Willingness to Offer HIV Self-testing Among Health Care Providers From Specialized HIV Care Services in the Northeast of Brazil

Background: The lack of knowledge of the serological status of people living with HIV is still a concern in Brazil. HIV self-testing (HIVST) has proved to have great potential for expanding testing, especially among more vulnerable populations. The large-scale distribution of HIVST by the Brazilian public health system has been ramped up in recent years. Methods: This study investigates the awareness, acceptability, and factors associated with willingness to offer HIV self-testing (HIVST) among health care providers (HCP) in northeastern Brazil. Cross-sectional study with HCP in 29 specialized care services (SCS) from 21 cities from state of Bahia. The inclusion criterion was that HCP had been working at least six months in the service. Sociodemographic, occupational, and behavioral data were collected using a questionnaire. Descriptive, bivariate and multivariate analysis, estimating the respective adjusted odds ratios (aOR) and 95% condence intervals (95%CI) by logistic regression, were performed. Results: The awareness and acceptability of HIVST, and willingness to offer it were 79.8%, 55.2%, and 47.1%, respectively. Few of the professionals reported that the SCS where they worked dispensed HIVST (3.6%), and 13.5% said they had received some kind of information or training on HIVST. Factors associated with willingness to offer HIVST were: HIVST acceptability (aOR = 9.45; 95% CI: 4.53 – 19.71), agreement to use it oneself (aOR = 4.45; 95% CI: 1.62 – 12.24), condence in offering HIVST to service users (aOR = 5.73; 95% CI: 2.26 – 12.72), and considering the general public eligible for HIVST (aOR = 2.88; 95% CI: 1.25–6.59). Conclusions: Although HIVST awareness among the HCP was high, its acceptability and their willingness to offer it was moderate. More training in HIVST is needed among HCP in Brazil.

Porto Seguro, Salvador (with ve services), Senhor do Bon m, Simões Filho, Teixeira de Freitas, and Vitória da Conquista. All the health workers who were available at the services during the data collection period were invited to take part in the study and selected by convenience sampling. Two hundred and fty-two health workers at the aforementioned services were recruited from the 490 registered in the National Register of Health Establishments (Cadastro Nacional de Estabelecimentos de Saúde) in March 2020, representing 51.4% of the total population.
The project was assessed and approved by the Research Ethics Committee of the Multidisciplinary Health Institute of the Federal University of Bahia (#3,523,832/2019). Signed informed consent was sought and obtained from all the participants.

Data collection and instruments
Data were collected between October 2019 and March 2020. A structured questionnaire was developed for this study and administered using in private rooms by interviewers who had been trained in how to conduct the interview. The questionnaire was piloted in one of the cities.

Study variables
The study's outcome variable was willingness to offer HIVST, which was structured according to the following question on the questionnaire: "Would you offer self-testing for the service users?" (no, yes).
Exposure variables that could explain willingness to offer HIVST were selected on the basis of a literature review. The sociodemographic variables were: sex (male, female), age (≤ 35, 35 to 50, ≥ 50), and education (high school graduate, college graduate, graduate diploma or higher). The training and occupation variables were: job (nurse, nursing assistant, physician, etc.), specialized in HIV/AIDS (no, yes), years of training (≤ 5, 5 to 10, > 10), length of time working at SCS (1 to 7 years, 8 to 13 years, 14 to 40 years), and type of employment contract (temporary, permanent). The HIVST-related variables were: prior knowledge of HIVST (no, yes), acceptability of HIVST (no, yes), willing to offer HIVST (no, yes), gives service users information on HIVST (no, yes), sources of information (training, media, coworkers, etc.), reasons for not offering HIVST (suicide risk, self-harm, incorrect use, etc.), prior training in HIVST (no, yes), HIVST is dispensed at the service where they work (no, yes), knows that HIVST is available via SUS (no, yes), knows HIVST is available at pharmacies (no, yes), trusts the HIVST diagnosis (no, yes), would use HIVST him/herself (no, yes), populations they believe HIVST should be distributed to (general public, key populations, etc.), sure about offering HIVST to the service users (unsure, quite sure, sure/very sure), believes HIVST results in risk compensation (no, yes), believes HIVST causes a reduction in high-risk sexual behaviors (no, yes), preferred form of dispensing HIVST (assisted testing, self-testing, as the user wishes), and other resources that should be offered together with HIVST (counselling, prevention materials, etc.).

Data analysis
A descriptive analysis was conducted of the knowledge, acceptability, and willingness to offer HIVST on the part of the health workers, and the supply of HIVST information to service users, as well as the other study variables, estimating the proportions with 95% con dence interval (95%CI). Next, the potential factors associated with willingness to offer HIVST were investigated using bivariate and multivariate analysis, estimating the respective odds ratios and 95%CI by logistic regression. The variables with p < 0.20 in the bivariate analysis, using the chi-squared test, were selected to include in the multivariate analysis. The Hosmer-Lemeshow test (p > 0.16) was used to assess the models' goodness of t, and the ROC curve estimation (0,88) was used to choose the nal model and the theoretical relevance of the variables. All data were analyzed using STATA, version 14.0.

