Introduction

The COVID-19 coronavirus (COVID-19) pandemic has arguably presented the most significant challenge to global public health in a century [1]. Non-pharmaceutical interventions, including social distancing measures, are an important means through which to control the transmission of such viruses. During the pandemic, in the UK, the social distancing measures that are the focus of this study included, but were not limited to: keeping physically separate (1–2 meters in the UK), only meeting with others in ways permitted under current legislation, and self-isolating when required to do so [2].

One of the key factors determining the effectiveness of such measures is the level of public adherence to them. As a pandemic draws on, is important to explore whether, and in what ways, the longer duration of measures may impact adherence and also the reasons for non-adherence amongst those who are not fully adhering to them. There has been much discussion in the media about whether ‘behavioural fatigue’ may potentially lead to increases in non-adherence [3]. However, this concept is controversial amongst academics, partly because and the term behavioural fatigue has been vaguely applied and ill-defined [4], and partly because of the lack of empirical evidence to support the fact that there is significant ‘behavioural fatigue’ around adherence [4]. In the UK, longitudinal public surveys generally note that adherence has been overall high over the course of the pandemic [5, 6]. However, researchers reporting on the survey data caution that there are high levels of self-reported confusion or a lack of understanding of guidelines [5]. Also, reported ‘complete adherence’ (where people report following all rules with no bending or even minor infringements) is considerably lower than reported ‘majority adherence’ (where people are following most but not all rules) [5]. Some rules (e.g. observing the 2-metre rule or visiting extended family when not permitted to do so) may be being flouted more than others (e.g. not self-isolating when advised to do so) [6]. However, a recent international study posited that ‘fatigue’ might explain an observed decline in adherence for ‘high-cost’ behaviours like social distancing (but not for ‘low-cost’ ones like mask-wearing). For high and middle income countries this decline was reversed later in the pandemic [7]. Further behavioural research is needed to add to our understanding of the potential existence and role of “fatigue” during the pandemic [8].

There is a rapidly growing body of research exploring factors associated with adherence and non-adherence to COVID-19 measures. Surveys of public behaviours suggests that some of the main factors contributing to high adherence include: high perception of risk of contracting COVID-19 [9–11], perceived social support [12], seeing others’ adhere [13], believing that you have already had COVID-19 [14], greater accessing of health-related information [15], trust in government [13] and political ideology [16]. Also, certain demographic factors have been found to predict high adherence, including: older age [5, 11] female gender [9], higher education level [9, 15]. Of course, adherence levels can vary across countries [15], due to a variety of reasons, including levels of pre-existing trust in authorities and the extent to which a country has a ‘tight’ or collective culture [17].

Much of this research on adherence to coronavirus measures has taken the form of quantitative surveys, which may give us an understanding of the extent of adherence and related factors, but often fall short on understanding the ‘why’ associated with such adherence. Qualitative research on experiences and perceptions of adherence can complement existing quantitative research by exploring in-depth some of the reasons behind instances of non-adherence, from the perspectives of the participants themselves, which can then be used to adapt policies to maximize impact. In this paper, we present results from an ongoing longitudinal qualitative research study looking at public experiences of, and attitudes to, COVID-19 measures put in place to reduce the spread of the virus at a population level [18, 19]. We focus on adherence to government mandates and regulation–particularly including social distancing measures such as keeping 1–2 metres apart from others, only meeting and mixing with others where and when permitted, but also including other mandated measures such as facemask requirements. In a previous paper, we explored experiences and perceptions of self-adherence and of adherence in others to social distancing measures [18]. We found that the majority of our participants reported high adherence to measures for themselves, but that reports of non-adherence in others were high [18].

In addition to the rapidly emerging empirical literature on COVID-19, we also drew on behavioural theory around rule-following and breaking, in particular Reason’s [20, 21] theory of human error. According to Reason [20], violations are deliberate deviations from safe practices, behaviours, standards or rules, whereas errors are the failed actions to achieve their desired goal. Errors include ‘slips’ and ‘lapses’ (failures of execution, e.g. through forgetting or misunderstanding information) and ‘mistakes’ (failures of intention, e.g. ineffective plans of action to achieve a desired goal) [20]. Additionally, Reason [22] distinguishes between two domains of failure: active and latent failure. Active failures refer to the unsafe acts (i.e. violations and errors) at the individual (person) level, whereas latent failures refer to the conditions under which such unsafe acts are enabled or not prevented (i.e. ‘error-provoking conditions’) at the institutional (systems) level [22]. Latent failures arise from decisions made by those designing, writing, managing, implementing and enforcing (safety) policies and procedures [21]. Deliberate rule violation versus errors in following rules are two very different forms of non-adherence, but literature on COVID-19 behaviour is yet to sufficiently distinguish between different types of non-adherence. Reason’s model enabled us to separate out the intention behind the behaviour into violations vs. errors. The present paper explores two main research questions: (1) What are participants’ perceptions of the extent of non-adherence to COVID-19 measures in self and others to COVID-19? (2) What do participants’ feel are the main reasons for non-adherence to COVID-19 measures, in self and others? The paper explores these questions in relation to behavioural theory including Reason’s model of human error [20–22]. The study therefore aims to contribute to existing knowledge around people’s experiences and perceptions of non-adherence to COVID-19 measures, with a view to contributing to the evidence base for the current pandemic as well as future infectious disease outbreaks.

