Accuracy of and preferences for blood-based versus oral-fluid-based HIV self-testing in Malawi: a cross-sectional study

Background HIV self-testing (HIVST) can use either oral-fluid or blood-based tests.

Studies have shown strong preferences for self-testing compared to facility-based services.

Despite availability of low-cost blood-based HIVST options, to date, HIVST implementation in sub-Saharan Africa has largely been oral-fluid-based.

We investigated whether users preferred blood-based (i.e. using blood sample derived from a finger prick) or oral fluid-based HIVST in rural and urban Malawi.

Methods At clinics providing HIV testing services ( n  = 2 urban; n  = 2 rural), participants completed a semi-structured questionnaire capturing sociodemographic data before choosing to test using oral-fluid-based HVST, blood-based HIVST or provider-delivered testing.

They also completed a self-administered questionnaire afterwards, followed by a confirmatory test using the national algorithm then appropriate referral.

We used simple and multivariable logistic regression to identify factors associated with preference for oral-fluid or blood-based HIVST.

Results July to October 2018, N  = 691 participants enrolled in this study.

Given the choice, 98.4% (680/691) selected HIVST over provider-delivered testing.

Of 680 opting for HIVST, 416 (61.2%) chose oral-fluid-based HIVST, 264 (38.8%) chose blood-based HIVST and 99.1% (674/680) reported their results appropriately.

Self-testers who opted for blood-based HIVST were more likely to be male (50.3% men vs. 29.6% women, p  < 0.001), attending an urban facility (43% urban vs. 34.6% rural, p  = 0.025) and regular salary-earners (49.5% regular vs. 36.8% non-regular, p  = 0.012).

After adjustment, only sex was found to be associated with choice of self-test (adjusted OR 0.43 (95%CI: 0.3–0.61); p -value < 0.001).

Among 264 reporting blood-based HIVST results, 11 (4.2%) were HIV-positive.

Blood-based HIVST had sensitivity of 100% (95% CI: 71.5–100%) and specificity of 99.6% (95% CI: 97.6–100%), with 20 (7.6%) invalid results.

Among 416 reporting oral-fluid-based HIVST results 18 (4.3%) were HIV-positive.

Oral-fluid-based HIVST had sensitivity of 88.9% (95% CI: 65.3–98.6%) and specificity of 98.7% (95% CI: 97.1–99.6%), with no invalid results.

Conclusions Offering both blood-based and oral-fluid-based HIVST resulted in high uptake when compared directly with provider-delivered testing.

Both types of self-testing achieved high accuracy among users provided with a pre-test demonstration beforehand.

Policymakers and donors need to adequately plan and budget for the sensitisation and support needed to optimise the introduction of new quality-assured blood-based HIVST products.