A recently published, Kenya-based, randomized controlled study (Masters & Thirumurthy/STIs) (M&T) evaluates a novel intervention that appears to combine in a fresh way elements of various innovative interventions for HIV prevention.  Recently published studies (e.g. Kissinger/STIs; Estcourt & Cassell/STIs) have explored the potential of ‘expedited’, or ‘accelerated’ partner therapy – where the partner of an index STI-infected individual is offered therapy through the infected individual directly, without a clinic attendance, or with only a telephone interview. They have also evaluated the benefits of ‘self-testing’ for HIV (e.g. Pai & Dheda/STIs).  The intervention trialled by M&T – partner delivered HIV self-testing – combines elements of these two types of interventions.  A third crucial element is that it takes advantage of the unique opportunity offered by antenatal and postpartum clinics in low resource settings (see Azeze & Haile/STIs; Sombie & Dabis/STIs; Ganiyu & Mason/STIs) to make contact with HIV-infected individuals who might otherwise have little access to health services.

In M&T’s study, participants attending antenatal and postpartum clinics (ANC) who were assigned to the intervention arm of the study (n=284/570) were offered two oral-fluid-based HIVselfing kits, one for themselves, one for their partner.  Discussion about HIV, self-testing, couple testing, and awareness of partner’s result were subsequently reported by index patients at monthly follow-up meetings over a three-month period.  This ANC-based, partner-delivered HIV self-testing intervention was evaluated against a control group (n=286/570) who were given an invitation card for conventional clinic-based HIV testing and encouraged to distribute the card to partners.

The primary outcome was reported partner HIV testing which was 91% for the intervention as against 52% for the control.  The intervention looked at couple testing (42% difference between intervention group and control), and awareness of partner’s HIV status (difference=39%).  0% participants in either group reported intimate partner violence (IPV) as a consequence of HIV testing.

Concerns around interventions of this kind tend to centre upon two things.  First, the absence of any direct contact between partner and health services, and consequent loss of an opportunity to test for the full range of STIs and help ensure integration into treatment (Estcourt & Cassell/STIs).  Second, worries as to the likelihood of IPV related to testing.  On the first, concern may be less where the self-test is for HIV, and the partner misses out on testing for the other STIs (Kissinger/STIs), than it would be where, as in many non-African settings, the self-test would generally be for Chlamydia or gonorrhoea and HIV could remain undiagnosed.  Regarding IPV, the authors note its lack of association with testing in the study (despite base-line reported rates of 27%).  They also point to the large proportion of eligible declining to participate (715/1,315).  This – taken together with zero cases of IPV – they interpret as a possible indication of the ‘agency’ of these women, and their capacity to make their own judgments regarding the possible impact of participation on their relationships.  The problem of ensuring engagement with care among those self-tested for HIV, however, is one that the authors acknowledge as still needing to be resolved.

As for the study itself, a primary limitation is self-reporting.  But here the authors note that misreporting is hardly likely to account for the differences between intervention and control arm – which is the evidence that the authors principally emphasize.