Materials and methods

Data from this study came from 12 focus groups with 51 participants were conducted between 25th September and 13th November 2020. Participants were all UK-based adults aged 18 years and above. The study period included a number of important policy developments, which varied across the for UK nations, that provide a context for perceptions of government mandated rules including: the introduction of local lockdowns in Wales (introduced on 27 September); Scotland’s 16 day ban on drinking alcohol in licenses premises (introduced on 7th October); the ‘three tier’ system of COVID-19 restrictions in England (introduced on 12th October); the 4 week closure of pubs and restaurants in Northern Ireland (introduced on 14th October); the 19 day ‘firebreak’ lockdown in Wales (introduced 23rd October); Scotland’s five-tier system (introduced 21st October); and England’s four-week national lockdown (introduced 5th November). Beyond these major developments, a full list and description of all the specific policy changes, including how they differed across the four UK nations is beyond the scope of this research, but full details and ongoing updates of pandemic policies can be found on various ‘policy trackers’ [23] (e.g. https://www.bsg.ox.ac.uk/research/research-projects/covid-19-government-response-tracker).

Participants and data collection

The methodology used in this study has been discussed in previous publications [18, 19]. Online focus groups were a necessary requirement of the pandemic, but have previously been shown to have the advantage of conveniently obtaining the views of a diverse group of geographically disparate group of individuals (although they potentially preclude those without the requisite communication technology) [24, 25]. Purposive sampling was used to obtain as diverse a sample as possible, including gender, age and race and ethnicity and geography (across the UK) (Table 1) Participants in the final sample were recruited from across England, Wales, Scotland and Northern Ireland, although the final sample was primarily from England and Wales (primarily an effect of the snowball sampling technique used). Recruitment took place via a combination of social media advertisements (targeted Facebook ads), online advertising via local community and volunteering sites (e.g. Facebook groups (local community interest groups), Gumtree volunteering groups) and social media snowball recruitment (e.g. via Twitter). Volunteering and community sites were chosen due to their common use as sites to advertise research studies (although this increases the potential risk of a biased sample- toward those who may be more inclined to take part in research studies). Although targeted social media ads attempted to recruit additional members of Black and Asian Minority Ethnic (BAME) participants and those aged over 50, the final sample included a large proportion of white participants aged under 50. As part of the initial recruitment, participants were told the study was looking at the UK public’s perceptions and experiences of social distancing measures during the pandemic.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic details reported by participants.

https://doi.org/10.1371/journal.pone.0258781.t001

Each focus group (average 4 participants per group) met virtually via a web videoconferencing platform (Zoom) for approximately one hour. Participants joined using both video and audio. Focus groups were organised and moderated by SW and KD. The topic guide for the focus groups was initially developed using existing literature on adherence to health behaviours (discussed above) as well as rapidly emerging research on COVID-19 public attitudes. Focus groups were semi-structured and contained a lengthy section on adherence, particularly focused on adherence to government mandates and legal and official measures around social distancing (as noted above, including keeping 1–2 metres physically distant, only meeting and mixing with others as and when permitted). In discussing social distancing measures, participants at times voluntarily discussed (non-)adherence to other rules, including mask mandates. The main topics for the focus groups included: what people thought about their own adherence and the adherence of others to COVID-19 social distancing measures, what they thought the reasons for adherence or non-adherence were, as well as broader topics related to people’s views and experiences around any impact of the pandemic on work life, social life and mental health–the focus of previous and future publications from the project. The focus group schedule is included as S1 File.

Ethical approval was received by Swansea University’s School of Management Research Ethics Committee. All participants gave informed consent, both written and verbal. All data were kept securely an confidentially in line with ethical requirements, and where data is presented below, all quotes are anonymised to protect participants’ identities.

