Risk communication and COVID-19 in Europe: lessons for future public health crises

Abstract Risk communication is key to engaging with the public on non-pharmaceutical interventions (NPIs) to promote acceptance, compliance and policy support. This article outlines the considerations needed for effective risk communication to promote support for and compliance with NPIs alongside building trust, before assessing the hazard characteristics of COVID-19 on public perceptions. Highlighting examples of risk communication successes and failures in five European case study countries: France, Germany, Sweden, Switzerland and the UK, this article aims to underline the impact of risk communication on public trust and confidence in respective government COVID-19 strategies, and outline recommendations for future public health crises.


Introduction
and sunbathing (Keller and Gollwitzer 2017;Slovic, Fischhoff, and Lichtenstein 1982). Conversely, risks perceived as less controllable such as landslides, nuclear power and chemical/technical food risks increase risk perceptions (Ho et al. 2008;Meagher 2019;Prati and Zani 2013). COVID-19 is characterised by Cori et al. (2020) as an uncontrollable risk to the individual. Maekelae et al. (2020) find a significant relationship between perceived lack of control of the outbreak and greater distress. However, Sobkow et al. (2020) find that higher perceived controllability was linked with more negative risk perceptions and greater individual action on COVID-19. Further research must unpick exactly what is meant by controllability, or 'command over the outcome of a risk' (Nordgren, Van Der Pligt, and Van Harreveld 2007, 533), as against volition, or 'command over exposure to the risk itself' (Nordgren, Van Der Pligt, and Van Harreveld 2007, 533), for COVID-19. For COVID-19 risk perceptions, this could be the difference between fear about the virus impact after infection (controllability) compared to fear about simply catching the virus, which may be perceived as more voluntary.
• Catastrophic potential: People fear multiple-fatality risks more than a series of smaller events (Renn 1989;Slovic 1987). The public displays higher risk perceptions of airplane crashes compared to motor vehicle accidents, even if many more people die in motor vehicles every year (Slovic, Lichtenstein, and Fischhoff 1984). COVID-19 is both objectively a catastrophic risk, and is perceived as a risk with catastrophic potential in a psychometric study by Gerhold (2020) in Germany.

Unknown risk characteristics of COVID-19
The unknown characteristics of risks also influence public perceptions (Slovic 1987). Unknown characteristics include observability and immediacy, levels of actual or perceived knowledge, and novelty: • Observability and immediacy: People worry less about visible hazards (Slovic 1987). Conversely, non-observable risks garner higher levels of worry. COVID-19 is not observable and can transmit asymptomatically between subjects who may not know they are infected (Gandhi, Yokoe, and Havlir 2020). Delayed risk effects can also increase perceptions (Slovic 1987). The full impact of COVID-19 may take some time to be realised and symptom onset is not immediate (He et al. 2020). Cases of extended symptoms of COVID-19 could increase fear and uncertainty in low-risk groups (Carfì, Bernabei, and Landi 2020). • Knowledge levels and novelty: Greater public knowledge of a risk can attenuate perceptions, and people tend to worry less about risks that are known to science (Slovic 1987). Low public knowledge can lead to greater reliance on trusted actors such as experts and scientists to inform decision-making (Siegrist and Cvetkovich 2000). Although coronaviruses are not new, the COVID-19 strain is novel. Public knowledge at the beginning of the pandemic was low, with high scientific uncertainty (Malecki, Keating, and Safdar 2021). Public uncertainty due to lack of knowledge caused by a novel virus has likely increased COVID-19 risk perceptions. This may change as COVID-19 becomes a more familiar risk (Caserotti et al. 2021;Fischhoff 2020).

