded early diagnosis, treatment and prevention of HIV transmission. To increase the number of people who discover their HIV serostatus, the Centers for Disease Control and Prevention (CDC) recommended HIV counselling and testing (HIV-CT) as a routine strategy in all health care settings for people aged between 13 and 64.2 The standard procedure of HIV-CT (SVCT) includes pretest counselling, venipuncture, screening for HIV with enzyme immunoassay (EIA), and confirming reactive EIA results with Western blot. People once tested need a second visit to receive their results and post-test counselling, usually one or two weeks after testing. Although this strategy effectively avoids delivering false-positive results, some people do not return for their HIV test results after testing. Previous
studies reported that approximately 40% of clients in public-funded clinics and 30% of risk clients in outreach settings did not receive their test results,3,4 impairing the efficacy of HIV-CT and wasting of medical resources. For improving the effectiveness of HIVCT, the 1998 CDC testing guidelines expanded the sphere of rapid HIV counselling and testing (RVCT), which can be completed within 20 min, to various settings by providing HIV preliminary positive test results 1 Department of Health Statistics, School of Public Health, Tianjin Medical University, Tianjin, China 2 Department of AIDS control and prevention, Tianjin Binhai New Area Tanggu Center for Disease Control and Prevention, Tianjin, China
Corresponding author: Wenli Lu, Department of Health Statistics, School of Public Health, Tianjin Medical University, Meteorological Station Road No 22, Heping District, Tianjin, 300070, China. Email: [email protected]
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
197
and eliminating return visits. This enabled the tested people to obtain definitive negative and preliminary positive results in a single visit, which increased the number of clients receiving their test results.5 Awareness of HIV status influences subsequent behaviours. A meta-analysis of HIV-CT showed that risk reduction counselling could be effective among HIV-positive persons but not among HIV-negative individuals.6 The 2006 CDC testing guidelines recommend that prevention counselling should not be a necessary part of HIV screening projects and could only be offered to HIV-positive persons.2 This may increase the risk of disinhibited post-test risk behaviours among persons who receive HIV-negative test results through RVCT. Compared with RVCT, clients in SVCT can interact longer with the counsellor, obtain more information on safe behaviours and have an interval to reflect their risk behaviour before learning their HIV serostatus. A previous meta-analysis reported that RVCT could increase receipt of HIV test results in pregnant women.7 However, the efficacy of RVCT in other medical settings was not confirmed. On the other hand, the risk of disinhibited behaviour after receiving a negative HIV result by RVCT remained uncertain. Therefore, we performed a meta-analysis of RVCT on receipt of HIV test results and post-test risk behaviours among people who accepted HIV-CT in different settings. For obtaining a rigorous estimate of intervention effect, we only included randomised controlled trials (RCTs) in our study.
Methods Eligibility criteria Inclusion criteria: (1) Type of studies: RCTs studying HIV-CT in health care or outreach settings published from 1990 to 2012 in English. (2) Type of participants: participants of any age who underwent HIV-CT were included. (3) Type of intervention: RVCT was defined as rapid HIV testing completed within 20 min, and receipt of HIV test result and post-test counselling completed in one visit. SVCT was defined as HIV testing completed within one or two weeks, and participants needed a second visit to receive the test result and posttest counselling. (4) Type of outcome measures: the primary outcome included receipt rate of HIV test result and incidence of risk behaviours after HIV-CT. Exclusion criteria: (1) Type of study: review (including systematic review), report of conference, etc., and studies with incomplete data of interest and duplicate publications. (2) Type of participants: those women who accepted HIV-CT in the perinatal period or while in labour.