Results

Discussion

Six themes emerged from our qualitative analysis across all focus groups: Alert fatigue; inconsistent rules; lack of trust in government; learned helplessness; resistance and rebelliousness; and (in later focus groups) reduced perception of risk due to the prospect of a vaccine. These themes are not mutually exclusive and are inter-related in complex ways. For example, in some instances the feeling of helplessness was attributed to the frustration of not being able to “keep up” with rules (alert fatigue) or the feeling they were inconsistent or didn’t “make sense”. Also, the resistance and rebelliousness reported was in part seen as a response to the frustration over inconsistent policy and a lack of trust in or respect for government, compounded by the perceived loss of agency and control (attributions linked to learned helplessness. Taken together these factors were associated with reduced adherence to rules either in the form of violations or errors. However, a key finding is how prominently latent conditions figured in participants accounts of non-adherence. Rather than violations and errors to stem from individual failings, they were instead seen to stem from systemic (latent) failures [20], including what were perceived as overly frequent, confusing and inconsistent rules set by what were seen to be untrustworthy decision-makers.

Recent survey evidence suggests that self-reported adherence (or at least ‘majority adherence’) in the UK has generally remained high and stable throughout the pandemic [5]. It may be the case that discrepancies exist between people’s perceptions of (high) adherence in self and their perceptions of (low) adherence in others. However, our findings suggest that, from the perspective of our participants, non-adherence in both self and others may be more apparent, as compared to earlier stages in the pandemic [18]. All six themes were observed in related to both self and other non-adherence. However, certain themes were more prominent in relation to participants’ accounts of non-adherence in self compared to non-adherence in others (and vice-versa). For example, participants were more likely to discuss non-adherence in self in terms of alert fatigue or subjective rule interpretation and others’ non-adherence in terms of resistance and rebelliousness and rule violations. A number of factors could account for this discrepancy, including social desirability bias (the desire to appear conscientious) or selection bias (those more likely to adhere might also be more likely to take part in a research study on adherence and might be more likely to be recruited via non-random sampling via social media). Perhaps one explanation for this is the fundamental attribution error, or specifically the actor-observer bias [27], which the holds that people tend to believe others’ actions stem from stable characteristics (e.g. rebelliousness) while seeing their own actions to also stem from contextual factors (e.g. an overload of complex information). It is important to note however that inconsistent policy and a lack of trust was pervasive in both self and others’ non-adherence.

Although, as discussed above, the notion of ‘behavioural fatigue’ is a largely vague and poorly evidenced concept [3], our findings do suggest that a very specific form of fatigue—alert fatigue—may be an important factor for non-adherence to guidelines. This phenomenon is explained as a result of cognitive overload (where too much information cannot be processed, or retained, effectively) or desensitisation (where repeated exposure to alerts leads to declining responsiveness to them) [36]. In their study of public health care providers, Baseman et al. [37] argued that “during a pandemic when numerous messaged are sent, alert fatigue may impact ability to recall when a specific message has been received due to the ‘noise’ created by the higher number of messages”. Our findings suggest that in the context of the current coronavirus pandemic, alert fatigue is a wider phenomenon being experienced by some within the general public. Indeed, the concept of alert fatigue has its roots in the much broader concept of “information overload” which was originally taken to be where a person develops a “blasé outlook” as a means to cope with a constant influx of stimuli [38]. Many of our participants described how they were no longer paying as much attention to—or even actively avoiding—what they perceived as “constant” news on COVID-19, including what was perceived as excessively frequent and detailed announcements (alerts) made by political leaders. As such, in their desire to provide “transparency”, and to communicate “the science” behind policy, government may have counter-productively “overloaded” the public with information to the point at which some are actively avoiding it. Such avoidance could be construed as a coping mechanism both in regard to the cognitive overload (of having to filter through the “noise”) and in regard to the anxiety or stress caused by the focus on the pandemic and the measures.

Of course, the frequency of the alerts was also related to the fact that policies were also changing frequently. Our findings suggest that participants were often confused over what they perceived to be constantly changing rules, which were sometimes inconsistent or didn’t “make sense” to them. Significantly, confusion was constructed not as an individual error but rather as a consequence of the latent failures or error-provoking conditions [20] created by confusing policy. The differing policy approaches and timelines between the four nations was a key source of confusion, as was the introduction of local lockdowns and, in England, the tier system. Significantly, the geographic variability, in addition to the variability over time, was perceived by some as being a reason as to why they were no longer able to, or no longer chose to, follow “the rules”, and can as such be seen as an example of a latent or system failure.

Related to both alert fatigue and inconsistency in rules was a general sense of a lack of trust in, or respect for, government. The relatively low confidence in the UK government has been documented in longitudinal surveys [5, 39]. Additionally, it is well-established that role modelling by people in positions of leadership is an important factor in motivating adherence, and research on COVID-19 is starting to explore the adverse impact of poor behavioural role modelling (the so-called “Cummings Effect” [40]). As seminal research in psychology has shown, trust and confidence in those in authority plays a key role in the extent to which they are followed [41], as does positive role modelling [42]. Milgram [41] found that polyphony—that is the disagreement between conflicting voices of authority—negatively impacts compliance by increasing the likelihood that people will have to decide for themselves what action to take. In the case of COVID-19 measures, there has been an increasing trend towards polyphony amongst political leaders across the UK, something participants noted when contrasting recent measures to the initial UK-wide lockdown in March. The existence of different announcements about different countries and regions coming from multiple voices (e.g. national and regional political leaders) likely compounds the general sense of alert fatigue caused by the frequency of announcements. Our findings suggest that, from the perspective of our participants, a more unified set of measures across the UK, or at least the appearance of a more collaborative and unified voice, would have resulted in an improved capacity to understand, and thus potentially follow, rules.

