Background
The United Kingdom (UK) National Risk Register details a broad range of threats to the public health and security of the UK incorporating infectious disease outbreaks (e.g., pandemics and emerging diseases), malicious attacks (e.g., terrorist incidents), and natural phenomena (e.g., extreme weather, earthquakes) [
1]. The risk of infectious disease outbreaks is so substantial that the UK National Risk Register ranks a pandemic outbreak as the number one high consequence civil emergency facing the UK (based on likelihood and probable impact [
1]. The coronavirus disease (COVID-19) pandemic which, at the time of writing has led to 19,440,423 confirmed cases and 722,706 deaths, presents a stark reminder of this public health threat [
2].
Outbreaks of infectious disease, particularly those for which little or no pre-existing immunity exists – such as the COVID-19 pandemic – represent a significant risk to public health. For example: the 2013–2016 Ebola outbreak in West Africa led to over 28,600 cases with 11,325 deaths [
3], while the ongoing outbreak in the Democratic Republic of Congo has led to over 2200 deaths thus far [
4]; since the identification of Middle East respiratory syndrome coronavirus (MERS-CoV) in 2012, 851 associated deaths have been reported with cases across 27 countries [
5], and; although less severe than expected [
6], the H1N1 pandemic was estimated as responsible for between 151,700–575,400 deaths worldwide during the first 12 months [
7,
8].
Even when controlling for confounding factors (e.g., improvements in surveillance, communication infrastructure, etc.), the number of infectious disease outbreaks has substantially increased since 1980 to 2013 [
9]. Similarly, deaths from terrorism have substantially increased from less than 200 in 1970 to over 26,000 in 2017 (peaking with over 44,000 in 2014 [
10]), and despite a decline in the number of individuals affected by natural disasters between 1994 and 2014, the average death rate has increased over the same time period [
11]. Given these trends, it is therefore critical to ensure that emergency preparedness, response and resilience is optimised to mitigate the occurrence and/or impact of these events.
Behavioural science represents one such broad method of mitigation. The importance of encouraging adaptive and protective behaviour change in response to public health emergencies is emphasised by the World Health Organisation (WHO), who provide risk communications guidelines designed to encourage individuals, families, and communities to act to protect themselves [
12]. This is echoed in the context of COVID-19, with Michie and colleagues stating that: “human behaviour will determine how quickly covid-19 spreads and the mortality. Therefore behavioural science must be at the heart of the public health response” [
13].
Research in the behavioural sciences has focused on identifying barriers and facilitators to maximising public compliance with recommended emergency response and infection prevention behaviours. For example, decontamination behaviour (e.g. [
14]), medication adherence (e.g., [
15]), hand washing (e.g., [
16]), social distancing/ avoidance behaviour (e.g., [
16,
17]), and vaccination (e.g., [
18,
19]), to name but a few. Furthermore, in the context of infectious disease emergencies, mathematical models are used to both: a) understand and map out the spread and control of disease (incorporating human-to-human transmission) and, b) calculate the potential effectiveness of interventions (including behavioural interventions) to reduce the spread of the disease [
20]. Considered together, the importance of human behaviour for emergency response – both in terms of developing interventions and its relevance for modelling the potential efficacy of said interventions – is clear.
However, there is still work to be done to optimise the incorporation of behavioural constructs in public health research, intervention design, and modelling. For example, despite Medical Research Council Guidelines recommending interventions be based on appropriate behaviour change theory [
21] (see also [
22]), reference to theory is often absent in such interventions [
22]. Indeed, Michie and colleagues, pioneers in the field of identifying and integrating behaviour change theory and techniques in the context of health promotion, note that much intervention design is based on the principle of “It Seemed Like A Good Idea At The Time”, rather than a systematic consideration and assessment of the most appropriate routes to behaviour change ([
23], p. 14).
Similarly, a limitation of traditional mathematical models is that they often do not allow for heterogeneous behavioural responses within a population [
24]. This assumption that human behaviour is homogenous can impact on the validity of these models. For instance, including a modest degree of fear-related flight behaviour (i.e., 10% of individuals in a model respond to fear of infection with flight) into a model in which fear of infection otherwise leads to hiding, caused projected disease incidence to rise to ~ 65%, up from ~ 30% in a model in which fear of infection led all individuals to hide [
25]. Although some recent infectious disease models do incorporate social and cognitive predictors of the kinds of self-protective health behaviours that are associated with infectious disease control and emergency response (e.g., vaccination uptake, social distancing etc.), they are more commonly informed by literature from behavioural economics than psychology [
26].
