Verification of HIV Self-Testing Use and Results: A Global Systematic Review

Abstract

HIV self-testing (HIVST) allows individuals to interpret and report their own test results, thus decentralizing testing. Yet, this decentralization can make it difficult to verify self-testing results, which is important for linkage to care and surveillance. The aim of this systematic review is to summarize methods for verifying HIVST use and results. We followed guidance from the Cochrane Handbook 5.1 on systematic reviews. We searched four journal databases (PubMed, Embase, Scopus, and Cochrane Library), one clinical trials database (ClinicalTrials.gov), two conference abstract databases (International AIDS Society and Conference on Retroviruses and Opportunistic Infections) and one gray literature database (OpenGrey). We included studies that verified opening of kits or test results. Two researchers independently screened articles and extracted data regarding HIVST location, method of verification, who performed verification, proportion of results verified, and primary or secondary kit distribution. The search yielded 3853 unique citations, of which 40 contained information on HIVST verification and were included. Among these 40 studies, 13 were in high-income countries, 16 were in middle-income countries, and 11 were in low-income countries. Seventeen studies included key populations and two focused on youth. Three methods verified results: supervision by a health provider, returning used test kits, and electronic transmission of photographs. One method verified opening of kits using Bluetooth sensors. Although HIVST has increased worldwide, strategies to verify self-testing results remain limited. These findings suggest a need for additional innovative strategies for verifying HIVST use and results and linkage of self-testing results to surveillance and care systems.

Introduction

HIV self-testing (HIVST) is recommended by the World Health Organization (WHO) as a complementary HIV testing strategy.¹ It is a process by which individuals collect their own blood or oral fluid sample, conduct the diagnostic test, and then interpret their results.¹ With 9.4 million people living with HIV (PLWH) unaware of their status,² HIVST is an important approach toward the first of The Joint United Nations Programme on HIV/AIDS (UNAIDS) 90-90-90 targets, that is, 90% of PLWH will know their status by 2020.³ HIVST may reach individuals at high-risk for HIV acquisition who require frequent testing and key populations who may avoid standard HIV testing services due to stigma and discrimination.^1,4 Moreover, it has several advantages compared to provider-delivered HIV testing services, such as enhanced confidentiality for testers⁵ and potential cost savings for health systems.⁶ This form of testing also provides greater convenience for testers by mitigating barriers related to accessing care, particularly traveling far distances to clinics^7,8 and long waiting times.^9–11 HIVST has the ability not only to expand uptake of HIV testing but also to increase the autonomy a person has over their care.

Despite the potential of HIVST in increasing testing access, there are challenges associated with verifying the results for individuals who self-test. Verification of testing and results involves methods that can link the test result to the individual performing the self-test. The shift of HIV testing away from health care facilities makes it difficult to verify self-testing results, which are keys for confirming HIV status, linkage to care, and routine surveillance. Many HIV projects have relied on self-reported outcomes of self-testing,^12–14 which has important limitations. Individuals may inaccurately report their HIVST use or HIV status,¹⁵ possibly due to social desirability bias, fear of stigma, or loss to follow-up. This can impact programming for HIV services and make it difficult to measure HIV prevalence in a population, to determine the proportion of persons who are aware of their HIV status, and to conduct epidemic surveillance.

We conducted a systematic review to consolidate global literature on reported methods that are used to verify HIVST, describe the different methods, and explore the appropriateness of each method based on testing conditions.

Methods

Search strategy

Our search strategy followed best practices as defined in the Cochrane Handbook version 5.1.¹⁶ We developed a comprehensive search, in collaboration with a health sciences research librarian, to identify studies that used HIVST verification methods published from January 1, 2008 to October 15, 2018. We searched four journal databases (PubMed, Embase, Scopus, and the Cochrane Library), one clinical trials database (ClinicalTrials.gov), two conference abstract databases (International AIDS Society and Conference on Retroviruses and Opportunistic Infections) and one gray literature database (OpenGrey). The search strategy consisted of words related to two key terms, (1) HIV or AIDS and (2) self-test or home test. We consolidated variations within each key term using the “OR” Boolean operator and combined the key terms using the “AND” operator. We include the full search strategy in the Supplementary Table S1.