Summary: impact of risk characteristics on risk perceptions
COVID-19 is characterised as a dread risk with debatable levels of controllability, with catastrophic potential, and as involuntary. COVID-19 was an unknown risk, especially at the start of the pandemic, with low knowledge levels and high scientific uncertainty about its impact and how it was spread due to its novelty (Malecki, Keating, and Safdar 2021). Alongside other non-risk factors, general risk perception of COVID-19 is high due to its individual risk characteristics, especially at the beginning of the pandemic. COVID-19 has therefore received a large amount of public attention, resulting in wide-sweeping policy and restrictions. Uncertainty about the spread and impact of the virus has led to mixed messages (Malecki, Keating, and Safdar 2021). 1 Government communication and policy has been highly precautionary in most nations, even if these NPIs are justified from a public health perspective (Aven and Bouder 2020;Wardman 2020;Warren, Lofstedt, and Wardman 2021).

Messenger attributes in the COVID-19 pandemic
Precisely who communicates to the public is important in ensuring public message acceptance, retention, and action (Bord and O'Connor 1990;Wynne 1991). The UK government has primarily communicated through public press briefings with a minister 2 and at least one scientist or expert 3 . This show of alignment with scientific expertise can promote trust and increase the legitimacy of the government strategy by promoting perceived competence and objectivity (Bogliacino et al. 2021;Jong 2020;Renn and Levine 1991). In Germany, Chancellor Angela Merkel filled the role of being both a politician and someone with a scientific background early on and was lauded for a trustworthy and effective communication style when explaining COVID-19 risks and the need for action (The Economist 2020; Miller 2020). Alongside Chancellor Merkel, decisions were informed by the Robert Koch Institute (RKI), a respected expert government public health agency (Jasanoff et al. 2021). The first phase of the pandemic birthed new celebrity scientists in the form of former Head of the Communicable Diseases for the Swiss Federal Office of Public Health (FOPH) Daniel Koch, and state epidemiologist of Sweden Anders Tegnell (Erdbrink and Anderson 2020; Fahy and Lewenstein 2021; Romy 2020).
Despite these individuals' and organisations' popularity, public opinion can shift quickly, with severe effects on willingness to comply with pandemic-related restrictions (Siegrist and Zingg 2014;Slovic 1993). Following the Swedish Public Health Agency's (PHA) guidance to relax restrictions while cases were increasing, amid great criticism from medical doctors, politicians instead began to take greater responsibility, becoming the new public face of the Swedish COVID-19 response in the place of Tegnell by November 2020 (Bjorklund and Ewing 2020;Claeson and Hanson 2021;Habib 2020). Although scientist popularity can promote trust and support for restriction measures, trust can easily be destroyed (see Battiston, Kashyap, and Rotondi 2021;Slovic 1993). Public support for individual communicators' persuasive messages can also only be as successful as the policy approach behind it: effective risk communication relies on popular, effective and considered risk management (Fischhoff 1995;Lofstedt 2005).
Public opinion of policy makers' and leaders' actions can also influence acceptance and adherence to restrictions, alongside creating new social norms (Schnall and Roper 2012;Van Bavel et al. 2020;Yaffe and Kark 2011). Leading by example has been a recommended practice throughout the COVID-19 pandemic ( Van Bavel et al. 2020;Wardman 2020). Exhibiting traits that show leaders as members of the public, displaying the mantra of we are 'all in this together' through following rules rather than acting in their own interest is strongly recommended in the context of COVID-19 and more generally (Haslam and Platow 2001;Haslam, Reicher, and Platow 2011;Hogg 2001;Van Bavel et al. 2020, 466). Several reports of UK leaders and experts breaking rules, such as political strategist Dominic Cummings, Professor Neil Ferguson, Scottish National Party MP Margaret Ferrier, and four senior Welsh politicians, have functioned as trust-destroying incidents (Neilan 2020;Sky News 2021). This has provoked a divided mentality of one rule for us, another rule for them between the public and the political class (Reicher and Stott 2020;Van Bavel et al. 2020). Since the Dominic Cummings scandal, public compliance to COVID-19 restrictions has decreased alongside confidence in the government to adequately manage the pandemic (Fancourt, Steptoe, and Wright 2020). Across Europe, even acting as a role model to promote NPIs has been lacking. This was seen early on in the pandemic with leaders including French President Emmanuel Macron and Prime Minister Johnson not wearing masks before they became mandated in these nations (Drylie-Carey, Sánchez-Castillo, and Galán-Cubillo 2020; Newton 2020).