That is not to say that the integration of theory is a silver bullet for the success of mitigation strategies and modelling. For example, there is mixed evidence concerning the efficacy of theory-based interventions (see [
22], p21 for a summary), inconsistency that may be based on the relevance of the chosen theory for the behaviour in question [
22]. To illustrate this point, according to Michie and colleagues, there are a total of 83 behaviour change theories across the behavioural and social sciences [
22]. Over the past three decades, multiple review papers and books have attempted to identify trends in theory use including those most frequently applied (e.g., [
22,
27‐
30]). Despite some commonalities in underlying psychological processes (Michie and colleagues cede that many of the 1659 constructs identified within their book were different labels for overlapping constructs, [
22]), this proliferation of competing theories and recommendations could indeed make it difficult for researchers, intervention designers, and modellers to decide which theories to use and in what context.
This unfortunately leads to a catch-22 situation: we wish to encourage non-specialists to use appropriate psychological theories and approaches within their own disciplines, yet we fail to recognise the complex and confusing landscape of psychological theory. Michie and colleagues have made great strides to simplify the process by which psychological theories are used to inform behavioural interventions [
23]. However, there are still a large number of behaviour change theories that were designed with specific applications in mind: for example, the Behavioural-Ecological Model of Adolescent AIDS Prevention [
31], the Integrated Theory of Drinking Behaviour [
32], or the Social Ecological Model of Walking [
33] Public health researchers, infectious disease modellers, and practitioners may therefore be understandably perplexed as to how best to model and examine or influence behaviour in the specific context of infectious disease outbreaks or emergency response.
This current paper therefore seeks to present a synthesis of the behaviour change theories that are most commonly applied within an emergency response or infectious disease outbreak context. That is, focused specifically on using behaviour change theories to understand and influence individuals’ engagement with protective health behaviours that are recommended during infectious disease outbreaks and public health emergencies. To identify these commonly applied theories, we conducted a scoping review of the existing literature, but with a particular focus on identifying reviews using behaviour change theory in an infectious disease or emergency response context. This approach is recognised as a method of distilling a substantial literature into a manageable summary of evidence for decision makers ([
34], see also [
35]). Although using this ‘review of reviews’ approach focused on secondary sources, which may have led to some relevant information being missed, it enabled us to reduce the quantity of papers identified in a large and highly diverse literature to a manageable level while still achieving a broad overview of the state of the art within the field. By dovetailing with Weston and colleagues’ recent review of the application of human behaviour within infectious disease models [
26], the outcomes from this current review will enable us to make useful recommendations as to how psychological constructs, theory, and research can be used by public health practitioners, researchers modellers, to improve our understanding of human behaviour within the contexts of infectious disease outbreaks and emergency response.
Methods
A scoping approach was adopted for our search. Scoping reviews are recommended as a mechanism by which a given literature might be summarised for policy makers or practitioners [
36]. As the aim of this review was to summarise and synthesise the psychological literature on behaviour change to inform recommendations for public health researchers and modellers, the adoption of a scoping review framework was a logical and appropriate choice.
Original search strategy
The literature search was conducted using PubMed, PsychInfo and Medline databases on the 6th January 2016. The databases were selected based on their coverage of discipline and context specific literature. Each database was searched individually to ensure that all Medical Subject Heading (MeSH) terms were used effectively. The search terms covered four major categories: behaviour, emergency response (e.g., infectious disease, preparedness, mass emergency), theoretical models, and reviews. Supplementary Information
1 provides the full list of search terms used for each database. Within the theoretical model category, we a priori selected several existing behaviour change models that were either: a) frequently cited within the literature, or b) adjudged to be of particular relevance within the context of infectious disease and emergency preparedness based on the authors combined expertise in these areas. In addition, generic phrases and subject headings for theoretical modelling were included within each search strategy to ensure that papers that do not cite the most common behaviour change models would still be captured within our search. Lastly, papers identified through other, non-systematic methods (e.g., some clearly relevant citations in papers, keyword Google Scholar searches) were also included to try to identify articles that were not indexed within these databases.