Data source Published literatures on effects of RVCT versus SVCT on Receipt Rate and Post-test Risk Behaviours were independently searched by two reviewers in electronic databases (Medline (Pubmed, 1966-till date), EBSCO (Academic Source Complete, Research StartersEducation, Research Starters - Sociology, ERIC, Library Information Science & Technology Abstracts, 1985-till date), Web of science ((Science Citation Index Expanded (1998-till date), Conference Proceedings Citation Index - Science (2004-till date)), Cochrane library (Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, Cochrane Central Register of Controlled Trials)) up to August 2012, complemented by manual search. Finally, the authors of possible relevant studies were contacted when more information or clarification was needed. The search terms were rapid* and HIV voluntary counselling and testing* and reception*; rapid* and HIV voluntary counselling and testing* and risk behaviours*. The reference lists of relevant reports and recent relative meta-analysis were also reviewed, but no additional study was identified.
Study selection Two review authors (Yuan Wang, Jian Guo), working independently and in parallel, scanned related abstracts and obtained the full text reports of studies when the abstract indicated or suggested the study was on SVCT or RVCT. After obtaining full reports of the candidate study (either in full peer-reviewed publications or press articles), the same review authors independently assessed eligibility of the studies for inclusion in the review. In the case of multiple publications of the same study or overlapping data sets, only data from the largest or most updated results were included. Any initial disagreement was resolved by consensus after additional review of the articles with the third author (Wenli Lu). Missing information was sought and unclear information clarified by contacting the corresponding authors of the eligible reports.
Data extraction The same two review authors extracted data from the eligible studies independently using a standardized form. Information was extracted from eligible articles on: (1) Characteristics of study: first author, year of publication, period of enrollment, location, and setting. (2) Type of intervention: approach of RVCT versus SVCT. (3) Type of outcomes: number of participants allocated, tested, and received HIV test result; risk behaviours (2 sex partners, unprotected sexual
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
198
International Journal of STD & AIDS 26(3)
intercourse). Any disagreement in data collection was adjudicated by the same third reviewer.
Assessment of risk of bias Risk of bias of eligible studies was assessed by the same two review authors independently using the Cochrane Collaboration’s tool.8 Six domains were evaluated: sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective outcome report and other sources of bias. The quality of evidence was assessed with GradePro. It contained study limitations of design or performance, heterogeneity of results, directness of evidence, precision of results and publication bias.
Analytic methods Relative risk (RR) and 95% confidence interval (CI) were used to estimate effect sizes. The pooled RR and 95% CI for overall effects of RVCT versus SVCT were constructed with random-effects model on an intention to treat (ITT) basis, because the variance across studies was conservative. Only two studies assessed risk behaviours after HIV-CT with a different approach; therefore, we emphasized on describing the results within qualitative synthesis rather than meta-analysis. Heterogeneity was evaluated by the Breslow-Day test, which was suggested by Higgins (termed I2).9 We explored the potential source of heterogeneity using subgroup analysis on risk of bias, research setting and source of funding. Because studies of outpatient and inpatient departments were associated with low HIV risk, these studies were classified into the category of hospital setting. Studies performed in high HIV risk settings including bathhouses for men who have sex with men (MSM), needle exchange sites and clinics for sexually transmitted infections (STIs) were classified as non-hospital settings. Forest plot was used for presenting meta-analysis results. Funnel plot was used for assessing publication bias. The symmetry was evaluated visually and formally with Egger’s test, performed with Stata12.0.10 This study was performed according to the PRISMA guidelines for meta-analysis.11 Data analysis was performed with Review manager 5.0.
excluding 465 duplicate publications, 2045 articles were reviewed of which 2024 articles were excluded as they did not meet the inclusion criteria. After reviewing the full text of remaining articles, five articles from four studies were included in this systematic review (Figure 1).