Study selection

We included original quantitative studies that had a focus on HIVST, described a method to verify HIVST results, were written in English, and were published between 2008 and 2018. We excluded HIVST studies before 2008 for three reasons: (1) the accuracy of HIVST kits before 2008 was not well-established and the technology has evolved in the past decade; (2) the policy environment actively discouraged self-testing compared to official WHO and national policies supporting self-testing now; and (3) the greater requirements for informed consent before HIV testing created obstacles for HIVST research before 2008.

We identified studies through database search and imported them into EndNote (X7.8), where we deleted duplicate articles. Two researchers (K.M.T. and J.J.O.) independently screened the titles of the remaining unique articles and selected studies for abstract review. We uploaded the EndNote file consisting of the selected studies to Covidence, an online screening and data extraction tool.¹⁷ We then screened the abstracts and full texts to determine which articles were eligible for study inclusion. The two researchers resolved any discrepancies during a meeting to discuss study selection.

Data extraction and analysis

We developed a data extraction form that was independently used by two researchers (K.M.T. and J.J.O.). We extracted the following general study information: author name, year of study, geographic location, study design, sample size, and sample demographics (sex, age, population), and type of HIVST. We also extracted the following HIVST verification details: clinic- or community-based testing, type of verification method, who performed the HIVST results verification, percent of results verified, and primary or secondary kit distribution. We collected data regarding primary or secondary kit distribution to evaluate the performance of verification procedures when kits are distributed directly to the study participant or to their partners or members of their social network. A third researcher (N.E.R.) was available to resolve any discrepancies regarding study inclusion or data extraction.

We conducted a cross-study synthesis of HIVST verification methods, which included a quantitative count and descriptive review of the current reported methods. We also conducted random effects meta-analyses to obtain weighted average percentages with 95% confidence intervals (CIs) for studies that reported the proportion of participants who sent results for verification and participants who were linked to care. We measured the variation in study results due to heterogeneity using the I² statistic and performed an Egger's test for small-study effects. Meta-analyses were performed in Stata version 15.¹⁸

Quality assessment

We assessed the quality of included studies using The Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data.¹⁹ The checklist contains nine criteria pertaining to study design, conduct, and data analysis. Each criterion was categorized as a “yes,” “no,” “unclear,” or “not applicable.” Studies that received a total score of five or less were considered low quality. We include the quality assessment in the Supplementary Table S2.

Registration

We registered the study protocol in PROSPERO (CRD42019124421), the International Prospective Register of Systematic Reviews.

Results

We identified 5421 citations, of which 1568 citations were duplicates. Among the 3853 unique citations that were screened by title, 248 records were identified for abstract review. Subsequently, 187 articles were selected for full-text review, during which 147 studies were excluded for one of the following reasons: no verification method was described (n = 119), study was a systematic or scoping review (n = 18), not written in English (n = 1), or the study was not focused on HIVST (n = 9). Ultimately, 40 studies were included in our analysis (Fig. 1).

FIG. 1.

Study flow diagram. HIVST, HIV self-testing.

Characteristics of included studies

Of the 40 studies included (Tables 1–3), 13 were conducted in high-income countries, 16 in middle-income countries, and 11 in low-income countries.²⁰ Fifteen studies were randomized controlled trials, 14 were cross sectional, 7 were prospective cohort, and 1 each for surveillance, pilot screening service, facility-based integration of HIV testing, and implementation research.

Table 1.

Characteristics of Included Studies That Verified HIV Self-Testing Results Reporting Through Supervision