Consistency and clarity of COVID-19 communication
Ensuring message consistency can promote public trust in the government's approach to managing risks including COVID-19 (Renn and Levine 1991;Wardman 2020;Warren, Lofstedt, and Wardman 2021). Perceived inconsistency may undermine a message, leading to confusion and public uncertainty about the real impact of risks such as COVID-19. Message inconsistency has been endemic during this pandemic in the face of uncertainty, changing information and policy priorities (Finset et al. 2020;Porat et al. 2020;Renn 2008). Independent SAGE (2020, 1) argue that message consistency and imprecision may result in 'mistakes, perceived inequities, frustration, and non-adherence' , so communicators should pay attention to avoid inconsistent strategies. Many of the nations discussed here have had inconsistent and changeable messaging strategies, including France, Sweden and the UK nationally, and Germany and Switzerland between federal states (Das Erste 2021; Demey 2021; Meyer 2020; Syal 2021; Warren, Lofstedt, and Wardman 2021).
The UK is a particularly bad offender on policy and communication inconsistency, with COVID-19 rules estimated to have changed over 60 times between March 2020 and January 2021 (Syal 2021;Warren, Lofstedt, and Wardman 2021). Despite Independent SAGE (2020, 1) advocating for a 'communications reset' in November, these government communication inconsistencies have continued into 2021. Recently, questions about summer holiday restrictions in 2021 have received contradictory replies from different ministers, outlined by Campbell (2021). Originally, despite Foreign Secretary Dominic Raab urging caution and warning against booking holidays on 17 January, Health Secretary Matt Hancock stated the next day that he planned on spending his summer holiday in Cornwall as part of a 'great British summer' . After widespread pessimistic statements from International Trade Secretary Liz Truss, Transport Secretary Grant Shapps, and Prime Minister Johnson in early February, Hancock appeared more cautious on 11 February, stating that it was too soon to begin plans for summer holidays. This example illustrates the kind of message inconsistency that has plagued the UK government's COVID-19 communications strategy. Sweden has seen similar communication inconsistency in recent months, with the PHA relaxing rules on public meetings, just for Prime Minister Stefan Löfven to overrule this NPI relaxation two weeks later (Bjorklund and Ewing 2020;Claeson and Hanson 2021;Habib 2020). In Germany, communication has been inconsistent due to varying rules and implementation across states, which has caused public uncertainty (Han et al. 2020;Warren, Lofstedt, and Wardman 2021).
Message clarity is necessary to avoid public uncertainty, which could lead to errors in judgment or refusal to adopt new COVID-19 measures (Finset et al. 2020), especially when interventions are recommended rather than mandatory and therefore rely on individual decision making (Fischhoff 2005). Messages must contain clear asks of the public, so that they understand exactly what needs to be done and how to do it (Fishbein and Ajzen 2011). An example of original success is in the UK, where the lockdown slogan of 'Stay at Home, Protect the NHS, Save Lives' offered a clear direction to people not to leave their home (Stay at Home), while also being clear about the rationale (Protect the NHS, Save Lives). This slogan is described by Wardman (2020Wardman ( , 1102 as 'simple, unambiguous, clear, easy to follow' , resulting in widespread compliance and around 90% of people understanding what actions were expected of them (Independent SAGE 2020;Skinner 2020). However, after restrictions were eased and the stay-at-home order was lifted, the slogan changed to 'Stay Alert, Control the Virus, Save Lives' , which was widely mocked on social media and seen as too ambiguous and unclear to outline what people should be doing to limit COVID-19 transmission (Ginnis et al. 2020;Independent SAGE 2020;Wardman 2020).

