Behavioural responses to Covid-19 health certification: a rapid review

Background Covid-status certification – certificates for those who test negative for the SARS-CoV-2 virus, test positive for antibodies, or who have been vaccinated against SARS-CoV-2 – has been proposed to enable safer access to a range of activities. Realising these benefits will depend in part upon the behavioural and social impacts of certification. The aim of this rapid review was to describe public attitudes towards certification, and its possible impact on uptake of testing and vaccination, protective behaviours, and crime. Method A search was undertaken in peer-reviewed databases, pre-print databases, and the grey literature, from 2000 to December 2020. Studies were included if they measured attitudes towards or behavioural consequences of health certificates based on one of three indices of Covid-19 status: test-negative result for current infectiousness, test-positive for antibodies conferring natural immunity, or vaccination(s) conferring immunity. Results Thirty-three papers met the inclusion criteria, only three of which were rated as low risk of bias. Public attitudes were generally favourable towards the use of immunity certificates for international travel, but unfavourable towards their use for access to work and other activities. A significant minority was strongly opposed to the use of certificates of immunity for any purpose. The limited evidence suggested that intention to get vaccinated varied with the activity enabled by certification or vaccination (e.g., international travel). Where vaccination is seen as compulsory this could lead to unwillingness to accept a subsequent vaccination. There was some evidence that restricting access to settings and activities to those with antibody test certificates may lead to deliberate exposure to infection in a minority. Behaviours that reduce transmission may decrease upon health certificates based on any of the three indices of Covid-19 status, including physical distancing and handwashing. Conclusions The limited evidence suggests that health certification in relation to COVID-19 – outside of the context of international travel – has the potential for harm as well as benefit. Realising the benefits while minimising the harms will require real-time evaluations allowing modifications to maximise the potential contribution of certification to enable safer access to a range of activities. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-021-11166-0.


Background
The current global pandemic caused by SARS-CoV-2 has resulted in wide ranging health, social and economic impacts, including many restrictions on daily movements, contacts, and activities. As testing and immunisation programmes are rolled out, one way of enabling increased access to a wide range of activities is certification of health status. This refers to the action or process of providing an official documenton paper, electronically or other approved mediumindicating that the holder is at low risk of acquiring or transmitting SARS-CoV-2. This could be due to a test-negative result for current infectiousness, a positive antibody test result conferring natural immunity, or vaccination(s) conferring immunity.
Health certification could have many benefits, through enabling greater and safer access to international travel, music, theatre and sports events, and to pubs, restaurants, hotels, and gyms. Allowing people to return to work, meet socially, and fulfil care obligations brings many social, emotional and economic benefits. Indeed, it might be considered unethical to restrict the movements of those who pose minimal risk to others [1,2]. Depending on how it is applied, health certification could also encourage vaccination uptake [3]. It also has the potential for harm. One concern from a behavioural perspective is that certification may foster an erroneous sense of no riskboth in those with and those without certificatesresulting in behaviours that increase risk of infection or transmission. In addition, immunity certification based on a test-positive result for antibodies could have a paradoxical effect on health protective behaviours whereby people deliberately seek infection in order to acquire a certificate [4][5][6]. Vaccination certificates could also increase opposition to vaccination in some groups [3]. Concerns have also been raised from ethical and legal perspectives. These include privacy [5], the removal of civil liberties [1,2], loss of social cohesion by the creation of a new hierarchy [1,6], discrimination against some social groups [4][5][6], and crime, including forgery, cheating, or obtaining documentation or data illicitly [4,5].
The use of health certificatesalso referred to as 'health passports' or 'vaccine passports'is not new. Printed health passes were used in Europe from the late fifteenth century to allow travel and trade while controlling the spread of plague [7]. They certified only that the bearer had come from a city that was free from plague [8]. The Vaccination Act of 1853 made smallpox vaccination compulsory in Britain for infants. Parents were given a blank certificate of vaccination when registering their child's birth, to be returned, signed, within three months. Failure to do so resulted in fines and imprisonment [9].
In relation to the current Covid-19 pandemic, certification has been used in China in the form of QR codes allowing entry into public spaces and a range of settings including workplaces, public transport, schools, airports, restaurants and grocery stores [10]. These codes amass data including exposure to places and people at higher risk of transmission. Certification was also used in Slovakia as part of population mass testing for infection. Those testing negative were given a paper certificate and released from strict curfew, thereby allowing return to all workplaces and visits to non-essential shops and restaurants [11,12]. In the UK, Covid-19 health certification is being planned or being used in limited number of areas, including visits to care homes [13,14], attendance at football games [15], and some music venues [16]. Israel has been operating a 'green pass' scheme in the form of an app which shows whether people have been fully inoculated or have already had the virus [3,17]. This allows access to gyms, hotels, theatres, and concerts. At the time of writing, similar schemes have been introduced in New York and in Denmark.
The main area where certification (for antigen testing) has been in active use is international travel. The EU has recently announced a 'digital green certificate' scheme, enabling those vaccinated, having a recent negative antigen test, or recovered from Covid-19 to travel freely and without quarantine between states within the bloc [18]. The International Air Transport Association has also been developing a digital health pass to "manage and verify the secure flow of necessary testing or vaccine information among governments, airlines, laboratories and travellers" [19]. A number of airlines are using digital health passports, mostly on a trial basis, including British Airways, Virgin Atlantic, and American Airlines [20].
Realising the benefits of health certification in the case of Covid-19 will depend in part upon understanding the possible behavioural and social impacts as a basis for designing systems that mitigate their potential harms. This paper describes the results of a rapid review to examine evidence for such impacts in four areas: (1) public acceptability, (2) effects on uptake of tests and vaccination, (3) impact on behaviours that affect transmission, and (4) crime.