Geographic location	Author (study year)a and ref. no.	Study design	Sample size	Population	Primary or secondary kit distribution	Type of HIVST	HIVST kit distribution	HIVST administration	Who is verifying the test result?
Australia	Birrell21	Surveillance	465	MSM	Primary	Oral	Gay entertainment venues, sex-on-premises venues	Gay entertainment venues, sex-on-premises venues	Outreach workers
Brazil and Peru	Volk37	Prospective cohort	103	MSM	Primary	Finger prick	Study site	Study site	Researcher
Hong Kong	Wang33	RCT	430	MSM	Primary	Oral	Mail	Communitynot specified	Nurse
India	Shivkumar38	Cross sectional	202	Pregnant women	Primary	Oral	Clinic	Clinic	Health care worker
Kenya	Kelvin (2017)34	RCT	750	Sex workers	Primary	Oral	Clinic	Clinic	Health counselor
	Kelvin (2015)35	RCT	84	Truck drivers	Primary	Oral	Clinic	Clinic	Health counselor
	Kurth (2013)36	Cross sectional	20	Mixed population	Primary	Oral	Clinic	Clinic	Study staff
Malawi	Choko (2010)22	Cross sectional	260	Mixed population	Primary	Oral	Home	Home	Health counselor
Singapore	Ng (2008–2010)23	Cross sectional	994	Mixed population	Primary	Oral	Clinic	Clinic	Health care worker
South Africa	Smith39	Cross sectional	224	Youth	Primary	Finger prick	Clinic	Clinic	Health care professional
	Pai (2018)40	Prospective cohort	490	Heterosexual	Primary	Not specified	Clinic	Clinic	Not specified
	Martinez Perez (2014–2015)41	Cross sectional	2198	Heterosexual	Primary	Oral	Clinic	Clinic	Health counselor
Swaziland	Pasipamire (2017)42	Prospective cohort	1462	Mixed population	Primary	Not specified	Clinic	Clinic	Study staff
Thailand	Jantarapakde (2015–2017)43	Implementation research	208	MSM, TGW	Primary	Not specified	Not specified	Community—not specified	Health counselor
Uganda	Asiimwe (2013)26	RCT	123	Mixed population	Primary	Oral	Clinic	Clinic	Study staff
United Kingdom	Saunders (2015)44	Prospective cohort	199	Mixed population	Primary	Finger prick	Clinic	Clinic	Study staff
United States	MacGowan45	Prospective cohort	22	MSM	Primary	Oral	Study site	Study site	Study staff
	Gaydos24	Cross sectional	478	Mixed population	Primary	Oral, finger prick	Clinic	Clinic	Study coordinator
	Carballo-Diéguez25	Cross sectional	27	MSM	Primary	Oral	Study site	Study site	Researcher
	Stephenson (2016–)46	RCT	400	MSM	Primary	Oral	Mail	Home	Health counselor
	Stephenson (2017–)47	RCT	200	Transgender youth	Primary	Oral	Mail	Home	Health counselor
Zambia	Floyd (2016–2017)48	RCT	13,267	Mixed population	Primary	Oral	Home	Home	Health provider
Zimbabwe	Mavedzenge27	Cross sectional	172	Heterosexual	Primary	Oral	Clinic	Clinic	Study staff

^aStudy year indicates the year the study was conducted.

HIVST, HIV self-testing; MSM, men who have sex with men; RCT, randomized controlled trial; TGW, transgender women.

Table 3.

Characteristics of Included Studies That Verified HIV Self-Testing Results Reporting Through Electronic Transmission Photographic Verification or Internet-of-Things

Verification method	Geographic location	Author (study year)a and ref. no.	Study design	Sample size	Population	Primary or secondary kit distribution	Type of HIVST	HIVST kit distribution	HIVST administration	Who is verifying the test result?	Incentive or reimbursement	Percent of participants who send results for verification
Photographic verification	China	Jin (2017)55	Cross sectional	879	MSM	Primary	Finger prick	Not specified	Community—not specified	Not specified	$5 deposit to buy kit was returned after uploading image of test result	78.0
		Operario57	RCT	60	MSM	Primary	Not specified	Not specified	Community—not specified	Not specified	None	Not specified
		Tang (2016–2017)58	RCT	1219	MSM	Primary	Oral, finger prick	Mail	Community—not specified	Study staff	None	23.0
		Tao (2012)32	Cross sectional	220	MSM	Primary	Finger prick	Mail	Home	Not specified	$10 deposit refunded to those who completed the test and reported their result	100.0
		Zhong (2015)59	Prospective cohort	198	MSM	Primary	Finger prick	Mail	Community—not specified	Not specified	US $23 deposit returned after participants took the test and sent an image of their result	90.0
	United States	Mustanski (2015)56	RCT	901	MSM	Primary	Oral	Mail	Community—not specified	Study staff	None	Not specified
Bluetooth sensors	United States	Wray (2016–2017)60	RCT	65	MSM	Primary	Oral	Mail	Community—not specified	Study coordinator	Participants could receive a total of $180 for completing all study assessments	Not specified

^aStudy year indicates the year the study was conducted.

