Can women safely distribute HIV oral self-test kits to their sexual partners? Results from a pilot study in Kenya

Introduction

Increasing uptake of HIV testing and counseling (HTC) and identifying HIV-infected individuals are a priority for all countries that subscribe to WHO/UNAIDS goal of an HIV-free generation by 2030 [1]. HTC is thus a crucial step towards achieving the promise of HIV treatment as prevention and achieving the first of the UNAIDS 90–90-90 targets [1]. However, globally and in sub-Saharan Africa, over 50% of HIV-infected individuals are unaware of their status [2]. Moreover, knowledge of HIV-positive status is lower in men than in women [3], with 61.9% of men and 47.8% of women in Kenya not knowing they have HIV [4].

New strategies are needed to increase HIV testing among men and other hard-to-reach populations. HIV self-testing (HIVST) is a strategy that is highly acceptable and feasible in many settings [6,7] and among diverse populations [8–14]. In particular, secondary distribution of self-tests in sexual networks is a promising approach to increasing partner and couples testing [11,14]. However, in the absence of face-to-face counseling, there is need to assess whether women who offer self-tests to their partners experience intimate partner violence (IPV) or other forms of social harm. This study describes the occurrence of IPV among women who accepted an intervention in which they received multiple self-tests for distribution to their sexual partners and compares IPV levels following the intervention to baseline levels of IPV.

Methods

We analyzed data from a larger observational cohort study that has been described elsewhere [11]. We enrolled female sex workers (FSW) from a Drop-in Center (DICE), a facility dedicated to serving key populations, and women seeking ante-natal care (ANC) and post-partum care (PPC) from a government health facility in Kisumu, Kenya. Eligibility criteria included being aged 18–39 years, HIV-negative, having at least one current sexual partner, and not anticipating violence from offering the self-test to partners.

Following informed consent, participants were administered a questionnaire that included questions about whether they had experienced IPV in the past 12 months. These questions were adapted from the Kenya Demographic and Health Survey [18], and participants were asked whether they experienced physical abuse, verbal or emotional abuse, and loss of economic support. In addition, each FSW was asked if in the past 12 months she had been forced by any partner or client to have unpaid sex or have sex without a condom, or to have anal sex when she did not want to. For the statistical analyses, IPV was determined to have occurred if participants responded affirmatively to any of the IPV questions.

Following the baseline questionnaire, participants were instructed on how to use the OraQuick Rapid HIV-½ antibody test (OraSure Technologies, Bethlehem, PA, USA), an oral fluid-based test. Research assistants, who were trained HTC counselors, performed a mock demonstration of self-test use and then asked participants to do a demonstration as confirmation that they understood the process well enough to test themselves and demonstrate to their partners. We developed a standard operating procedure for research assistants to prepare participants on how to minimize IPV risk when distributing self-tests to their partners. FSW were given 5 test kits while ANC and PPC women were given 3 test kits, each accompanied by written and pictorial instructions on usage in English, Kiswahili and Dholuo languages. Participants were also given a study helpline number and contact information for locally available IPV services. Participants were then followed monthly for 3 months and asked about distribution of self-tests to sexual partners and about IPV. Those reporting any IPV since enrollment were asked if this was related to self-test distribution.

Statistical analyses

The primary outcome variable was experience of IPV due to self-testing in the 3 months following enrollment. We conducted descriptive analyses of the primary outcome and of IPV incidence during the 12 months prior to enrollment. We then used chi-squared tests to determine whether participants who experienced IPV during the follow-up period were more likely to also have reported IPV during the 12-month period prior to enrollment.

Results

A total of 280 participants (61 ANC, 117 PPC, 102 FSW) were enrolled and 265 (94.6%) completed follow-up interviews. Baseline characteristics of participants have been described previously [11]. A large majority of the participants with a primary sexual partner who were followed-up distributed a self-test to their partner (53/58, 91% ANC; 91/106, 86% PPC; 64/85, 75% FSW) [11]. Overall, IPV at baseline ranged from 27% in the ANC group to 46% in the PPC group. IPV occurrence at follow-up was comparatively lower, ranging from 14% in the ANC group to 21% in the FSW group. Among all the study participants, 2 in PPC, 2 FSW and none in ANC reported experiencing IPV due to offering a self-test to a male partner.