Top-up search strategy
As the initial search was run in 2016, a condensed follow up search strategy was devised to identify seminal works in the field published since this date. Given time and resource constraints in conducting this search, the strategy was designed to identify literature that closely corresponded to the output from the original selection process. For example, the strategy was simplified based on the broad search categories used in the original search, and as literature concerning Human Immunodeficiency Virus (HIV)/ Sexually Transmitted Infections (STIs) was excluded from the original data extraction (see Inclusion/ Exclusion criteria section), the decision was taken to exclude these papers at the search strategy stage here. On 24/10/19 the following search was conducted on Google Scholar, sorted by relevance, with a custom date range of 2016–2019:
“review* AND behavio* AND theor*, OR models AND infectious disease*, OR emergenc* -HIV, -STD, -STI”Given time and resource constraints, only the first 20 pages of Google Scholar were screened, first for title, then for abstract, and finally full-text screening was conducted on any remaining papers.
Optimised top-up search strategy
Due to limitations concerning the use of wildcard operators (*) on Google Scholar in the initial top-up search, and the potential for COVID-19 related review papers to have been published in the intervening period, a further optimised top-up search strategy was developed. This strategy consisted of the following two searches (specified for emergency response and infectious diseases respectively), conducted on Google Scholar on 16th – 17th May 2020:
review emergency theory behavior OR behaviour.
review disease theory behavior OR behaviour.
As for the previous top-up search, the first 20 pages of Google Scholar were screened for each search (for a total of 400 results).
Study selection
For the original search, duplicates were removed electronically, and all remaining papers were subjected to title and abstract screening by one author (AI) using the inclusion/ exclusion criteria. All papers retained for full text assessment were screened independently by two researchers (AI and DW) to increase the reliability of the selection process. Any inconsistencies between the researchers were resolved through a joint discussion.
For the top-up search, individual title, abstract, and full-text screening stages were conducted by the first author (DW). As for the original search, all papers retained for full text assessment were screened independently by two researchers (AI and DW) with any inconsistencies between the researchers resolved through a joint discussion. The inclusion/ exclusion criteria employed in the top-up search were the same as those used for the original search.
For the optimised top-up search, individual title, abstract, and full-text screening stages were conducted by the first author (DW). Due to time constraints imposed by the COVID-19 pandemic, full text screening for this search was conducted by the first author alone. As for the initial top-up search, the same inclusion/ exclusion criteria were employed as used in the original review.
Inclusion/exclusion criteria
The following inclusion/ exclusion criteria were used:
(i)
Type of article: Reviews (systematic, scoping and narrative) and meta-analysis.
(ii)
Theoretical model/theory: The papers needed to present or apply a model or theory of behaviour change. Leniency in this criterion was initially applied in so far as papers which clearly applied constructs that were adapted from theories/models, were also retained, but this was subsequently restricted to focus specifically on the presentation of behaviour change models/ theories (see Original Study Selection section below).
(iii)
Context: The papers needed to present or apply the theory/model to explain human behaviour in the context of emergency or infectious disease outbreaks. As per [
26] any reviews focusing on diseases that are not transferred from human-to-human (e.g., vector borne) were excluded.
(iv)
Target behaviour: Preventive health behaviours during an emergency or outbreak (e.g. social distancing, vaccination and reducing social ties) were included in the review.
(v)
Other: There were no restrictions on the date reviews were published or the population in question. Reviews were included if they were written in the English language and involved human behaviour.
For simplicity, papers exploring STI-related health behaviours (total: 53) were pragmatically excluded wholesale following the initial screening of the original search as the majority were deemed either: irrelevant according to the above criteria (30), of unclear relevance to the researcher (AI) (6), or inaccessible to the researcher (AI) (15). For consistency, papers exploring STI-related health behaviours have also been excluded during subsequent screening of the top-up searches.
Based on data extracted as part of previous review work in this area (e.g., [
27,
28]), the following information was extracted from the included papers in both the original and top-up stages:
(iii)
Number of studies included in the review
(vi)
Key outcomes/ conclusions regarding the utility of behaviour change models
Data concerning the theories employed was identified within the included papers using the original reviews’ own definitions or conceptualisations of theory. That is, if an included review referenced a particular theory or model, it was subsequently included in our synthesis. Reference to theory was either found in specific citations of theories used by individual papers incorporated within the review (commonly included in summary tables within the included papers), or as a broader framework used by included reviews to collate and synthesise the identified literature. Key outcomes and conclusions regarding the utility of behaviour change models were identified similarly, using review authors’ references to the theories they cite within their Results and Discussion sections.
