In-depth interviews with key informants and focus group discussions elucidated a variety of preference-relevant characteristics of HIV testing options, which map conceptually to three domains: confidentiality, quality, and accessibility.
Not surprisingly, as others have clearly documented [21–23, 30], concerns about confidentiality were preeminent and likely to affect HCT utilization. Our data highlight a level of unease about potential inadvertent disclosure of HIV status or HIV testing in association with specific testing venues, including free-standing and mobile VCT sites and testing at home. Similar apprehension was expressed about HIV testing conducted by young counselors, who were perceived to be possibly less discreet than older ones. Because the desire for confidentiality is linked to stigma, our findings highlight the importance of addressing both confidentiality and stigma in the design of newer approaches to HCT.
Concerns about accuracy were related to specific venues, the training of test providers, and the type of test. Some respondents feared that private or free-standing testing sites were prone to providing false positive results, whereas counselors with more experience, especially those working in large hospitals, are likely to deliver more accurate results. Variation in perceived accuracy was also related to differences in tests and testing procedures across venues. Home testing did not appear to be a popular option, even among those who had previously tested, in part due to confidentiality concerns, but also because the quality of home testing was potentially suspect. It has been previously documented that perceived unreliability of test results and distrust of HIV testing technologies can discourage uptake of HIV testing [22, 31, 32].
The findings highlight the conceptual overlap between testing venue, counselor, and test characteristics and suggest that individuals make important trade-offs in considering testing options. Some prefer to test at venues that see more clients, where they perceive the accuracy of the test to be better. Others, concerned more about confidentiality, may be willing to sacrifice perceived more accurate testing at a high-volume testing center for a perceived lower risk of being seen testing at testing sites with fewer clients The heterogeneity of preferences and the complex links between the domains of quality, confidentiality, and accessibility should be accounted for in the design or re-design of testing options.
To address both confidentiality and quality concerns, the integration of testing services into a hospital or health center setting may be preferable to isolated testing services offered at free-standing facilities. As attendance at free-standing HIV testing sites appears to be declining , plausibly because clients prefer to access HCT services within health facilities, HCT policy makers should examine ways to re-define the roles of free-standing VCT sites in this context.
The focus group discussions demonstrated heterogeneous preferences with respect to the accessibility of testing. For many participants, distance was not a significant barrier due to the local availability of a variety of testing options. For others, traveling seemed advantageous, as testing done farther from home is more likely to be confidential. Disparate views were expressed regarding direct payments as means of increasing accessibility of testing. We note that in the context of selected studies that addressed a slightly different question, high value conditional cash transfers, given in exchange for testing negative for sexually transmitted infections, were associated with reduced infection [34–36].
With the introduction of newer approaches of HCT delivery such as the use of community based lay counselors , couples counseling and testing [7, 8, 38], provider-initiated [39, 40], home based [5, 11, 12], and mobile HCT [13, 23, 41], it is important for planners of HIV testing interventions to recognize that many barriers are inextricably linked. Some are overlapping and reinforcing (e.g. concerns about both quality and confidentiality with home testing), and others demand that patients make trade-offs as they choose to test (e.g. paying for travel to reduce risk of disclosure within one’s community). Novel approaches to HCT delivery must weigh the benefits of addressing heterogeneous preferences against the costs and complexities of addressing the multi-faceted and interlinked barriers.
Our study is subject to important limitations. IDIs and FGDs afforded an opportunity to identify a variety of characteristics of HIV testing options associated with preferences, and to begin to understand which features are most important. However, our study suggests significant preference heterogeneity among participants, which precluded the development of a consensus regarding the relative importance of specific characteristic to participants, or to differentiate the preferences of individual sub-groups. The existence of heterogeneous preferences has been confirmed by a quantitative follow-up study in the area .
Second, it is not clear how participants’ stated preferences relate to actual testing decisions. The preferences and characteristics of individuals are likely to interact with characteristics of testing options (e.g. venue, method for obtaining the sample for the HIV test) to influence actual testing decisions. Further, due to the focus of FGDs and IDIs on characteristics of testing options, several important elements of the decision process could not be explored in detail, including differences between first-time and repeat testing, external motivators, such as social support for testing, and internal barriers to testing, such as fears of knowing the result.
Finally, an inherent weakness of this qualitative study is that the findings may not be representative of the population in the study area and may not be applicable to other settings. Because study participants were recruited from an urban setting with comparatively wide-spread access to a variety of HCT services, our findings may not be as relevant in rural areas.