Study characteristics Three studies recruited male and female participants,12–14 and one study recruited only female participants.15 The study settings included public STI clinics, clinic hospital, needle exchange and MSM bathhouse. The duration of research ranged from 4–10 months (Table 1). In the report of Speilberg_2005, there were two studies performed in different settings: MSM bathhouse and needle exchange, so that this report was classified into two studies: Spielberg_2005A (Needle exchange), Spielberg_2005B (MSM bathhouse). 1. Participants: the eligible studies contained 4 412 participants. All of the participants were above 14 years old, were proficient in English and were unaware of HIV status before study entry. 2. Intervention and control: all studies were conducted between 1999 and 2007 in the USA. The following procedures were contained in RVCT arm: singlesession counselling, streamlined counselling and standard counselling; HIV test with Single Use Diagnostic System for HIV–1 test (SUDS), Orasure or Oraquick. In SVCT arm, clients would need a second visit for their HIV test results after their first visit including brief two-session counselling and standard counselling; and HIV test with EIA or Orasure. Although some studies (Metcalf_2005, Spielberg_2005A, Spielberg_2005B) used oral fluid test (Orasure) in SVCT arm, their clients needed a return visit for their HIV test results. 3. Outcomes: the outcomes evaluated were receipt rate of HIV test result and incidence of risk behaviours after HIV-CT. All of the five RCTs compared the result receipt rate. Only two RCTs assessed risk behaviour after HIV-CT, but with a different approach.
Risk of bias within studies Results Study selection A total of four studies containing five RCTs were identified for inclusion in our study. The search from electronic databases provided 2 510 articles. After
Three RCTs were considered as having adequate sequence generation (Metcalf_2005, Spielberg_ 2005(A,B)), indicating a low risk. In the other two RCTs (Wurcel_2005, Anaya_2008), the process of sequence generation was mentioned but not clearly described, and their risks were judged to be high.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
199
Figure 1. Flowchart of included studies.
Only one RCT (Metcalf_2005) used allocation concealment which blinded assignment to laboratory staff and data analysts, but not to participants and study staff. None of the remaining RCTs reported implementing concealment or blindness. Because receipt of HIV test result was unlikely to be influenced by awareness of allocation, the risks of bias about concealment and blindness among all RCTs were judged to be low. Two studies (Metcalf_2005, Spielberg_2005B) reported close proportions of incomplete outcome between RVCT and SVCT, indicating a low risk. Other RCTs had a disproportional loss of those participating in HIV-CT (Spielberg_2005A, Anaya_2008, Wurcel_2005,) which was not discussed and were considered as high risk. All RCTs were free of selective reporting, and the risks were judged to be low. One RCT (Metcalf_2005) was free of other bias because the characteristics of RVCT and SVCT were adequately reported and the study design was appropriate, indicating a low risk. The sample of Wurcel_2005 and Anaya_2008 was relatively small, which perhaps did not adequately power the outcomes. In addition, the RCTs of Spielberg_2005A and Spielberg_2005B did not report the baseline characteristics between the two groups. Thus, other sources of bias in these four RCTs (Spielberg_2005A, Spielberg_2005B, Wurcel_2005, Anaya_2008) were judged to be high. Overall, the studies of Metcalf_2005 and Spielberg_2005B were considered to have
low risk of bias. The risks of bias for Wurcel_2005, Spielberg_2005A, and Anaya_2008 were judged to be high.