Methods
A rapid review of the literature was undertaken in accordance with PRISMA criteria for systematic reviews [21] to identify the potential impact of enabling access to activities through certificating for one of three outcomes in relation to covid-19 status: (a) negative test results for the virus; (b) positive results on a test indicating immunity; (c) vaccination against Covid-19.

Search strategy
The search strategy was applied to four peer-reviewed databases --Web of Science (Core Collection, BIOSIS Citation Index, BIOSIS Previews, KCI-Korean Journal Database, Medline, Russian Science Citation Index, SciELO Citation Index), Ovid (Journals@Ovid, Global Health), Scopus, and APA PsycINFO --and four preprint databases --SocArXiv, MedRXiv, PsyRXiv, and SSRN. These databases were selected based on their coverage of public health topics.
For the grey literature, a search was conducted through the websites of public polling companies such as YouGov and Ipsos MORI; websites detailing public, private and third-sector research projects into Covid-19; and academic websites. Many of these websites were initially identified through a web search using Google Advanced. References and forward citations of relevant articles were also searched.

Inclusion and exclusion criteria
The following inclusion criteria were used: Given the relative paucity of evidence, we took a liberal approach to the inclusion criteria, which allowed us to add a small number of studies judged to be relevant that were not identified in the search (e.g., a study on the phrasing of test results).

Risk of bias
Risk of bias was measured using the Mixed Methods A p p r a i s a l T o o l ( M M A T h t t p : / / mixedmethodsappraisaltoolpublic.pbworks.com; see [23]) evaluating studies on five dimensions based on the study method. Studies were rated as good quality if they scored four or more out of five; moderate quality if they scored three out of five; and poor quality if they scored two or less out of five. See Supplementary Information for details: https://osf.io/357kt/?view_only=475cd0776a2 74e6bbc74f95e1eecd0e0