Results
Two hundred and fty-two health workers from SCS were interviewed, 78.2% of whom were female, 18.3% were up to 35 years old, 54.4% were 35-50 years old, and 27.4% were over 50 years old. Most (84.5%) of them had a college degree, around half (51%) had a graduate level diploma, 7.2% had a master's degree, and 2.4% had a doctorate. As for their professions, 25.8% were nurses, 12.3% were pharmacists, 11.9% were nursing assistants, 11.9% were physicians, and 9.5% were social workers. Overall, 23.4% said they had specialized training in HIV/AIDS and over half had trained in the area over ten years previously (74.2%). As for their experience, 78,6% stated they had been working at the SCS less than 14 years, and 66.3% had a permanent employment contract (Table 1).  Table 2). Few of the professionals reported that the SCS where they worked dispensed HIVST (3.6%), and 13.5% said they had received some kind of information or training on HIVST. Just 23% of the professionals were aware that HIVST was available via SUS, but 61.9% knew it could be purchased at a pharmacy. Just over half (55.6%) of the respondents expressed con dence in the diagnosis given by the HIVST, and 75.4% would use it themselves. As for how con dent they felt about offering HIVST, 38.1% said they felt uncon dent, 28.6% felt quite con dent, and 33.3% felt uncon dent or very uncon dent. Almost half (42,9%) believed that HIVST could result in risk compensation, while 28.7% said they believed access to it could reduce high-risk sexual behaviors. Over half (56.8%) felt that assisted testing at the health facility was the best option ( As for the resources that should be dispensed together with HIVST, a positive association was found with whatever materials the user prefers (OR = 3.5; 95%CI: 1.71-7.2) ( Table 3).  In a systematic review among key populations, eight of the fourteen studies found high acceptability (≥ 67%), ve found moderate acceptability (34-66%), and one found low acceptability (≤ 33%) of HIVST [17]. In our study, we found the acceptability of HIVST among the health workers to be moderate, which differs from the reports of high acceptability found in systematic reviews of speci c populations and age groups in different parts of the world [17,21,27]. Also in Brazil, a study carried out in 2016 with MSM in 12 cities found a similar level of acceptability of HIVST (47.3%), which was even lower among the MSM who had never done an HIV test (42.7%) [28]. Although this study did not involve health workers, the coincidence of the levels may be indicative of the incipient nature of self-testing in the country.
In our study, assisted testing at the health facility was the preferred means of administering HIVST, which could indicate the health workers interviewed could still be operating from the traditional perspective of voluntary counselling and testing (VCT), which assumes users voluntarily seek out services. This approach is different from provider-initiated counselling and testing (PITC), which is marked by the routine proactive offer of testing by health providers at all consultations. Both assume a particular posture on the part of the health worker in presenting the user the opportunity to get an HIV test [29,30].
Considering the Brazilian MoH current proposal to expand HIV testing using HIVST and promote user autonomy, the traditional approach of VCT could constitute an obstacle to the test kit's distribution, while approaches more akin to PITC could help generate demand for testing.
In this study, fewer than half of the health workers demonstrated willingness to offer HIVST. The main reasons for this relate to potentially negative mental health outcomes, such as suicide risk, self-harm, and harm to others in response to a positive test result. Studies with key populations such as MSM, transgender people, and sex workers in different countries have found similar misgivings about HIVST use, but to date there is no evidence in the literature that HIV self-testing is associated with such outcomes [31,32].
As for the practice of counselling, the traditional format still seems to prevail among health workers, since the absence of post-test counselling for negative test results was also considered a reason for not offering HIVST. In a systematic review of qualitative data collected between 1998 and 2018, Njau et al. [32] found a similar concern among health workers and administrators in ve African countries. A qualitative study conducted in a STI/AIDS counselling and testing center in Maceió, Brazil, in 2017 found that some health workers felt HIVST could diminish their professional role and could threaten their work, potentially hampering the continuity of care dispensed to users [33]. The data about the lack of availability of counselling suggests both a concern on the part of health workers unwilling to offer HIVST about users' care, while also reinforcing the space of power occupied by these health workers in the userprofessional relationship, leaving little space for lay subjects to express their autonomy.
The denial of HIVST because of lack of counselling could also reduce the capacity to diagnose HIV infection among stigmatized key populations who face multiple barriers to accessing care. Previous experience shows that unconventional testing and counselling formats, such as approaches using online technologies, are related to enhanced HIVST acceptability or an enhanced experience of self-testing by users [34][35][36]. However, HIVST use should not be turned into a free-for-all. It should be used according to the latest scienti c evidence and the legal provisions of SUS, which include access to online post-testing guidance, a toll-free 24-hour telephone hotline for any HIVST registered in Brazil [42], and explicit user support guidelines provided by the MoH, such as con rmation of diagnosis and assurance of linkage to health services [14].
It is important to be aware that social changes should be taken into account and novel ways of providing information and counselling could be needed, which are still a challenge to the widespread introduction of self-testing [32,43,44]. Quicker approaches mediated by easy-to-access digital technologies, or options that may not even require real-time user-professional interaction are potential routes forward, provided user access is enabled in such a way as to prevent the risk of excluding precisely those key populations among whom the epidemic is disproportionately high [40,45].
Prior knowledge of HIVST on the part of health workers was not associated to increased willingness to offer it, which could suggest that such willingness may be in uenced by other concerns vis-a-vis this testing option. Conversely, indicators that suggest greater familiarity with HIVST, such as working at a service where HIVST is already on offer, having received training in HIVST, and knowing about its distribution by SUS seem to be important in increasing the willingness to offer it, indicating the need for a broader repertoire of awareness-raising options.
Our study also found that willingness to offer HIVST was strongly associated with acceptability of HIVST. In a systematic review, Sekhon et al. [46] propose a de nition of acceptability that involves a subjective evaluation of a health intervention by both the individual who delivers the intervention and the individual who experiences it. These aspects are based not only on prior knowledge, but also on prior practical experience of the intervention. As such, the level of speci c training in a new health technology could be a critical factor in uencing its uptake [47,48]. It is therefore important for continuing education and training to be provided on HIVST at SCS.
In our study, the health workers who reported they would use HIVST themselves were four times more likely to offer it than those who said they would not use it. This could be a good indicator of con dence in the test and its use. Multi-center studies in Kenya in 2009 and 2010 [49] and in Ethiopia in 2012 [50] with health workers found HIVST training and experience using HIVST, respectively, to be associated with willingness to offer it.
Willingness to offer HIVST was also higher among those who felt very or quite con dent about offering HIVST than those who did not feel con dent about doing so. Around 68% of the health workers reported the potential failure of users to use the test or read the result correctly as a reason for not offering HIVST, which shows their concern at having laypersons perform a medical test. The same concerns have also been reported elsewhere [24,51,52]. Although errors in conducting the nger-prick test are among the concerns relating to HIVST [53], recent tests indicate that users and health workers may have the same performance when doing this kind of HIVST [54], and that the nger-prick test is more precise than a saliva test, insofar as it involves analyzing a blood sample [45].
Being con dent about HIVST could be impacted not just by the capacity of the test to give a reliable result, but also by concerns about the healthcare service being able to provide support for users who have a positive test result. For example, in a systematic review of factors that help and hamper the diffusion of HIVST, Musheke et al. [55] identi ed studies showing that lack of trust in health systems could be a barrier to HIV testing and treatment.
Additionally, our study results indicate that willingness to offer HIVST is higher among health workers who agree that the general public should be eligible to receive the test. This indicates that although the vulnerability of certain social groups and the importance of their having access to HIVST is recognized, the test could be well accepted among professionals for large-scale distribution in a comprehensive health system like the Brazilian one. However, the MoH [15] guidelines have focused on the provision of HIVST through strategies geared towards people using PrEP, people whose sexual partner is living with HIV, and key and priority populations for the HIV epidemic in Brazil [14].
The limitations of this study include its sampling process, which was not completely probabilistic, and the fact that willingness to offer HIVST was analyzed by health workers from SCS that both dispensed and did not dispense HIVST. As the outcome variable has to do with future willingness, those professionals who worked at services where HIVST was dispensed could have a higher proportion for this outcome. To overcome these limitations, the municipalities with SCS were put in a draw, considering the nine health districts in the state of Bahia, and just one of the centers at which HIVST was on offer was included in the study.