Results of meta-analyses and qualitative analyses All of five RCTs contained in the four studies reported that the receipt rate of HIV test results in the RVCT was significantly higher than in the SVCT. The pooled effect also showed that RVCT increased the receipt rate significantly compared with SVCT (N ¼ 4 412, RR ¼ 1.74, 95% CI ¼ 1.47–2.07) (Figure 2) Two RCTs (Metcalf_2005, Anaya_2008) assessed the risk behaviours after HIV-CT. However, the behaviours were measured with a different approach in the two studies; therefore, we emphasized on describing the results within qualitative synthesis rather than metaanalysis. The first RCT, Meltcalf_2005, compared the incidence of risk behaviours three months after HIV-CT (N ¼ 2 495). The rate of follow-up was 76.4% in RVCT and 75.0% in SVCT. There was no significant differences on the behaviours of 2 sex partners (RVCT: 33.7%, SVCT: 30.0%; RR ¼ 1.11, 95% CI ¼ 0.99– 1.25), any unprotected sex (RVCT: 64.2%, SVCT: 62.5%; RR ¼ 1.03, 95% CI ¼ 0.97–1.09), any unprotected sex with non-primary partner (RVCT: 18.7%, SVCT: 15.8%; RR ¼ 1.18, 95% CI ¼ 0.99–1.41), any unprotected sex when drunk or high (RVCT: 23.7%,
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
EIA, Orasure
EIA
SUDS
Oraquick
Oraquick, Orasure Standard counseling
MSM bathhouse
Inpatients, outpatients of hospital Hospital, outpatient clinic
Seattle
Massachusetts
Southern California –
EIA
EIA, Orasure SUDS
Standard counseling; pretest written material Standard counseling; pretest written material Standard counseling Needle exchange
1999–2000 (122 days) 1999–2000 (122 days) 2003.9–2004.6
SVCT: 22.2%; RR ¼ 1.07, 95% CI ¼ 0.92–1.23), sex with new partner on day first met (RVCT: 6.7%, SVCT: 7.4%; RR ¼ 0.90, 95% CI ¼ 0.67–1.20), and one-time sex partner (RVCT: 20.1%, SVCT: 18.7%; RR ¼ 1.07, 95% CI ¼ 0.91–1.26). In addition, the Meltcalf_2005 study assessed the incidence rate of STI and did not find any significant difference between the two groups at 12 months after HIV-CT.12 The second RCT, Anaya_2008, evaluated the post-test reduction of sexual risk behaviours, and 81.0% of clients in RVCT and 73.6% of clients in SVCT completed the four weeks follow-up (N ¼ 191). There was no significant differences on reduction of risk behaviours between the two groups (RVCT: 48.5%, SVCT: 46.3%; RR ¼ 1.05, 95% CI ¼ 0.77–1.43).15
Risk of bias across studies Significant evidence of heterogeneity was found (I2 ¼ 93%, P < 0.001). To explore the heterogeneity, we conducted subgroup analysis according to risk of bias, setting and funding of research. 1. Risk of bias: The low-risk RCTs (Metcalf_2005, Spielberg_2005B) and the high-risk RCTs (Spielberg_2005A, Wurcel_2005, Anaya_2008) were associated with a higher RR (low: N ¼ 3 734, RR ¼ 1.37, 95% CI ¼ 1.33–1.41; high: N ¼ 678, RR ¼ 2.35, 95% CI ¼ 1.43–3.85). The I2 of the low and high group was 0% and 90%, respectively (Figure 2) 2. Setting: the non-hospital setting RCTs of Metcalf_2005, Spielberg_2005A, and Spielberg_ 2005B (N ¼ 4 058, RR ¼ 1.40, 95% CI ¼ 1.31–1.51) had a narrower interval than the hospital setting of Wurcel_2005 and Anaya_2008 (N ¼ 354, RR ¼ 2.85, 95% CI ¼ 1.86–4.37). The I2 of the non-hospital setting and the hospital setting decreased to 68% and 70%, respectively (Figure 3). 3. Funding: The Melcalf_2005 was funded by CDC (N ¼ 3 297, RR ¼ 1.38, 95% CI ¼ 1.33–1.42). The I2 was 97% for the remaining four RCTs funded by other organizations (N ¼ 1 115, RR ¼ 2.03, 95% CI ¼ 1.25–3.29) (Figure 4).
Anaya_2008
Wurcel_2005
Spielberg_2005B
Spielberg_2005A
Meltcalf_2005
Sensitivity analysis
Study ID
Approach of counselling
EIA, Orasure SUDS Brief two-session counseling
Single-session counseling Single-session counseling Single-session counseling Standard counseling Streamlined counseling STI clinics
Denver, Long beach, Newark Seattle 1999.2–2000.12
Setting Location Period of enrollment
Table 1. Summary of RCTs evaluating the receipt rate of HIV test results between RVCT and SVCT.