Data from the included studies were synthesised to identify: a) the behaviour change theories most commonly employed to understand and influence protective health behaviours during public health emergencies and infectious disease outbreaks and, b) any (in) consistency in the reported utility of different behaviour change theories.
Discussion
To the best of the authors’ knowledge, this scoping review represents the first attempt to systematically collate review data concerning the application of behaviour change theories within a broad infectious disease and emergency response context. In this review we have synthesised the health behaviours, theories, and applications presented across 13 review papers drawn from both peer reviewed journals and grey literature in the context of infectious disease and emergency response. This synthesis enables us to provide some key recommendations and suggestions for infectious disease modellers, researchers, and public health professionals looking to apply behaviour change theory to understand and influence behaviour in this context.
Looking across the 13 reviews included in our synthesis provides a clear picture of the typical use of behaviour change theories across the current context. Firstly, many papers included within these reviews do not seem to be explicitly based on a specific theory of behaviour change. Secondly, whether we take a high-level approach (i.e., number of theories cited by multiple reviews) or a more granular approach (i.e., number of citations per theory across all reviews) to the synthesis, the conclusions are broadly the same. As per Michie and colleagues, only a small number of theories were most commonly cited despite the broad number of theories available (83 theories detailed within [
22], and 44 distinct theories cited at least once within our included review papers, although these may not all be present in Michie and colleagues work [
22]). Specifically, three theories stand out as the most commonly applied: Health Belief Model, Theory of Planned Behaviour, and Protection Motivation Theory. Another four theories are also repeatedly, but less frequently, cited: Precaution Adoption Process Model, Extended Parallel Process Model, Theory of Reasoned Action, and Social Cognitive Theory.
Of these seven theories, four are consistent with the most frequently used theories as identified by Michie and colleagues [
22] (Theory of Planned Behaviour, Health Belief Model, Precaution Adoption Process Model, & Social Cognitive Theory), and with some of the theories cited as frequently used across other health behaviour contexts e.g., [
27,
28] (Health Belief Model, Theory of Planned Behaviour, & Social Cognitive Theory). Of the remaining theories, one (Theory of Reasoned Action) is closely linked to the Theory of Planned Behaviour (the latter having developed from the former [
57]). The final two most cited theories – Protection Motivation Theory and the Extended Parallel Process Model —are closely related (the latter builds on the former [
61]) and both are particularly concerned with the processes underlying fear appeals and threat messaging, a focus with clear relevance in the context of this review.
Having identified the commonly applied theories, we subsequently conducted a rapid synthesis of the included reviews’ key outcomes with regards to the utility of behaviour change theory. On the basis of this synthesis, we make two key conclusions. Firstly, the most frequently cited theories do find broad (though not universal) support within the cited literature. Secondly, despite this broad applicability, several reviews cited herein highlight the relative absence of behaviour change theory within research conducted in this context [
40,
41,
44,
46,
50]. Indeed, several of the reviews included herein explicitly advocate further work to study the thorough application of both individual theories and a range of multivariable theories, considering the interrelationship between model factors/ components (e.g., [
37,
38,
40,
41,
44‐
46,
50]).
Overall, a clear take-home message from the current review is that there are a range of commonly applied behaviour change theories with broad support for their use within an infectious disease and emergency response context. In the following section, these findings are used to form the basis for recommendations concerning the use of behaviour change theory by researchers, practitioners, and modellers in both research and practice.
Recommendations
Based on the results of our synthesis, there are two broad categories of recommendations for future work applying and incorporating behaviour change theories into both: 1) research and practice (i.e., understanding behaviour, and developing & deploying effective interventions), and; 2) infectious disease modelling.
Furthermore, although most of the literature screening and synthesis for this review was conducted prior to the COVID-19 pandemic, we are aware of some recent and relevant work detailing recommendations and guidance for the use of behavioural science to tackle COVID-19 in practice. Some of this work has also been incorporated into this section to signpost interested readers to additional material of relevance.