Search results
The search of peer-reviewed databases identified 6292 citations; searches of pre-print databases identified a further 18 citations. Of these, 1133 were duplicates and were removed, with 5178 citations remaining. A search of the grey literature identified 25 additional citations. After title, abstract and full-text screening of all citations, 33 were judged to meet the eligibility criteria. Additionally, 1 article was identified through backward referencing (see Fig. 1).
32 of the 33 studies used quantitative methods, with one study using qualitative (narrative) methods. Of the quantitative studies, the majority were cross-sectional surveys (n = 29), with the remaining being experimental in design (n = 3). Studies were conducted in a variety of countries: Germany (9), UK (10), US (3), Australia (2), Canada (1), Nigeria (1), Poland (1), Romania (1), Spain (1) and Switzerland (1). Three studies drew large samples from several countries (15, 19 and 11 respectively). Of the types of intervention that were the focus of 1 A recent systematic review of mandatory vaccination for children recently summarized findings as follows: 'Quantitative studies found little evidence for any factors being consistently associated with support for mandatory vaccination. Qualitative studies found that parents perceived mandatory vaccination schemes as an infringement of their rights and that they preferred universal, compared to targeted, schemes' [22] studies, 2 concerned test-negative result for current infectiousness, 14 concerned test-positive for antibodies conferring natural immunity, and 17 concerned vaccination(s) conferring immunity. The majority of studies related to Covid-19, with one concerning yellow fever, and one other concerning flu vaccinations.

Risk of Bias analysis
Using the MMAT, the mean average risk of bias score was 1.5 from a maximum of 5 (where a higher score means lower risk of bias). In many cases authors did not describe studies in sufficient detail for an evaluation to be made (see Supplementary Information: https://osf.io/357kt/ ?view_only=475cd0776a274e6bbc74f95e1eecd0e0). Based on the available information, 15 of the studies were rated as low quality, 14 as medium, and three as high.

Overview
We present a narrative analysis of the results on the impacts of certification in four areas: (1) public acceptability; (2) effects on uptake of tests and vaccination; (3) impact on behaviours that affect transmission and (4) crime. All results are summarized in Table 1.
Various terms were used to refer to health certification documents, including 'certificates', 'passes' and 'passports', referring to infection, virus, antibodies, immunity and vaccination. The terms used in this section are infection certification (based on test-negative results for infection, whether lateral flow test or qPCR) and immunity certification (based on either a test-positive result for antibodies or a completed COVID-19 vaccination).

Public acceptability
Ten studies of public opinion regarding health certification were found. Some asked about access to particular activities while others simply asked about the use of health certification in principle. In addition, eight studies examined attitudes towards mandatory vaccination.   Immunity certificate •22% of respondents would "probably" or "definitely" seek infection if earning immunity gave access to various opportunities: 14% to go to gatherings greater than 25 people, 13% to visit eldercare facilities, 12% to visit foreign countries, 10% to visit hospital patients, 11% to maintain or access employment at an eldercare facility.
•Younger age was significantly positively associated with willingness to seek infection.
•29% of gig workers reported they would seek selfinfection to maintain or access employment in eldercare.•51% of respondents "strongly" or "somewhat" agree that eldercare facilities should be allowed to require immunity certificate from employees.  Immunity certificate • The majority of respondents did not object to the idea of immunity passports, with over 60% of respondents supporting the idea to varying extents. • Over 60% of respondents wanted an immunity passport for themselves. • Around 20% of respondents considered immunity passports to be unfair and opposed them completely. • 79% of respondents would not consider at all deliberate self-infection to obtain an immunity passport, around 21% considered doing so to varying degrees. • Increased age, greater perceived risk of the disease, greater trust in government were positively associated with acceptance of immunity passports whereas gender had no effect. Immunity certificate • 60% of participants reported that immunity certificates should be offered to the general population. • The contexts where certificates would be perceived as most useful were taking a plane (73%) and entering a country (72%); fewer participants agreed with them being useful for participating in large gatherings (55%) or the right to work (32%). • 55% of participants thought a vaccination should be mandatory and 49% thought a vaccination certificate should be mandatory. • 68% felt there was a potential risk of discrimination.

Infection certification
One study surveyed plane passengers (n = 4700) from 11 countries in August 2020. 84% were in favour of infection certification for air travel [24].