COVID-19 communication methods
The method by which public leaders communicate influences public reception, acceptance and compliance with message requirements (Rickard 2021;Simcock et al. 2014;Warren and Foulds 2020). Top-down, one-way communication methods are commonplace in fast-paced crises (Lofstedt 2003), and this has been evident throughout the COVID-19 pandemic in all nations discussed. Risk communication in Sweden has remained very much top-down through the entire COVID-19 pandemic. In the first wave, this was more appropriate for PHA to undertake as public trust remained high (Lofstedt 1996(Lofstedt , 2003. This strategy did not change, resulting in no thought-through risk communication strategy despite public confidence in the government approach declining since March to record lows of 30% in late January 2021 (Sjöström 2021). The PHA and Government simply believed the public would just trust the guidance that was being continuously offered as before (Habib 2020;Lofstedt 2005). Typically high trust in public figures has drastically reduced, and top-down communication alongside an original avoidance of enforcing stringent restrictions has not helped Sweden to avoid a second COVID-19 wave (Habib 2020;Sjöström 2021). A bottom-up focus, understanding what people need, integrating the public in decision making and creating a dialogue rather than simply providing facts through a top-down approach is a more effective method to engage the public on COVID-19 and increase efficacy (Nazione, Perrault, and Pace 2021;Porat et al. 2020;Richards and Scowcroft 2020). This is especially important in nations with low levels of social trust in government or access to social resources (Ihm and Lee 2021;Kasperson 2014), such as the UK or France.
Message format can also influence effectiveness (Rickard 2021). One example of this is through the use of visual imagery to communicate risk (Lipkus 2007). France was one of the first nations early on in the pandemic to use a traffic light heat map to display geographical variations in COVID-19 risk. Visualising the local risk level on a map based on three main indicators 4 allowed the public to easily understand their local area's rating in terms of safety and the NPI rationale. The traffic light system has become a commonplace visualisation of risk as it is easily understandable (Hieke and Wilczynski 2012), often more so than other visualisation methods (Roberto et al. 2012). Many including the RKI in Germany and the Swedish PHA through a live-tracker (Folkhälsomyndigheten Sverige 2021; Robert Koch Institut 2021), the UK Chief Medical Officer during press briefings (Cabinet Office Briefing Rooms 2020) and the Swiss FOPH in their weekly briefing (Bundesamt für Gesundheit BAG 2021) have since adopted risk maps to visualise geographical spread of COVID-19 cases and data, however, none are as clear and understandable as the original French version from May 2020.

