Our survey shows that there is potential for the acceptability of SARS-CoV-2 self-testing in Brazil. In our inquiry, acceptability was conceptualized as a composite of the public’s values toward self-testing, including agreement with the concept of self-testing (73.91% and 60.09% of female and male respondents were in agreement, respectively); willingness to pay for self-testing (if available at an average price of 5.64 USD (SD 4.73) for the 69% of respondents who would pay for a self-test device); willingness to serially self-test (44.57% of females and 39.48% of males expressed willingness to perform weekly self-tests); and likelihood to use self-testing (12.95% and 36.21% of respondents were “very likely” or “likely”, respectively, to use a self-test). Although satisfactory, the rates of likelihood to use a self-test were not as high as those found in Indonesia [26], Nigeria [27], or Kenya [28], other countries where surveys of people’s values and attitudes towards self-testing were conducted in 2021 using the same methodology we used in São Paulo.
People’s attitudes toward the acceptability of self-diagnostics are context-dependent, and can be mediated by cost, design, accuracy, accessibility, and health authorities’ endorsement of self-testing for infectious diseases, among other factors. Our survey was designed to assess what the predictors of acceptability might be in a variety of countries [25]. As per our regression analyses, individuals with a secondary education or who are working full-time for an employer might have higher odds of being likely to using a self-test when needed, while those living in a household with people at increased risk of severe COVID-19 disease might have lower odds of being likely to using a self-test. The regression analyses also suggested that individuals aged ≥ 56 years and those self-employed part-time may have lower odds of paying for a self-test device, while individuals with higher education, those who are working full-time, those who perceive themselves to be at high-risk of COVID-19, and those living in a household with individuals at increased risk of severe COVID-19 disease might have higher odds of paying for a self-test device. The predictors we detected might be helpful for those planning community- and primary healthcare-based testing services to map the profiles of the population groups who might be more attracted to using (or not) self-testing and, thereafter, to decide who should be targeted via the promotion of self-testing in São Paulo. Nevertheless, we also warrant caution in considering only the predictors detected by our analyses in future self-testing promotion planning. As suggested by HIV self-testing experiences [20, 21, 23, 29], the more that SARS-CoV-2 self-testing programs are diverse, inclusive, civil society-endorsed, and decentralized, the more likely it will be that such programs will efficiently meet the needs of different sectors of the public. To be responsive to these needs, programs must take full consideration of the intersectionality of populations’ barriers to accessing testing and care with their personal cultural, clinical, and socio-economic characteristics.
The drivers for the acceptability of self-testing that our survey explored are dependent on the studied population’s historical, sociocultural, and epidemiological context. Factors that may mediate the general public’s acceptability of self-testing in Brazil, such as access to facility-based SARS-CoV-2 testing and vaccination programs, and the epidemiological evolution of the pandemic, have undergone frequent changes since February 2020. These are factors that, for different persons and in different moments and geographies, might act as deterrents or as drivers of the acceptability and uptake of self-testing. The rapidly changing epidemiological scenario and public health authorities’ responses since the first SARS-CoV-2 infection was reported in Brazil are among the reasons why for our survey methodological approach we considered respondents’ characteristics, and not their context, as predictors of acceptability.
It must be noted that, regarding our respondents’ context, as of January 2022, after our survey had already ended, Brazil’s regulatory authorities (i.e., Anvisa) were beginning the process of accepting companies’ requests for approval for distribution of their SARS-CoV-2 self-tests [24, 30, 31]. In mid-February 2022, the Anvisa approved the first device (i.e., the CPMH® COVID-19 antigen self-test) for distribution [32]. In this context, depending on how self-testing is introduced and explained to the Brazilian general public and healthcare workforce, individuals’ values and preferences for access and usage of this case detection approach may be impacted. Indeed, the more user-friendly SUS health facilities are to self-testers, the more likely it will be that self-testers react favorably to a reactive result. The more transparent the government is in providing evidence that self-testing can decrease the incidence of COVID-19-related morbidity and mortalitu, the more likely it is that the public, and especially daily laborers, education center attendees, and those interested in traveling or in attending social gatherings, might want to self-test more frequently.
Of the self-testing acceptability studies that have been conducted, our survey findings are aligned with the results of studies conducted in Germany [9], Indonesia [26], Nigeria [27], Kenya [28], the United Kingdom [33], and Greece [34, 35] and Cyprus [34], where the study populations also manifested a willingness to use self-testing. Of these studies, only the inquiries in Indonesia [26], Kenya [28], Greece [34, 35] and Cyprus [34] targeted the general public. Comparing our survey with that of Goggolidou et al. [34] and Mouliou et at. [35], it should be noted that different contextual factors might have mediated the respondents’ favorable opinions exhibited toward self-testing in each study. In Greece, the health authorities had approved self-testing, distributed self-tests free of charge, and had made educational materials for end-users available via a government website [11]. These efforts could have promoted favorable public opinion toward self-testing in Greece. Nevertheless, it is worth noting that the survey carried out by Mouliou et al. [35] in mainland Greek reported that almost half of the total sample (n = 614) considered self-testing ‘dangerous’, and that only one in five respondents declared that they would buy a self-test. And, to our knowledge, this survey in Greece [35] is the only comparing attitudes towards self-testing in populations with both exposure to and experience of self-testing and facility-based testing for SARS-CoV-2.