Immunity certification from antibody testing
Four surveys carried out in Germany in May 2020 (ns between 925 and 1014) found that between 45 and 49% disagreed with the introduction of an "immunity pass", with around 26% agreeing [25][26][27][28]. Two surveys carried out in Australia in April and May 2020 (ns = 1169 and 449) found that~11% did not support immunity 'passports' or 'certificates' at all but~75% supported them slightly to fully [29,30].
Other studies asked about attitudes to immunity certificates for different purposes. Across five studies (n~1000 to~1700) conducted in four countries between April and December 2020, a majority of participants (54 to 73%) were in favour of the use of immunity certificates, particularly in the context of international travel [31][32][33][34][35]; a minority (15-20%) strongly opposed their use. One study (n~1000) conducted in Germany in May 2020 found the opposite, with more people opposed to than supporting "immunity cards" [36]. A UK survey carried out in December 2020 (n = 1706) reported that while 44% of respondents found vaccination certification acceptable for going to the cinema, this fell to 39% for going to a restaurant [37]. In another UK survey carried out in December 2020 (n = 5396), 22% of respondents said that those who have been vaccinated should not be subject to any more coronavirus restrictions while 68% disagreed [38]. The percentage in favour of immunity certificates for use for the right to work was much lower than in the case of travel. Across three studies in three countries carried out in April -September (n ranging from 1000 to 1500) support ranged from 20 to 51% [31,33,39].
There was little information in most studies on how any of the attitudes described above varied across social groups. In the UK, one study found that acceptance increased with age, greater trust in government, and higher perceived risk of COVID-19 [31].

Immunity certification from vaccination
Only one study of attitudes towards vaccination certificates specifically (n = 4311) was retrieved, conducted in the UK in November 2020, which assessed attitudes towards their use on international flights. 72% supported their use (42% strongly) and 11% strongly opposed them [40]. Support was strongest in older age groups, and unrelated to gender or socioeconomic status.

Mandatory vaccination
The terms 'mandatory' and 'compulsory' vaccination were used in studies to refer to a general requirement by governments for all citizens to be vaccinated, but with the means by which this could be achieved usually left unspecified. A UK survey published in November 2020 (n = 2000) found that 45% of respondents thought the Covid-19 vaccine should be mandatory for everyone, with 35% disagreed entirely [41]. Of those who did not want to be vaccinated, 19% said they would do so if they could go to the pub, 35% if they could go on holiday abroad, and 28% if they could go to sporting, music or other events. A UK survey carried out in December 2020 (n = 5351) also found that 37% supported compulsory vaccination [37]. A survey carried out in Germany in June and July 2020 (n = 851) found that 51% of respondents were against and 49% in favour of mandatory vaccination. The approval rate was significantly higher among those who would get vaccinated voluntarily (59%) than those who would not (27%) [42]. An American survey carried out in September 2020 (n = 2730) found that acceptance of mandatory vaccination was positively associated with non-Black respondents and those with a bachelor's degree [43]. An international survey (15 countries) carried out in August found that support for mandatory vaccinations was generally strongest in Brazil, Mexico, and India [44]. A survey in Canada (n = 1000) found that support for mandatory vaccinations fell from 72% in July to 61% in October 2020 [45]. Similarly, a survey in Germany (n = 1169) found that support for mandatory vaccination declined from 73% in April 2020 to 36% in December of the same year [46].

Uptake of tests and vaccination
Few studies addressed the possible impact of certification on uptake of vaccines or tests. A number suggested that intention to get vaccinated would vary with both the activity enabled by this and the source recommending vaccination.

Infection certification
No studies were found.

Immunity certification from antibody testing
An online experiment carried out in the UK in April 2020 (n = 1204) found that 85% would definitely (56%) or probably (29%) have an antibody test if offered [47].