COVID-19 message attributes: framing and emotive messaging
Changing communication framing can influence emotions and affective response to COVID-19 (Heffner, Vives, and FeldmanHall 2021;Rickard 2021;Roeser 2012). Different emotions influence risk perceptions, decision making likelihood, strength and speed of action, and may focus public attention on certain characteristics of a risk (Slovic 2010;Van Bavel et al. 2020). There are many potential emotions that can be evoked, and this article will explore three broad groups in the context of COVID-19 communication messages, namely: guilt and fear (negative affective messaging), humour, and solidarity and empathy (positive affective messaging). 5 The UK government has primarily undertaken an information-based communications strategy between spring and autumn 2020, best exemplified through the Hands. Face. Space advertising campaign released in early September 2020 (Corbishley 2020). A new self-described 'hard hitting' fear-and guilt-based emotional television advertising campaign replaced it in January 2021, where the government asks: 'Can you look them in the eyes and tell them you're helping by staying at home?' while faces of patients on ventilators and healthcare staff are shown (Department of Health and Social Care 2021). The campaign aimed to promote more protective behaviours and to give the public a 'massive jolt' , in the words of Chief Medical Officer Chris Whitty, required due to a lack of behaviour change thus far (Shipman and Wheeler 2021). It shifted responsibility and blame onto the individual for their actions as hospitalisation rates remained high (Cunliffe 2021; Department of Health and Social Care 2021), to elicit a negative affective response. Despite Keller et al. (2006) finding a link between increased negative affect and perceived risk, Peters et al. (2013) find that when this frame is used without information on how to limit risk then this kind of message is often ignored by the public. This communications approach to COVID-19 has specifically been warned against by Stolow et al. (2020, 532), who warn that negative affective messages including fear can actually backfire through 'unanticipated, negative, reactionary behaviors' . This should especially be avoided as pandemic weariness has recently increased in several European nations, reducing the impact of fear-based messaging on capturing public attention (Cevipof 2021; Sutton et al. 2020).
Although negative affective messages are commonplace, positive affective messages have also been employed during the COVID-19 pandemic. One example involves a humour-based positive social media campaign from the German government titled '#specialheroes -together against corona' , a public health message to promote patience and solidarity, where in a documentary-style interview an old man is asked to reflect on surviving the pandemic (Bundesregierung Deutschland 2020). He states: The fate of this country lay in our hands, so we mustered all our courage and did what was expected of us. The only right thing. We did… nothing. Absolutely nothing. (Bundesregierung Deutschland 2020).
Although some criticised the advert for being aimed at a very small demographic (Oelze 2020), the hashtag was widely shared through social media and received praise for its positive style and focus on young adults (Kluth 2020). Although communicating through humour can backfire, it has been found to raise awareness of risks and promote behavioural change, and stimulate the processing of negative events and emotions (Eisend 2009; Kaltenbacher and Drews 2020; Murthy and Gross 2017; Skurka et al. 2018). The #specialheroes hashtag shows a more positive messaging technique, using social media memes and viral video content to promote awareness, coping strategies and action against risks (Ross and Rivers 2019), and in this case COVID-19 (Cauberghe et al. 2021).
Promoting social solidarity and collectivism is significant when requiring or requesting that groups undertake altruistic actions or accept limitations to freedoms to protect other groups from risks (Bierhoff and Küpper 1999;Fincher et al. 2008;Roos et al. 2015). For COVID-19, this may involve younger individuals staying at home and physically distancing to stop the spread of a disease to which they are not as susceptible compared to older adults or individuals with underlying health conditions (Porat et al. 2020;Yanez et al. 2020). Avoiding an us versus them mindset is important as it can lead to counterproductive actions (Porat et al. 2020;Van Bavel et al. 2020), such as panic buying seen in early 2020 in many nations (Habersaat et al. 2020). A communications approach that promotes solidarity is recommended by researchers from many academic disciplines (see for example Drury, Reicher, and Stott 2020;Elcheroth and Drury 2020;Habersaat et al. 2020;Porat et al. 2020;Van Bavel et al. 2020). At the beginning of the pandemic, the Swiss government was particularly effective in promoting solidarity through the 'Wir/Nous/ Noi/Nus' campaign. As part of an extensive communications campaign by the FOPH, Health Minister Alain Berset personally encouraged Swiss people in a video shared on social media to show what they were doing to keep safe, using a hashtag in three national languages. This campaign has developed in different ways, including projecting the Swiss flag with 'Wir/Nous/ Noi/Nus' onto the Matterhorn mountain on 31 March. The solidarity and 'all in this together' approach was, at the time, a fundamental part of the FOPH's overall communications approach to COVID-19. Switzerland displayed a high level of adherence to rules in the early stages of the pandemic, with strong public support for further stringency since late March (Hermann 2020). A cohort study found 87% of young adults, often seen as the group most likely to flout rules, adhered to the prescribed non-mandatory health measures to stay at home (Nivette et al. 2021).