Researchers and practitioners
Behavioural science can play a critical role in combatting the effects of a global pandemic, such as COVID-19, both through informing our understanding of public perceptions of the virus, and through developing interventions to reduce barriers and facilitate uptake of recommended behaviours [
13]. Indeed, in the context of COVID-19, West and colleagues indicate that in the absence of robust intervention data, behaviour change theories and constructs should be used to inform the development of policy and practice for increasing uptake of self-protective behaviours [
67].
Unfortunately, however, multiple reviews cited herein lament the absence of behaviour change theory in work conducted to date [
40,
41,
44,
46,
50], and advocate for the more in-depth study of various behaviour change theory within this context [
37,
38,
40,
41,
44‐
46,
50]. Our primary recommendation therefore reinforces that advocated in the literature cited herein. Specifically, we recommend that researchers and practitioners working in the context of infectious disease and emergency response, should: a) draw on the available theoretical literature and; b) work with experts in behavioural science to inform both empirical work to understand behaviour, and the design and implementation of interventions to affect behaviour. This primary recommendation is supported by several additional recommendations, which are unpacked in the subsequent paragraphs.
In describing the way forward for behaviour change theorising, Michie and colleagues [
22] note that the most popular behaviour change theories are relatively context agnostic, and that there may be important insights to be taken from models that are more context dependent. In order to enhance the translation of research into public health practice, we therefore recommend that future research and intervention development should consider both general theories and theories that most closely fits the context of study. The theories identified within this review represent a mix of both context agnostic (e.g., Health Belief Model, Theory of Planned Behaviour) and some more context specific (e.g., Protection Motivation Theory and Extended Parallel Process Model) theories, and would therefore seem to be a good starting point for researchers and practitioners working in this area. Furthermore, although not a theory identified within the current review (see Limitations section), the COM-B (capability, opportunity, motivation and behaviour) model [
68] has been advocated as a key starting point for interventions to reduce the transmission of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) during the COVID-19 pandemic (e.g., [
67]). We would therefore further recommend considering the application of this theory when either conducting research or delivering research into practice.
While these recommendations are consistent with the Medical Council Guidelines [
21], and echo recent work by Michie and colleagues (e.g., [
22,
23]), it is critically important to approach the incorporation of behaviour change theory –particularly within intervention design—in a systematic fashion [
23]. Indeed, one review included within our synthesis found that theory had a positive impact on the success of interventions when used at the design and development stage, relative to theory used only at the evaluation stage [
37]. To facilitate the systematic and appropriate use of behaviour change theory, we therefore strongly recommend that researchers and practitioners involve expert behavioural scientists in the design and implementation process, and also make use of available guidance within the behaviour change literature. For example, we are aware of a thorough guide to designing interventions using the Behaviour Change Wheel [
23] that may be of use to practitioners.
We also echo Michie and Prestwich’s recommendation for researchers to use their Theory Coding Scheme (a list of 19 items for coding the use of theory within intervention design) to both: a) allow for researchers and practitioners to systematically assess the incorporation of theory within existing interventions and, b) facilitate transparent reporting of the incorporation of theory within novel interventions [
69]. Using these resources will enable researchers and practitioners to overcome the limited use of theory acknowledged by the papers included in this review [
40,
41,
44,
46,
50] while still avoiding “It Seemed Like A Good Idea At The Time” ([
23], p14) interventions.
Lastly, we include a note on the role of behavioural science in combatting the COVID-19 pandemic. Helpfully, a range of prominent behavioural scientists have developed guidelines and recommendations for the application of behavioural science within the context of COVID-19. For example, Michie and colleagues advocate four principles to help reduce transmission by effecting behaviour change [
70]; the British Psychological Society have compiled a list of nine recommendations to optimise the effectiveness of changing policy and communication/ guidance [
71], and; West et al. draw upon COM-B to provide an account of components that need to be addressed in order to increase uptake of specific COVID-19 transmission-reduction behaviours [
67]. While this is by no means an exhaustive list, it does reinforce the importance of explicitly and systematically incorporating behavioural science into research and practice within the context of infectious disease and emergency response. Alongside our recommendations detailed above, we therefore further advocate that interested readers explore these principles and guidelines in more detail.