The mean and median sample size of participants who opted or were assigned to HIVST in the studies were 2454 and 330, respectively. Twenty-two studies involved community-based HIVST, 17 involved HIVST at a clinic or study site, and 1 study provided the option for participants to self-test in the clinic or in the community. Two studies used secondary kit distribution to reach partners or peers of the participants.

Two studies focused on youth. Seventeen studies focused on key populations, including 14 for men who have sex with men (MSM) and 1 each for female sex workers, transgender people, and MSM and transgender women together. One study in Kenya focused on truck drivers, as they are considered to be a key population in Africa due to their high risk for HIV. The remaining 21 studies focused on heterosexuals, mixed populations, and pregnant women. No studies were targeted for prisoners or people who inject drugs. The number of published studies or protocols on HIVST increased over the years: 5 studies in 2010–2012,^21–25 7 studies in 2013–2015,^26–32 and 28 studies in 2016–2018.^33–60

Verification methods

We identified three methods that verified HIVST results and one method that verified opening a self-test kit, each with their own strengths and limitations (Table 4). Twenty-three studies had a clinician or study team member supervise participants as they took an HIVST (Table 1).^{21–27,33–48} Ten studies requested used HIVST kits to be returned to the clinic or study site (Table 2).^{28–31,49–54} Six studies requested participants to take a photo and upload their test results to an online or mobile portal (Table 3).^32,55–59 One study equipped HIVST kits with Bluetooth sensors to detect the opening of kits and notify the study team (Table 3).⁶⁰ Below, we describe each of these verification methods in greater detail.

Table 4.

Strengths and Limitations of Each Verification Method

Verification of HIVST test results				Verification of opening HIVST kit
	Supervision/observation	Return of used kits	Electronic transmission of photographs	Bluetooth sensors
Strengths	Quickest way to verify HIVST results Not difficult to link HIV-positive individuals to care	Multiple ways to return kits Maintains privacy of testing Convenient option for testers Appropriate method for settings without strong technological infrastructure	Global ownership of phones is widespread Maintains privacy of testing Convenient option for testers	Real-time detection of opened kits Immediate notification of opening to a health provider Maintains privacy of testing Convenient option for testers
Limitations	Removes confidentiality of self-testing Facility-based supervision/observation does not eliminate barriers to standard HIV testing services	Delayed or no return of kits Difficult to ascertain if the lack of returned kit means the test was not used or that it was just not returned Do not have confirmation of who used the test kit, just who received the test kit	Requires access to the internet Issues regarding network security Do not have confirmation of who used the test kit, just who received the test kit	Does not verify that an individual took a self-test Do not have confirmation of who used the test kit, just who received the test kit

Table 2.

Characteristics of Included Studies That Verified HIV Self-Testing Results Reporting Through Returning of Used Kits

Geographic location	Author (study year)a and ref. no.	Study design	Sample size	Population	Primary or secondary kit distribution	Type of HIVST	HIVST kit distribution	HIVST administration	Who is verifying the test result?	Incentive or reimbursement	Percent of participants who send results for verification
Malawi	Choko (2012)28	Prospective cohort	16,660	Mixed population	Primary	Oral	Community	Home	Health counselor	None	74.6
	Choko50	RCT	800	Heterosexual	Secondary	Oral	Clinic	Home	Health counselor	Male partners who test and link to care can receive US $3, $10 or $30, depending on their randomized trial arm	Not specified
	Choko29	Prospective cohort	13,966	Mixed population	Primary	Not specified	Community	Community—not specified	Not specified	None	89.0
	Dovel (2017–)51	RCT	6000	Mixed population	Primary	Not specified	Clinic	Clinic	Study staff	None	Not specified
	MacPherson (2012)31	RCT	8194	Mixed population	Primary	Oral	Community	Home	Health counselor	None	64.9
Uganda	Nanfuka53	Cross-sectional	95	Heterosexual	Secondary	Oral	Community	Community—not specified	Not specified	None	98.9
United Kingdom	Brown (2016–2017)49	RCT	8999	Mixed population	Primary	Finger prick	Mail	Community—not specified	Not specified	None	54.2
	Greaves30	Pilot screening service	150	MSM	Primary	Oral	Public sauna	Community—not specified	Not specified	None	10.0
Zimbabwe	Hatzold (2016)52	Facility-based integration of HIVST	6278	Mixed population	Primary	Oral	Clinic	Clinic, community not specified	Health care professional	None	100.0
	Sibanda (2016)54	Cross sectional	7510	Heterosexual	Primary	Oral	Community	Community—not specified	Study staff	None	68.2