Approach of testing
SVCT RVCT SVCT
International Journal of STD & AIDS 26(3)
RVCT
200
A fixed-effects model was conducted for testing the robustness of our results. Compared with the random-effects model, the fixed-effects model provided less conservative results and identical patterns of results.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
201
Figure 2. Meta-analysis result of receipt rate stratified by risk of bias.
Figure 3. Meta-analysis result of receipt rate stratified by research setting.
Publication bias To assess the presence of publication bias, a funnel plot was drawn (Figure 5), and the Egger test was used to
investigate funnel plot asymmetry. No evidence of publication bias was observed in the receipt rate of HIV test result studies (Egger test: interception ¼ 0.548, P ¼ 0.514).
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
202
International Journal of STD & AIDS 26(3)
Figure 4. Meta-analysis result of receipt rate stratified by funding source of research.
Figure 5. Funnel plot for assessing publication bias.
Discussion Summary of main results A total of four studies involving five RCTs were eligible for inclusion in the review. The overall effect of the
meta-analysis suggested that the result receipt rate was significantly higher in the RVCT than in the SVCT. On the other hand, the evidence from the two studies comparing the risk behaviours after HIV-CT did not show significant difference between RVCT and SVCT.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
203
Therefore, it was not robust enough to determine the effect of RVCT on post-test risk behaviours.
Implications for practice An effective strategy for HIV prevention is HIV-CT. The main purposes of HIV-CT were to prompt people to know their HIV serostatus, to provide medical care to the infected clients, to provide counselling to reduce clients’ risk behaviour for avoiding infection and to prevent infected clients from spreading HIV.16 The transmission rate of people who were unaware of their positive status was higher than people who were aware of their infections.17 In the USA, approximately 14,000 new infections per year could be attributed to unawareness of HIV infection.18,19 In low-income countries such as sub-Saharan Africa, this proportion ranged from 30% to 70%.20 Failure to return for HIV test results was a possible reason for unawareness of infection. As an alternative strategy of SVCT, RVCT can effectively elevate awareness of HIV serostatus. Because most persons test negative, RVCT allows clients to receive counselling and HIV test results in one visit. This saves time and cost of a return visit and increases the proportion of receiving HIV test result among the tested clients, especially for the less reluctant ones. Rigorously designed RCTs were associated with adequate sequence generation, complete outcome data and comparable baseline, indicating reliable conclusions. This led to low risk of bias and homogeneity of the pooled effect (I2 ¼ 0%), suggesting high quality of evidence. The heterogeneity in studies with high risk of bias (I2 ¼ 90%) may be because of the relative small sample sizes. Therefore, studies with larger sample sizes are needed in future research. On the other hand, heterogeneity may indicate that the pooled effect was more vulnerable to variations in study location, setting and population. Early diagnosis of HIV infection helped clients access timely medical care and reduce risk behaviours, which avoided further transmission.21 In the area with high prevalence but low receipt rate, the counsellor or physician could use RVCT to increase the number of clients who become aware of their HIV serostatus, which reduces the difficulty of locating infected clients. The subgroup analysis of research setting also showed that RVCT increased the receipt in both hospital and non-hospital settings. However, the effect of RVCT on increase of receipt rate was lower in non-hospital settings than in hospital settings, indicating that RVCT may be more efficient in hospital settings. The potential reason could be that compared with patients in hospital, persons in high-risk settings, such as bathhouses, needle exchange sites and STI clinics were more likely
to return for test results because of their risky behaviours and fear of being infected. The result receipt rate of SVCT was clearly lower in hospital settings than non-hospital settings in our studies, reducing the RR between the two settings. On the other hand, the utilization of different modalities of rapid HIV testing (venipuncture and oral fluid) may be another explanation for the significant difference of pooling effects between hospital settings and non-hospital settings. Because oral fluid testing was more acceptable than blood testing, the unbalanced distribution of testing approaches may also affect the effectiveness of RVCT in the two settings. The CDC was responsible for widespread dissemination of HIV counselling and testing. The Meltcalf_2005 study was a multicentre study in STI clinics; funded by CDC it accounted for the