Any IPV at baseline was strongly associated with any IPV at follow-up in each study group (Table 1; p-value<0.001). Participants in the FSW and ANC group who had experienced IPV at baseline were 40% more likely to have experienced IPV during the follow-up period. Among PPC participants, the same comparison indicated a 20% higher risk of experiencing IPV during the follow-up period. Very few participants (4% FSW, 2% PPC, and 6% ANC) who experienced IPV during the follow-up period reported no IPV at baseline. Of the 4 IPV cases that were reported to be associated with self-testing, 2 were among women who had experienced IPV at baseline and 2 were among women who had not. These IPV cases involved physical abuse, verbal abuse, or loss of economic support.

Table 1:

IPV at follow-up stratified by baseline IPV and study group

	ANC		PPC		FSW
	No IPV at baseline	IPV at baseline	No IPV at baseline	IPV at baseline	No IPV at baseline	IPV at baseline
Did not experience IPV at follow-up1	43 (98%)	9 (56%)	59 (94%)	39 (74%)	55 (96%)	25 (57%)
Experienced IPV at follow-up1	1 (2%)	7 (44%)	4 (6%)	14 (26%)	2 (4%)	19 (43%)
Difference in follow-up IPV by baseline IPV2	-	41% (17% to 66%)	-	20% (7% to 33%)	-	40% (24% to 55%)
P-value3		<0.001		<0.001		<0.001
Number of participants that distributed a self-test to their primary sexual partner	41	12	49	42	38	26
Experienced IPV due to testing at follow-up1*	0 (0%)	0 (0%)	1 (2%)	1 (2%)	1 (3%)	1 (4%)

Abbreviations: ANC, antenatal clinic; PPC, postpartum clinic; FSW, female sex workers; IPV, intimate partner violence.

Notes:

¹- N (%) presented;

²- % (95% CI) presented;

³- P-value from test of proportions;

^*- Percentages calculated from the number of participants who distributed a self-test to their primary sexual partner.

Discussion

Despite high underlying IPV prevalence in the study setting, a majority of women reported safely distributing self-tests to their primary sexual partner. A central finding of this study is that IPV risk during the 3-month intervention period was greatest among those who reported IPV prior to the intervention, and moreover, only four of the women who experienced IPV during the intervention period attributed it to the offer or distribution of self-tests to their sexual partners. These results underscore the reality that women are at very high risk of violence even in the absence of HIVST interventions and that the introduction of self-tests by women to their partners does not necessarily heighten the risk of IPV. Studies of other HIV testing strategies among couples show similar results. In Zambia, clinic-based couples testing did not lead to greater occurrence of IPV than individual testing [21]; in South Africa, there was no increase in IPV pre-post couples testing [22]; in Malawi there was no IPV reported post couples testing but women who anticipated violence did not invite partners back for testing [23]; and in the United States, there was no difference in IPV pre-post intervention between individual testing and couples testing [24].

Details of the four IPV cases that were related to study participation indicated that the violence experienced by women may have occurred even after regular, clinic-based HIV testing in the presence of a counselor [19]. Given the high underlying prevalence of IPV in this population, a key lesson that emerges is that IPV experienced by women receiving self-tests should not directly be attributed to HIVST per se but rather to various relationship dynamics that arise when individuals suggest testing or learn their partner’s status – dynamics that may arise even when counselor-administered HIV testing takes place.

An important factor that should be considered when interpreting our results is that women who self-selected to participate in the study may have been those who felt comfortable offering self-tests to their partners. This self-selection process is likely to closely resemble real world outcomes once self-tests are widely available, as only the women with a high degree of comfort offering self-tests to their partner are likely to do so by obtaining self-tests. Consistent with other studies [8,20] our results dispel fears regarding safety of HIVST and further the case that women attending services at health facilities can offer self-tests to men, thereby increasing HIV testing among them.