^aStudy year indicates the year the study was conducted.

There were 23 studies (Table 1) involving supervised HIVST. Supervision of HIVST involved a health professional, study staff, or community outreach worker watching participants self-test for HIV. Sixteen studies supervised participants in clinics or study sites, four of which focused on key populations. Seven studies utilized supervision outside of the clinic, of which four involved online counseling with HIVST.^33,43,46,47 In five of the seven studies where supervision was conducted outside of the clinic, the study participants consisted of key populations. Out of the 23 studies that utilized supervision as a verification method, 20 were conducted in middle- or high-income countries. All 23 studies involved primary kit distribution.

Ten studies requested participants to return their used HIVST kits (Table 2) in person,^28,31,50,51 through mail,^30,49 or at a designated drop box.^28,30,52,54 Seven of these studies distributed oral HIVST kits to participants. One study focused on key populations, particularly MSM. Eight of these studies were conducted in Malawi, Uganda, and Zimbabwe. The only two studies in the review to focus on secondary kit distribution both utilized returned used kits as a verification procedure. Eight studies reported the percentage of participants who returned used kits. The weighted average percentage of used kits return was 75% (95% CI: 60–87%, I² = 99.9%). The p value of small-study effects was 0.44, suggesting that the test for small-study effects did not provide evidence of publication bias. Two studies sent participants reminders, one through short message service (SMS) and one through phone calls, to return their used samples.^49,50 In the study that used SMS, the overall return percentage was 54.2% (95% CI: 53.2–55.2%).⁴⁹ This study found that SMS reminders that included behavioral insights—messages that prompted participants to set aside time to self-test, encouraged self-efficacy, reminded participants to return their used kits, and included a deadline for self-testing—increased the kit return percentage.

Photographic verification of self-test results, which was utilized in six studies (Table 3), involved participants taking a photo of their HIVST results and submitting the picture to the study team. Three studies allowed participants to submit their photos through an online database^55,56,59 and one study provided the option to submit through a website portal or e-mail.³² Two studies allowed participants to submit their photos through WeChat, an instant messaging service popular in China.^57,58 Of the six studies that used electronic transmission of photographic verification, five were conducted in China and one was conducted in the United States. All six studies focused exclusively on MSM and involved primary kit distribution. Four studies reported the percentage of participants who submitted photographs of their results. The weighted average percentage for electronic transmission of photographs was 78% (95% CI: 38–100%, I² = 99.7%). The p value of small-study effects was 0.40, suggesting that the test for small-study effects did not provide evidence of publication bias.

One study equipped HIVST kits with Bluetooth sensors and installed mobile applications on participant's smartphones to detect when kits were opened (Table 3). The Bluetooth sensors connected to the smartphones, which then sent a notification to the study team that a kit had been opened. This allowed investigators to reach out to participants (within 24 h in this study protocol) to provide counseling and referrals. The test kit sensors detected that 73.4% of the study kits had been opened over the course of entire study period; 100% of detected opened kits resulted in post-test counseling. Researchers believe that failure to detect kit opening could have been due to participants deleting the study mobile app from their phone or switching to a different phone altogether. This study involved MSM in the United States.⁶⁰

Secondary outcomes

Fifteen studies that verified HIVST results also reported the proportion of identified HIV-positive participants linked to care after self-testing.^{26,29,31–33,37,40–43,48,52,54,55,59} Eight of those used supervision to verify HIVST results,^{26,33,37,40–43,48} of which three were community-based.^33,43,48 Four studies verified results through returning kits^29,31,52,54 and three studies used electronic transmission of photographs.^32,55,59 None of the studies reported referral to HIV prevention services for participants who tested negative for HIV. All 15 studies involved primary kit distribution. The weighted average percentage for linkage to care was 85% (95% CI: 53–100%, I² = 99.5%). The p value of small-study effects was 0.02, possibly indicating evidence of publication bias. The percentage for linkage to care may not have been solely influenced by the verification method utilized. Other factors, such as stigma, fear of disclosure of HIV status, or denial of HIVST results could have affected linkage to care.