Large UK surveys captured increases in depression and anxiety during the first wave of the pandemic, that largely improved over the summer, and worsened again during the second wave [5]. Our previous qualitative research, conducted during the first wave of the pandemic, supported these findings and explored people’s experiences of social isolation. We found that social distancing was already leading to feelings of ‘loss’. Practical losses, (e.g. social interaction; income; structure and routine) and psychological losses (motivation and self-esteem) which in turn were having negative impacts on mental health (increased depression and anxiety [18]. Learned helplessness theory proposes that repeated loss of control or agency over time can lead to depressed mood [32]. Learned helplessness in our study appears to have emerged as a result of the sustained social sacrifice and social suffering (a sustained sense of loss), coupled with a growing sense of a lack of control, as the pandemic has drawn on. Participants reported perceiving the pandemic as stable (something that would never end or cease to have an influence) and uncontrollable (the increasing levels of the pandemic despite sacrifices being made). Participants tended to make external attributions, seeing the government as causing their loss of control due to inconsistent measures and lack of clarity.

Participants were, in a sense, being ‘conditioned’ to feel as though they had little control over the outcome of the pandemic and over their lives in general, and (drawing a comparison to seminal experimental studies [43] they had been experiencing a ‘failure to escape’ social restrictions loss of freedoms). Similar to ‘giving up’ in the context of health behaviours, such as giving up trying to stop smoking after repeated failures, repeated failure in the context of coronavirus measures meant ‘giving up’ trying to adhere to the rules. Such learned helplessness stems from cognitive, affective and motivational deficits [44] which some of our participants reported in the form of a perceived lack of ability to understand, or motivation to adhere to, rules.

Our findings suggest that resistance and rebelliousness, as an overt form of rule-violation, may be becoming more observable–at least in regard to people’s perceptions of non-adherence in others—as resilience to coronavirus measures is increasingly tested. Indeed, resistance can be seen as a natural corollary of control particularly where such control is felt to be excessive or sustained [45]. Our findings reveal the ways in which some people may be performing ‘resistance through rituals’ [46], by for example eschewing or misappropriating new rituals (e.g. failing to wear masks or wearing them improperly) or by intentionally engaging in rituals that were once normative but which are in the context of a pandemic considered deviant (e.g. handshaking or not keeping socially distant). Although a form of violation, this rebelliousness might also be characterised as ‘mis-behaviour’ that ‘arises from an impulse to take control rather than to be always subject of control … and seldom comes simply from a desire to break rules’ per se [47].

Finally, our results suggest that further research is needed on the impact of vaccination programs on perceptions and behaviours related to adherence to measures. For some, the prospect of a vaccine might have an adverse impact on adherence, either by lowering the overall perception of risk in the general public. Research evidence on COVID-19 suggests that one of the key drivers of adherence in the subjective perception of risk [9–11]. However, for others, vaccine programs may not substantially change pre-existing attitudes to adherence. This theme, as with the others identified in our study, will warrant investigation in our ongoing research.

Implications and recommendations for policy and practice

Further research is needed to explore possible approaches to promoting adherence (e.g. through incentives for adherence), and other potential factors impacting behaviours, such as situational events (e.g. public holidays), new threats and variants, and pharmaceutical developments such as vaccine rollouts. However, findings from our research provide insights that are helpful for policy development, and the introduction and implementation of various health security measures, educational campaigns, and health promotion activities during the current pandemic as well as future threats to public health. Our findings on alert fatigue, inconsistency in messaging, lack of trust in government and learned helplessness support our recommendations that government bodies and health officials should to consider ways to: (1) Strike a balance between open and transparent communication around measures and ‘overloading’ the public with information around rules and rule changes; (2) plan for measures that are as unified as possible within and between countries in the UK and use as consistent a message (voice); (3) work to rebuild public trust, through exemplary adherence to rules amongst those in positions of authority. This study provides insights that are helpful for policy development, and the introduction and implementation of various health security measures, educational campaigns, and health promotion activities during the current pandemic as well as future threats to public health. Our findings suggest that latent and systemic failures—in the form of policy decisions that are commonly experienced as too changeable, inconsistent and confusing, and policy makers that are commonly perceived as untrustworthy–may play a significant role in creating the conditions that enable or encourage non-adherence.

Supporting information