Immunity certification from vaccination
One US study (n~1000) conducted in September 2020 assessed 'vaccine rules that would resonate' [48]. The activities requiring vaccination certification for which most people said they would get a Covid-19 vaccination were: visit a hospital or nursing home (likely uptake rate of 70%), travel to another state (70%), air travel (68%), work (60%), attending large non-religious gatherings (59%), attending large religious gatherings (55%), and attending school (51%). However, a Polish study carried out in June 2020 (n = 1066) [49] found that of those who did not plan to get vaccinated, 51% were not swayed by any reasons. Indirect evidence that certification of vaccination for access to work could reduce uptake of vaccination is provided in a survey of 13,426 adults in 19 countries carried out in June 2020. A baseline of 71% reported that they would be very or somewhat likely to take a COVID-19 vaccine, compared with 61% if the vaccine was recommended by an employer [50]. However, an American survey carried out in September 2020 (n = 2730) found that slightly more respondents found employer-enforced employee mandates acceptable (47.7%) than unacceptable (38.1%) [43]. Those reporting higher levels of trust in information from government sources were more likely to accept a vaccine and take their employer's advice to do so [50].

Mandatory vaccination
Two studies with experimental designs carried out in Germany (ns = 993 and 297) found that if a vaccination were to be presented as compulsory this led to anger (compared to voluntary vaccination) which then had a negative effect on willingness to accept a subsequent vaccine [36,51]. A UK survey carried out in November 2020 (n = 2090) found that, for mandatory vaccination, the numbers saying they would or would not get a vaccination did not vary depending on legal penalty (65% tõ 25% in each case) [52].

Impact on behaviours that reduce transmission
The evidence for possible behavioural outcomes of certification is summarised below first, amongst those with a certificate, and second, amongst those without a certificate.

Those with a certificate
Infection certification An online experiment (n = 4765) conducted in November 2020 in a UK sample found that intentions to fully follow guidance were 61% for those receiving a negative test result but 56% for those receiving a certificate alongside their negative test result [53]. For those not asked to imagine they had undergone testing, 63% reported fully following guidance.
Immunity certification from antibody testing Another UK online experiment (April 2020, n = 1204) assessed the impact of describing a positive test indicating presence of antibodies on risk perception and protective behaviours [47]. Using the term 'immunity' as opposed to 'antibody' increased the proportion who erroneously perceived they would have no risk of catching coronavirus in the future given an antibody-positive test result, from 9.8 to 19.1%. Perceiving no risk of infection with coronavirus given an antibody-positive test result was associated with an intention to wash hands less frequently.
Immunity certification from vaccination A UK survey carried out in December 2020 (n = 1706) found that 50% of respondents said they would continue to follow coronavirus rules and restrictions just as strictly after having a vaccination; 29% less strictly; and 11% not at all [37].

Those without a certificate
Having failed an immunity test The majority of participants in a Swiss survey said they expected that tests showing an absence of antibodies would encourage people to take more precautionary measures such as wearing of face coverings (76%) and respect for social distance measures (87%) [33].
Having not applied for a test Six studies in four different countries conducted between April and June 2020 (n > 1000 each) reported between 19% [39,54] and 31% [29][30][31][32] of participants saying that they would likely expose themselves to infection in order to get a certificate.
More students compared to other groups reported that they might deliberately infect themselves (58%) [31]. In another study, those who were younger and those who worked in the "gig" economy (29%) were more likely than others to report that they would seek self-infection to maintain or access employment [39]. However, a survey study carried out in Germany in May (n = 1007) found that no respondents reported they would intentionally get infected in order to receive an 'immunity pass' (though no data was shown to confirm this) [25]. A further study (in Switzerland) examined expectations of others' behaviour and found that 28.6% thought that others might self-infect (respondents were not asked how they themselves might respond) [33].

Crime
One report [55] described the use of counterfeit certificates for yellow fever. In December 2018, Nigeria and other countries introduced machine-readable yellow fever cards, but cards could still be obtained without evidence of vaccination. More outbreaks were predicted as people continue to carry fake vaccination certificates throughout Africa.