Discussion

This systematic review identified three methods to verify HIVST results: self-testing supervised by a study team member, returning of used kits, and electronic transmission of photographs. The review also found one method that used Bluetooth sensors to verify the opening of an HIVST kit. These methods have been implemented in various settings and are associated with a broad range of return rates. This study extends the literature on HIVST by focusing on verifying self-testing results beyond self-reporting and evaluating linkage to care after HIVST. Each verification method identified in this review has strengths and limitations.

Supervised HIVST may be more likely to ensure that test results are correctly linked to testers. Supervision by health care professionals or lay health workers may enhance linkage to care. However, supervision removes important benefits of HIVST, including convenience and confidentiality of testing and results. Sixteen of the 23 studies that verified HIVST through supervision had verification done at a clinical site, which maintains barriers typically associated with facility-based testing. There are many groups that may fear seeking facility-based HIV testing services, including undocumented migrants, people living in remote areas, and some sexual minorities.⁶¹ This fear of facility-based services could also make supervised HIVST verification not feasible.

Returning used self-test kits was used in eight studies in three low-income countries (Malawi, Uganda, and Zimbabwe). Returning self-test kits could be done at a community site, clinic, or through mail. Returning the used test kit at a community site could alleviate some of the transportation barriers that are known to discourage HIV care seeking.⁶² Moreover, self-testers returning their used HIVST kits to a community location may decongest health facilities because of fewer individuals who do HIV test. Given limited technological access, returning used kits may be an appropriate and useful verification method for low-resource settings and populations. However, this verification approach does not guarantee that the individual returning the used kit is the same individual who self-tested for HIV.

Photographic verification of HIVST was also used in one upper middle-income country (China) and one high-income country (United States). Photographic verification has a tester take a photograph of the completed test kit and send it through a private, secure transmission to a test organizer. All photo-verification studies were among MSM. Data from the United States⁶³ and China⁵⁸ suggest that men are comfortable taking a photograph of HIVST test results and sending it through a mobile platform. Further implementation science on photographic verification of HIVST among other key populations is needed. In addition, one study used Bluetooth sensors to detect opening of kits and notify study counselors. This verification method is an example of internet-of-things technology, in which devices communicate with one another through the internet to share and receive data. Internet-of-things technology is a method that can link self-test results to surveillance programs, and allow for real-time monitoring of self-testing procedures and timely reporting of information. Further research is needed on internet-of-things technology to better understand their role in public health surveillance, preventative care, and treatment. However, internet-of-things technology is not currently linked to completion of the test itself. Integration of internet-of-things to test completion could be a major advance for HIV surveillance, follow-up services, and linkage.

None of the HIVST verification methods identified guaranteed results notification. Testers may fail to verify results due to possible fear of disclosure if HIV positive or the stigma of being outed as being at-risk for HIV infection. Testers may also forget or delay verification procedures, contributing to subsequent delays in linkage to services. Verification procedures should emphasize to testers the advantages of verified HIVST results, particularly enhanced linkage to clinical services.

Given the broad scale-up of HIVST in a variety of settings, these results are timely and have public health and programmatic implications. Current HIVST verification methods are limited, but implementation of HIVST services should incorporate verification procedures to improve HIV test result capture. Ideally, these verification technologies would be performed in real time and connect patients to appropriate post-HIVST services, while retaining some of the important benefits of HIVST such as autonomy and confidentiality. Moreover, these procedures would be pertinent to the settings and populations, in which verification is performed to ensure uptake, feasibility, scale-up, and sustainability. Verification methods that not only verify results but also encourage patients to be engaged in their own care can better facilitate linkage to HIV care and related HIV services.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by NICHD UG3HD096929 and NIAID K24AI143471.

Supplementary Material

Supplementary Table S1

Supplementary Table S2