It can be hypothesized that the public response to the Brazilian government’s behavior with regards to the COVID-19 pandemic in Brazil might have influenced people’s willingness to self-test, especially as our data collection was conducted prior to the surge of the Omicron variant of SARS-CoV-2, which dramatically increased the local (as much as the global) demand for rapid antigen-detection tests for SARS-CoV-2. In Brazil, the government’s challenges to providing mass screening and testing have been acknowledged [3, 4]. In mid-January 2022, a phone survey revealed that—in the midst of the Omicron variant wave—more than 8.1 million Brazilians had tried and failed to obtain a SARS-CoV-2 test [36]. These challenges to accessing testing have driven many Brazilians to resort to private healthcare. In a previous study of self-testing for HCV, informants reported that they would prefer to either self-test or go to a private practitioner for HCV testing rather than to go to an SUS facility [23]. Similarly, now that Anvisa is receiving requests from manufacturers for the registration of self-tests [32], many people may opt to purchase a SARS-CoV-2 self-test via a private provider rather than trying to access facility-based testing. Although accessing self-testing in the private healthcare sector may alleviate the burden in overstretched SUS facilities, attention needs be paid to private facilities’ capacities to inform their clients on the risk of false results when SARS-CoV-2 incidence rates drop in Brazil [37]. The individuals’ clinical status and their vaccination and exposure history to SARS-CoV-2, or the prevalence of SARS-CoV-2 infections in the community at the time of self-testing are other factors that can lead to false results and that must be considered in self-testing delivery models [37].
Future studies may provide a more thorough indication of what the reasons might be for the acceptability of self-testing in contexts where these devices have not previously been widely deployed. Future studies will also have to discern what attitudes are triggered by the intrinsic advantages of self-testing, and which attitudes are triggered by health system-related failures to cater for individuals at risk of SARS-CoV-2 acquisition and of severe COVID-19 disease. While our survey findings are optimistic (i.e., 88.49%, 97.60%, and 96.64% of respondents would communicate their result, self-isolate, and warn their contacts, respectively), in actuality, post-self-testing behaviors might be different if no social, labor or family support is provided. If social safety nets are not provided, self-isolation and reporting of a SARS-CoV-2 infection might be neither feasible nor desirable for affected people. As other self-testing studies in Indonesia and Nigeria have suggested, self-isolation might only be guaranteed if there are provisions in place to ensure that those who use a self-test do not lose their job or social position [26, 27].
Our survey findings have other implications for practice. Ideally, self-tests should cost less than 5.64 USD, to enable people to afford them. Further education on the risk of false-negatives must be provided, as 89.45% of our respondents expressed that they would stop self-isolating if they self-tested negative, even if they were symptomatic and had been in contact with a case. It is possible that an emphasis on frequent testing might be needed, to counterbalance the effects of the likely lower sensitivity of some SARS-CoV-2 self-testing devices (despite the Anvisa requirement that self-tests for distribution in Brazil must have at least 80% sensitivity and 97% specificity [30]), so that individuals who suspect they might have SARS-CoV-2 repeat a self-test every 24 h and monitor their symptoms before deciding to stop self-isolating. While self-testing may have an added value to identify asymptomatic SARS-CoV-2 carriers—and, while self-testing programs may be useful to educate the public on the possibility to transmit the virus to other persons even if the carriers are asymptomatic-, culturally-grounded and less value-laden communication materials will be necessary to sensitize self-testing users on the implications of being, and sharing spaces with, an asymptomatic carrier.
Considering our findings, it could be argued that facilitating access to self-testing may be a useful case detection approach to halting or slowing the transmission of SARS-CoV-2. Self-testing has the potential to reduce the burden on SUS facilities, which should be attending to those who are most seriously ill. It also has the potential to be scaled-up in educational, religious or working environments, where large numbers of individuals regularly congregate. Self-testing could also be useful in the hands of civil society-based grassroots organizations that can promote community-based testing for SARS-CoV-2 in poverty-stricken favelas or in areas where indigenous populations are in urgent need of improved access to testing.
It should be noted that this survey had some limitations. First, the findings might be representative of the inhabitants of São Paulo city but not representative of people who live in rural areas of the state. The intention was to conduct the study throughout the entire state, but for logistical reasons and because of restrictions on social movement due to COVID-19, this was not possible. Also, it must be noted that to avoid security incidents within São Paulo city, recruitment in some favelas took place in the areas’ main avenues. It is not possible to know whether the results may have varied slightly if the interiors of these neighborhoods had not been avoided. Despite these impediments, we managed to recruit a diverse sample, with a broad representation of self-expressed ethno-racial identities, education levels and employment statuses.