Discussion
In response to the Covid-19 pandemic, health certification is being used or considered for use to enable increased access to a wide range of activities for leisure, work and travel while minimising risk of transmission of the virus. Public attitudes were generally favourable towards the use of immunity certificates (based on vaccination or on antibody tests) for international travel, protecting the vulnerable (e.g., in a care home setting), but generally unfavourable towards their use for access to work, educational or religious activities or settings. A significant minority was strongly opposed to certificates of immunity -whether based on antibodies or on vaccination -for any purpose. A minority supported mandatory vaccination. A number of studies suggested that intention to get vaccinated varied with the activity enabled by certification or vaccination (e.g., international travel). There was no evidence in the review that mandatory vaccination including sanction would increase uptake. Some studies suggested that health certification might reduce Covid protective behaviours, including social distancing and handwashing. Making access to settings and activities conditional on antibody test certification may lead to deliberate exposure to infection in a minority, especially among young adults and those in precarious employment. No studies were found suggesting effects of Covid-19 health certification on crime.
Both the quality and quantity of studies was low thus limiting the certainty of any conclusions. The potential benefits of Covid-19 health status certificatesthrough enabling greater and safer access to international travel and other activitiesneed to be considered in the context of their potential for harm. At the most general level the evidence reviewed suggests the potential for harms of certification but the nature and scale of these remains uncertain. Also uncertain is how any harms might most effectively be mitigated. The evidence reviewed on the potential impact of certification or mandates on vaccination rates suggests this would not increase vaccination rates and might even reduce them. Mandating vaccinations through various means to reduce or eliminate choice is controversial and much debated particularly in the context of childhood vaccination programmes. While effective in some contexts, other approaches to increasing uptake in children can be as or more effective [56,57]. Amongst adults, a recent review of vaccination policies found that in 17 of 42 European countries some form of mandate or regulation was used [58].
The limited evidence reviewed here that health certification might reduce Covid-19 protective behaviours is consistent with concerns expressed by WHO that those who believed they had had COVID-19 would reduce their adherence to protective behaviours [59]. It is also consistent with more recent research on behavioural responses to rapid antigen tests and vaccinations against Covid-19. A study of rapid antigen tests in the UK found that around 13% of those receiving a test-negative result reported increasing their interactions with others [60]. Around 40% of those aged over 80 in England reported breaking Covid-19 restrictions in place at the time after receiving their vaccinations [61]. In Israel, the rapid vaccination of much of the adult population was accompanied by a short term rise in Covid-19 infections [62]. These findings are consistent with those vaccinated or certificated as having had the virus reducing their adherence to protective behaviours [63,64]. Group processes have the potential to amplify these behavioural effects. When those with certificates reduce their protective behaviours, such changes can be seen as normative, leading others in their ingroupincluding those without certificatesto do the same [65][66][67].
Regardless of the basis for any Covid-19 status certificate issued, certification will indicate that the holder has been deemed to pose a lower risk of infection and perhaps transmission of the virus than those without a certificate. At the time of writing, evidence suggests that protection from further infection having had the virus is very high for~7 months [68,69], and that current vaccines can cut household transmission of known variants by half [70]. Vaccination and prior infection therefore mean risks may indeed be significantly lower. But a residual risk will remain until the virus is eliminated.