Mathematical Modellers
In our sister review [
26] we find that although the ‘gold standard’ for incorporating protective behaviour into infectious disease modelling involves the incorporation of a range of cognitive and social constructs, there is very little explicit reference to well-recognised theories of behaviour change (indeed, only five of 42 included papers make any such reference [
26]). Acknowledging the necessary trade-off between accurately modelling human behaviour and the computational demands of such modelling [
72], Weston and colleagues echo previous recommendations for modellers to familiarise themselves with the relevant behaviour change literature in order to improve their awareness of the main factors underlying human behaviour [
24,
73]. As a result, the recommendation was made for infectious disease modellers to closely consult with the psychological literature concerning the predictors of health behaviour when developing their models [
26].
Based on the outcomes of this current review, we can provide further guidance for infectious disease modellers on both where to begin with this familiarisation, and how/ where to involve behavioural scientists in the process. Firstly, in both the current review and Weston and colleagues work, the Health Belief Model is the most commonly cited behaviour change theory. We therefore agree that the Health Belief Model represents an appropriate base on which to build infectious disease models incorporating human behaviour [
26]. Nevertheless, as noted in the modelling review, there are a broad range of additional factors—including emotional and social constructs—that should be more fully considered when representing infection prevention behaviour [
26].
Given the emphasis on Protection Motivation Theory and the Extended Parallel Process Model within the literature reported in the current review, we first suggest that infectious disease modellers should consider these models alongside the Health Belief Model to help improve the modelling of emotional responding and defensive avoidance behaviour (but see also Bish & Michie for further recommendations concerning the use of parallel processing models to incorporate cognitive and emotional constructs, [
38]). Similarly, we would also recommend infectious disease modellers use Social Cognitive Theory, identified as a prominent public health behaviour change theory in the current review, as another starting point for behavioural model formulation. By more fully considering these theories and associated literature, infectious disease modellers will be well prepared to accurately and precisely model a range of relevant social, cognitive, and emotional constructs that may be associated with behavioural responses to a public health emergency.
Although this familiarisation exercise will be invaluable in helping modellers to develop a deeper understanding of the factors underlying behaviour change and is consistent with recommendations from previous literature as outlined above, it is pertinent to echo a recommendation made by Michie and colleagues in the context of intervention design. That is, it is important to ensure that the theory selected is appropriate for the type of behaviour in question. For example, if a behaviour is more likely to be influenced by habitual factors then models concerning deliberative and reflective processing are less likely to be relevant [
22]. As this current review has focused on identifying the broad state of the art for incorporating behaviour change theory within infectious disease and emergency contexts, a full consideration of the appropriate theories for each individual behaviour in each specific disease or emergency context is unfortunately outside scope. However, we recommend that infectious disease modellers work closely with behavioural scientists in the design and development of their models to ensure that the most appropriate theories are being consulted and incorporated for a given target behaviour or context. Through this greater collaboration between modellers and behavioural scientists, the discipline will be able to develop a more in-depth understanding of the requirements for behavioural theory and the computational limitations to their incorporation within infectious disease models.
Limitations and future research
Although the review reported herein represents an impressive effort at addressing a herculean task (namely, the collation of psychological behaviour change theories applied across infectious disease and emergency response contexts), as with any large-scale review there are inevitably trade-offs and potential limitations that should be considered when interpreting the outcomes and recommendations from this review.
Firstly, although other emergency contexts are represented within the current review, we acknowledge that the papers included within this review are predominantly focused on infectious diseases. Although the search strategy did include terms relating to other specific civil emergencies (e.g., terrorism), and emergencies generally (e.g., emergency response, emergency resilience), there were more search terms relating to infectious disease outbreaks/ pandemics. We therefore recommend that future reviews in this area should utilise search strategies optimised more clearly to reflect the full breadth of public health emergency contexts.
Secondly, and similarly to the first limitation, we are aware of some prominent behaviour change theories (e.g., COM-B) that are not represented in the reviews cited herein. Although this may represent a limitation of our search strategy, generic phrases and subject headings relating to theoretical models were included to ensure a breadth of focus. Furthermore, as our focus was on identifying review articles that have themselves collated primary research involving behaviour change theories in the context of infectious disease and emergency response, it follows that any prominently applied theories should have also been represented within our sample regardless of the specific search terms we used. We are therefore confident that the theories identified herein represent an accurate overview of the most commonly cited theories within this specific context over the period in question (i.e., pre COVID-19).