high proportion of participants in the meta-analysis. The network of CDC provides support for recruiting large number of participants in three locations including Denver, Long beach and Newark. Although investigators may tend to favour interventions or products of the sponsor, no statistical difference was found between pooling effects of the CDC-sponsored study and nonCDC sponsored study. However, the relative smaller sample size of studies funded by non-CDC organizations might limit the generalizability of the findings. Costs of the HIV-CT procedure, prevalence and receipt rates after testing determine the entire costs of the HIV-CT approach. Compared with SVCT, the process of RVCT was more expensive because of the cost of the test kit and further counselling for the infected clients.22 However, considering the acceptance rate of testing and the receipt rate of results, the cost of informing an infected client was lower with RVCT than with SVCT in STI clinics and emergency departments.22 A costeffectiveness study also recommended that RVCT could be used as a routine HIV-CT approach for adults in areas where the prevalence of undiagnosed HIV infection is higher than 0.2%.23 Therefore, when planning HIV-CT policy, RVCT is an ideal alternative strategy that provides opportunity for people with potential for infection to learn their HIV serostatus. Our study was insufficient to confirm the difference in post-test risk behaviours between RVCT and SVCT. Previous meta-analysis focusing on the effect of HIVCT on sexual risk behaviour revealed that HIV-CT could reduce sexual risk behaviours among the HIVpositive clients but not among the HIV-negative ones.21 Some investigators argued that according to the Information-Motivation-Behaviour Skills model, perception of risk was the precondition of modification of behaviour.24 Anxiety of being infected was a critical motivation for people who seek HIV test.25,26 After learning the negative test result, the anxiety of being
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
204
International Journal of STD & AIDS 26(3)
infected was eliminated for clients in both RVCT and SVCT, which might lead to similar effect on modification of risk behaviours.
Quality of the evidence No evidence of publication bias was observed in the receipt rate of HIV test results; however, it should be noted that the low number of RCTs limits the potential for evaluating substantial publication bias. The quality of evidence presented was discussed on the basis of the GRADE approach. The receipt of test result from the included studies was graded as moderate. However, the risk behaviours after RVCT can not be evaluated based on present studies, indicating a very low grade. The chief objective of HIV-CT is to help people learn whether they are infected or not. Awareness of HIV serostatus determines the client’s future behaviours. Some people in the SVCT fail to return for the test results, but RVCT can solve this problem and increase the number of people becoming aware of their HIV serostatus. At present, although the effect of RVCT on the post-test risk behaviours is uncertain, we evaluate RVCT as a strong contender for increasing receipt of HIV test results.
Limitations There are several limitations in our study. First, as with other systematic reviews, the study population and setting of research was inconsistent across studies. This led to a high level of heterogeneity in our meta-analysis, which could be because of the research setting in the subgroup analysis. Second, the four studies other than the Metcalf_2005 study provided a limited sample size. It should be noted that a meta-analysis of several small studies does not predict the results of a single large study. The inequality may affect extrapolation of study conclusion. In future research, RCTs with larger sample sizes are needed to increase representativeness of different circumstances. Third, because of the limited financial and database resources, our search process did not cover databases such as Embase. The included studies were restricted to reports published in English owing to limitations in translating other languages. Finally, all of the studies included in our meta-analysis were performed in the USA. It limited the application of the result in other countries with different socioeconomic background and HIV prevalence.
Conclusions In our study, we recommend RVCT as a favourable method to increase the receipt of HIV test results.
Only two included studies assessed the modification of risk behaviour after HIV-CT in a different manner; also, the sample size was small in the current meta-analysis. In future research, it is necessary to confirm the effect of RVCT on disinhibition of post-test risk behaviour. Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by The Ministry of Education Humanities and social science research (Project No. 13YJCZH188). It was designed at the school of Public Health in Tianjin Medical University and will carry out at Tianjin Binhai New Area Tanggu Center for Disease Control and Prevention.