Maximising benefits and minimising harms
Health certification could enable greater and safer access to a wider range of activities and locations for many people. To realise these benefits while minimising the harms, health certification schemes should be implemented with an evaluation designed in from the outset, and, in keeping with the principles of open science, to include the use of pre-registered protocols. Such schemes should also be designed within a transparent ethical and legal framework to protect privacy, equity and minimise fraud.
Evidence from both testing and vaccination suggests that increased inequalities would be a possible harm of health certification. Participation in NHS Test & Trace is lower in marginalised groups [71,72] and in areas of high deprivation [73]. The Liverpool mass testing pilot found that uptake in the most deprived areas (16.8%) was half that in the least deprived areas (33.4%) [74]. Data from the UK and other countries suggest that those with lower incomes or education and from minority ethnic groups have lower intentions to undergo COVID-19 vaccination than others [75][76][77]. In part these differences in testing and vaccination uptake reflect higher mistrust in government amongst marginalised communities [71,[78][79][80]. Stigmatisation, discrimination and racism might also reduce migrants' and ethnic minority communities' willingness to come forward [71]. In addition, certification will likely be most readily available as a digital record, which has the potential to exclude those without access to electronic platforms [72]. In summary, disadvantaged groups are underrepresented in those getting tested and vaccinated and would therefore be disproportionately excluded in any Covid certification scheme.
Some suggest that antibody certification could provide advantages to those who are otherwise most disadvantaged by greater exposure to Covid (due to work and living conditions) by providing safer workplaces [81]. The alternative to this individual-focused approach is Covidsafe certified workplaces, with legislation properly enforced in the same way that food outlets are legally required to be checked for safety [82].
Use of the social rewards associated with health certification to encourage take-up of the Covid-19 vaccine [3,83] might work well with some groups but could backfire with those who are already mistrustful of the authorities. While the authorities in Israel see an incentive-based approach as an alternative to coercion, the scheme has already led to conflict at workplaces [3,17]. The issue of enforced exclusion of many people from significant areas of social life raises broad issues of justice and fairness and could mobilize a wide constituency. In the nineteenth century, resistance to the Vaccination Act included violent protests from the working class [9] which contributed to a change in the law allowing exemptions on the basis of conscience [84].
Minimizing the potential harms of certification will require the following. First, there should be equality and equity of access to tests, vaccinations and certificates. Second, there needs to be clear and open communication that is accessible to different communities of the meaning of any results and certificate, the residual risks of infection and transmission, and the implications for individual behaviour. National and local leaders, including community members and community organisations, should be involved in this communication campaign, in line with engagement and public inclusion principles [85,86]. Finally, practical steps are needed to ensure that no group should lose access to an everyday activity or setting by requiring certification, particularly if income, health or education will be impacted.

Strengths and limitations of the review
This review included 33 studies pertinent to understanding the possible effects of health certification on public behaviour. To the authors' knowledge it provides the first overview of studies in this area, with implications for practice and policy.
The review was limited both in scope and quality of studies retrieved. The focus was upon the behaviour of general populations and not upon the behaviour of other relevant actors such as employers or those managing or organising venues and events, entry to which may be dependent upon health certificates. The behaviour of these other actors will also be important in realising benefits of health certification to ensure, for example, that measures designed to reduce transmission at a venuesuch as physical distancingare seen as additional and not substitutes for entrants having a health certificate [87].
Only four of the studies included in this view were judged to be high quality. The main reasons for being judged low quality were that it was unclear whether there was a non-response bias; like many surveys conducted during the pandemic, most of the studies featured in this rapid review relied heavily on convenience samples which were not representative [88]. Only three of the studies were peer reviewed at the time of this rapid review. While three were available on pre-print servers, most were unlikely to be published in peer reviewed journals and were often released as public opinion surveys. However, it is worth noting that, where they addressed similar questions, findings from the highquality studies [31,33] were similar to those from studies judged to be medium [32,34,35] or low [39] quality (i.e., of greater risk of bias).
All the studies concerning Covid-19 studies relied on self-report measures of behaviour and in response to hypothetical scenarios. This was because these studies were carried out before the introduction of certification.
Most of the studies were from high income countries. Most of the studies did not take process or demographic measures. This restricts what we can conclude about the underlying reasoning behind attitudes such as opposition to covid health certification (e.g., whether privacy concerns vs inequality implications were more important).
Finally, public attitudes in 2020, when certification schemes were not widely discussed or implemented and populations had less experiences of living with restrictions due to higher prevalence of the virus, are very likely to change over 2021 where such schemes are introduced or actively considered as an approach to controlling transmission of the virus. For example, a representative poll carried out in the UK in March 2021 [89] found higher levels of support for vaccine passports for a variety of activities than was found in the 2020 surveys in the present rapid review. And in the USA, studies published in March and April 2021 found divided views and levels of support for types of immunity certification ranging from 17 to 80% [90][91][92].
Mindful of these limitations, this review nonetheless provides a starting point for anticipating the potential harms of health certification as a basis for mitigating these to realise the benefits with minimal harms.