Thirdly, we acknowledge that several of the reviews included herein were authored by some of the same individuals [
38‐
41,
48,
50], with at least one [
39] explicitly cited as an update of another [
38]. Although this may influence the macro-representation of theories across reviews (i.e., at review-level), our decision to also examine the frequency of theory use at a micro-level (i.e., at individual cited study level, excluding repeat citations across reviews) with similar results mitigates the likely impact of this.
Nevertheless, we acknowledge that frequency of theory use – a key outcome within the current review – is not the same thing as contextual-relevance of theory within these contexts. However, the purpose of this review was to identify the most commonly employed theories of behaviour change within the infectious disease and emergency response contexts. Given this focus, we believe that the emphasis on theory frequency within the current review is well founded. Nevertheless, we do also provide a synthesis and summary of the key outcomes and conclusions in order to further guide researchers and mathematical modellers to the points of commonality and divergence within the extant review literature. We hope that this review will therefore serve as a jumping off point for further research and modelling work building on our outcomes as detailed in the preceding Recommendations section.
Finally, given the breadth of available primary literature, the proliferation of available reviews of behaviour change theories, and the specificity of our research question (i.e., to identify commonly applied theories of behaviour change within a public health emergency context) we elected to conduct a review of reviews rather than a systematic review of all literature. Similarly, although driven by pragmatic concerns, our decision to conduct our top up searches using the first 20 pages of Google Scholar searches may have limited the number of potentially relevant manuscripts for screening. Furthermore, given both the unclear relevance /lack of access to a number of papers within the current review, we elected to wholesale exclude reviews concerning sexually transmitted infection. Although these decisions and outcomes may have narrowed our focus, we argue that the close parallels between the theories identified herein and those remarked as commonplace within previous review work (e.g., [
22]) render this claim ill founded. Indeed, when combined, our top-up searches (which were sorted by relevance) allowed us to search through 600 citations published since 2016. The number of unique citations across papers included in our review (see Table
2) further suggests that we have succeeded in drawing together a broader range of literature than the independent systematic reviews themselves managed.
Nevertheless, we consider the current work to be an initial attempt at identifying and integrating the literature applying behaviour change theories in the context of infectious disease and emergency response. We therefore invite and encourage the conduct of a full and systematic review of all primary literature concerning the application of behaviour change theories across public health emergency contexts using our search strategy and extraction terms as a guide.
Conclusion
Behavioural science can play a critical role in combatting the effects of an infectious disease outbreak or public health emergency, such as the COVID-19 pandemic. However, the proliferation of available theories, with either general or specific application, can made the landscape confusing for researchers, infectious disease modellers and public health practitioners alike. In an effort to simplify the considerable behaviour change literature for ease of use by public health emergency researchers, we conducted a systematised scoping ‘review of the reviews’ concerning the application of behaviour change theories in infectious disease and emergency response emergency contexts. Our search strategies revealed 13 relevant review papers from which we were able to identify and collate the seven most commonly cited and applied behaviour change theories in this context: Health Belief Model, Theory of Planned Behaviour, and Protection Motivation Theory as most commonly applied, followed by Precaution Adoption Process Model, Extended Parallel Process Model, Theory of Reasoned Action, and Social Cognitive Theory.
Following a synthesis of the key theory-related outcomes and conclusions, we conclude that while there is broad support for the use of the most commonly cited theories within this context, the previous application of these theories within the literature is patchy. That is, much research in this context has not drawn on relevant theories of behaviour change.
Based on these identified theories and our synthesis of review outcomes, and in conjunction with a recent review by Weston and colleagues [
26], we make recommendations to assist researchers, intervention designers, and mathematical modellers to incorporate psychological behaviour change theories within infectious disease and emergency response contexts. First, we echo previous recommendations that future research and intervention design within this context should be based explicitly and systematically on relevant behaviour change theories, and in close consultation with experts in behavioural science. The theories identified herein represent an excellent starting point for this work, and we further signpost the reader to both general materials to aid in intervention design, and guiding principles for practitioners and researchers working on COVID-19. Second, we recommend that mathematical modellers should consult the theories identified herein, and work closely with behavioural scientists to familiarise themselves with the key factors underlying behaviour change within an infectious disease and emergency response context.
Considered together, the results and recommendations reported herein therefore represent an important resource to enable researchers, modellers, and practitioners working in the context of infectious disease and emergency response to better incorporate a systematic and evidence-based consideration of human behaviour into their work.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.