References 1. Centers for Disease Control and Prevention (CDC). HIV prevalence estimates—United States, 2006. MMWR Morb Mortal Wkly Rep 2008; 57: 1073–1076. 2. 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. 3. Valdiserri RO, Moore M, Gerber AR, et al. A study of clients returning for counseling after HIV testing: implications for improving rates of return. Public Health Rep 1993; 108: 12–18. 4. Sullivan PS, Lansky A, Drake A, et al. Failure to return for HIV test results among persons at high risk for HIV infection: results from a multistate interview project. J Acquir Immune Defic Syndr 2004; 35: 511–518. 5. Centers for Disease Control and Prevention (CDC). Update: HIV counseling and testing using rapid tests – United States, 1995. MMWR Morb Mortal Wkly Rep 1998; 47: 211–215. 6. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89: 1397–1405. 7. Minnier KS, van der Walt SJ and Klopper HC. A systematic review of counselling for HIV testing of pregnant women. J Clin Nurs 2009; 18: 1827–1841. 8. Higgins JPT and Green S. Cochrane handbook for systematic reviews of interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, www.cochranehandbook.org (2011). 9. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. 10. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
Wang et al.
205
11. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med 2009; 151: W65–W94. 12. Metcalf CA, Douglas JM Jr, Malotte CK, et al. Relative efficacy of prevention counseling with rapid and standard HIV testing: a randomized, controlled trial (RESPECT2). Sex Transm Dis 2005; 32: 130–138. 13. 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. 14. Wurcel A, Zaman T, Zhen S, et al. Acceptance of HIV antibody testing among inpatients and outpatients at a public health hospital: a study of rapid versus standard testing. AIDS Patient Care STDS 2005; 19: 499–505. 15. Anaya HD, Hoang T, Golden JF, et al. Improving HIV screening and receipt of results by nurse-initiated streamlined counseling and rapid testing. J Gen Intern Med 2008; 23: 800–807. 16. Centers for Disease Control and Prevention (CDC). Technical guidance on HIV counseling. Center for Disease Control and Prevention. MMWR Recomm Rep 1993; 42: 11–17. 17. Marks G, Crepaz N and Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 2006; 20: 1447–1450. 18. Hall HI, Song R, Rhodes P, et al. Estimation of HIV incidence in the United States. JAMA 2008; 300: 520–529.
19. Campsmith ML, Rhodes PH, Hall HI, et al. Undiagnosed HIV prevalence among adults and adolescents in the United States at the end of 2006. J Acquir Immune Defic Syndr 2010; 53: 619–624. 20. Department WHOHA. Progress report 2011: Global HIV/AIDS response epidemic update and health sector progress towards universal access WHO, UNICEF, UNAIDS. 2011. 21. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89: 1397–1405. 22. Farnham PG, Hutchinson AB, Sansom SL, et al. Comparing the costs of HIV screening strategies and technologies in health-care settings. Public Health Rep 2008; 123: 51–62. 23. Paltiel AD, Walensky RP, Schackman BR, et al. Expanded HIV screening in the United States: effect on clinical outcomes, HIV transmission, and costs. Ann Intern Med 2006; 145: 797–806. 24. Fisher JD and Fisher WA. Changing AIDS-risk behavior. Psychol Bull 1992; 111: 455–474. 25. Takahashi S, Yamazaki Y and Kawata C. Client anxiety from awareness of personal HIV risk and relation to taking HIV test at public health centers in Japan. Nippon Koshu Eisei Zasshi 1999; 46: 275–288. 26. Earl A and Albarracin D. Nature, decay, and spiraling of the effects of fear-inducing arguments and HIV counseling and testing: a meta-analysis of the short- and longterm outcomes of HIV-prevention interventions. Health Psychol 2007; 26: 496–506.
Downloaded from std.sagepub.com at UNIV OF MICHIGAN on July 2, 2015