Public trust in government in the wake of COVID-19
Overall, public trust is vital to effective risk communication (Cvetkovich and Lofstedt 1999;Lofstedt 2005;McComas 2006;Slovic 1993). Public trust has a significant impact on public acceptance of interventions and likelihood to comply with government COVID-19 NPIs (Clark et al. 2020;Soveri et al. 2020;Wright, Steptoe, and Fancourt 2020;Nivette et al. 2021). Despite some evidence of communication that follows recommendations from scientific bodies, much of the COVID-19 communication and risk management strategy in the five nations discussed has not fostered public trust, decreasing public confidence especially from autumn 2020 (Warren, Lofstedt, and Wardman 2021).
UK public trust in government has always been relatively low. Even in April 2020, although 95% of respondents agreed that the government should have power to enforce restrictions, only 52% believed the government was effective in managing the pandemic and 42% trusted that the government 'always or mostly' told the truth on COVID-19 -this is actually higher than seen in previous years, attributed to a 'crisis effect ' (Enria et al. 2021, 15-16). By late November, public distrust in the government's ability to control the pandemic reached 57% (Allington et al. 2020). However, following positive press on the vaccine rollout, belief that the government is handling the COVID-19 pandemic 'well or very well' reached 38% in February 2021, up from 30% in October 2020 (Skinner et al. 2021, 22).
The French public has displayed high levels of government distrust since the start of the pandemic. In an Ipsos (2020) study in April 2020, only 35% of French respondents trusted the national government. Public trust has remained stubbornly around this level, with a February 2021 study by Odoxa-Backbone Consulting for Le Figaro and France Info showing 39% public trust and 60% distrust in the government's ability to manage the COVID-19 health crisis (Odoxa 2021).
German public likelihood to trust the government reached its highest level in 5 years in July/ August 2020, at 61% 'rather trusting' (Statista 2020). Since, happiness with the federal government approach has waned, reaching its lowest level yet in February 2021, where only 42% of respondents were either very happy or happy, and 22% being not at all happy (Infratest dimap 2021).
Public trust in the Swedish government has greatly reduced following the first wave, where 64% trusted the government's approach to the COVID-19 pandemic, and 75% trust in the PHA at the end of March 2020 (Sjöström 2021). By late January, this had consistently decreased, to 30% and 50% respectively (Sjöström 2021).
In Switzerland, public trust in the Federal Council has also dramatically declined. After starting strongly, with 61% of respondents from all language regions either self-declaring high or very high trust in the Federal council in March 2020, this has plummeted to 32% in January 2021 (Bosshard et al. 2021).
Overall, declining public trust in all countries highlights the task facing respective governments to promote public confidence in their approach to tackling the COVID-19 pandemic. Growing mistrust in politics has also led to an increase in COVID-19 conspiracy theory belief (Cevipof 2021).

Conclusions and recommendations
Overall, COVID-19 is perceived as a dread risk with unknown qualities (Cori et al. 2020), especially at the beginning of the pandemic. This has resulted in high public perceptions of risk, great media scrutiny and thus more effective risk communication strategies are needed (Wardman 2020;Warren, Lofstedt, and Wardman 2021). To achieve this, communicators must consider the messenger, audience, and message characteristics (Rickard 2021). With COVID-19 risk communication, the use of scientists and trusted messengers is widely used in the nations discussed. Leading by example in a crisis is vital (Schnall and Roper 2012;Van Bavel et al. 2020), however, examples from the UK and France highlight where this can go astray, with damaging effects. Message consistency and clarity can also encourage public engagement and support for government COVID-19 interventions, although examples from across Europe suggest that this is not occurring. In a crisis, top-down risk communication can be a tempting method to use (see Lofstedt 2004;Olsson, 2014). However, alongside the theoretical critique of this approach, especially in cases of low public trust in government (Leiss 1996;Lofstedt 2003), its effects after over a year of the pandemic suggest that one-way communications employed throughout Europe are not functioning as desired. Finally, using frames to elicit emotions can promote behaviour change, but it is important to apply them correctly. The use of framing has seen varying applications and success during the COVID-19 pandemic.
Based on the findings from the five nations studied, the following recommendations are offered for risk communicators in future public health crises: • Ensure communication is consistent over time, and U-turns are avoided if possible; • Keep instructions clear and actionable to ensure public compliance and to avert disengagement; • Embody leading by example to promote solidarity and to avoid trust-destroying incidents; • Avoid the use of fear-and guilt-based frames in communications campaigns, especially after the initial phase of a crisis; • Move away from orthodox top-down communication approaches, towards bottom-up tailored risk communication that accounts for public values, perceptions and situation; • Pre-test communications campaigns beforehand and evaluate them afterwards to increase effectiveness and avoid potential unexpected outcomes.

1.
For example, the original focus on communicating the need to wash hands over the use of face coverings in the early stages of the pandemic in the UK (Sample 2020

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
The first author would like to acknowledge the financial support of the King's College London Geography Department's Research Fund and Impact Fund respectively which made the research and writing up of this article possible.