Conclusions
This study found a high proportion of prior knowledge of HIVST, but only a moderate proportion of acceptability and willingness to offer it. Furthermore, some information gaps about HIVST were identi ed, expressed as reasons for not offering it, which drew on distorted beliefs about what it is, including suicide risk, self-harm, or harm to others after using it, and the idea that HIVST use could be related to high-risk sexual behaviors. There was also some lack of con dence in the capacity of the health system to provide treatment for PLHIV and a tendency to perpetuate health practices that reinforce the power of the professional and limit individuals' autonomy in their choice of testing.
The large-scale implementation and distribution of HIVST, as is intended for Brazil, will depend on investment in the training of the health workers who interact directly with the public to be bene tted by this strategy. In order to be effective, concerns about this testing strategy will have to be challenged.
Qualitative studies could help develop a more in-depth understanding of the conceptions of professionals about HIVST and serve as a basis for interventions designed to adapt them in light of proven international experience.
Finally, it is believed that HIVST distribution should be accompanied by structural improvements designed to assure simple, clear information is available about how to use HIVST to ensure it is used correctly and the results are read and interpreted correctly, thereby facilitating access to treatment for PLHIV. Consent for publication "Not applicable" Availability of data and materials The data that support the ndings of this study is available in "dataverse.harvard.edu": https://dataverse.harvard.edu/privateurl.xhtml?token=bb9c73db-6801-41c1-965c-eb9621ce2daa Competing interests The authors declare that they have no competing interests.  Figure 1 Knowledge, acceptability, willingness to offer and supply information on HIVST for service users on the part of health workers at specialized HIV care services in Bahia, Brazil, 2019-2020.