Background:People are encouraged to respond swiftly to digital health invitations, but they can be (rightly) skeptical about their legitimacy. Objective:Drawing on digital communication theory and psychological science, we hypothesized that easy-to-read web links that facilitate participants' ability to identify the health organization as the website host would improve trust in and user engagement with digital communications. Methods:In 3 double-blind, randomized online experiments, adult UK residents were recruited via the online platform Prolific (experiment 1: N1=569) or Qualtrics (experiment 2: N2=596; experiment 3: N3=1993). In each experiment, participants read a hypothetical email invitation for a COVID-19 vaccine from the UK's National Health Service (NHS). Participants reported trust in the email (primary outcome), whether the web link was easy to read, who they thought the website host was, and their intention to book an appointment via the web link. We manipulated the booking web link (between-participants, double-blind allocation via the Qualtrics randomizer). Across experiments, the control group read the email containing a deactivated NHS vaccination booking web link: "accurx.thirdparty.nhs.uk/r/aafwaczmd5." In experiment 1, participants read the email with the control link (randomized n=300, analyzed n=286) or an experimental clear link ("vaccine-booking.nhs.uk"; randomized n=301, analyzed n=283). In experiment 2, participants read the email with the control link (randomized n=203, analyzed n=201) or one of 2 experimental links: a shortened web link ("https://bit.ly/3GtTL0c"; randomized n=201, analyzed n=201) or a text-embedded link ("book here"; randomized n=196, analyzed n=194). In experiment 3, participants read the email with the control link (randomized n=655, analyzed n=655), the clear link (randomized n=669, analyzed n=668), or the text-embedded link (randomized n=670, analyzed n=670). Results:Across experiments, the control web link was poorly perceived, with most participants (729/1142, 63%) unsure or unlikely to use it. Relative to the control web link, the clear web link improved trust (β coefficientExpe1=0.26, 95% CI 0.18-0.36; β coefficientExpe3=0.24, 95% CI 0.19-0.30). The text-embedded web link also improved trust (β coefficientExpe2=0.18, 95% CI 0.09-0.27; β coefficientExpe3=0.17, 95% CI 0.11-0.22), but the shortened web link did not (β coefficientExpe2=0.01, 95% CI -0.09 to 0.11). Across experiments, improved host identification and ease of reading explained increased trust, which was significantly associated with increased booking intention. The clear and text-embedded web link (vs control) effects were robust when controlling for demographics. Conclusions:We extend digital communication research by investigating how web link design can reduce people's justified suspicion in health messaging. Moving beyond the previous research focus on health message content, we experimentally test the role of web link wording. We bring causal evidence that easier-to-read links that have easy-to-identify host health institutions increased trust and intention to use the link. We provide simple, practical design guidance to improve real-world engagement with digital health communications.

Background If the most evidence-based and effective smoking cessation apps are not selected by smokers wanting to quit, their potential to support cessation is limited. Objective This study sought to determine the attributes that influence smoking cessation app uptake and understand their relative importance to support future efforts to present evidence-based apps more effectively to maximize uptake. Methods Adult smokers from the United Kingdom were invited to participate in a discrete choice experiment. Participants made 12 choices between two hypothetical smoking cessation app alternatives, with five predefined attributes reflecting domains from the theoretical domains framework: (1) monthly price of the app (environmental resources), (2) credible source as app developer (social influence), (3) social proof as star rating (social influence), (4) app description type (beliefs about consequences), and (5) images shown (beliefs about consequences); or opting out (choosing neither app). Preferences and the relative importance of attributes were estimated using mixed logit modeling. Willingness to pay and predicted uptake of the most and least preferred app were also calculated. Results A total of 337 adult smokers completed the survey (n=168, 49.8% female; mean age 35, SD 11 years). Participants selected a smoking cessation app rather than opting out for 90% of the choices. Relative to other attributes, a 4.8-star user rating, representing social proof, was the strongest driver of app selection (mean preference parameter 2.27, SD 1.55; 95% CI 1.95-2.59). Participants preferred an app developed by health care–orientated trusted organization (credible source) over a hypothetical company (mean preference parameter 0.93, SD 1.23; 95% CI 0.72-1.15), with a logo and screenshots over logo only (mean preference parameter 0.39, SD 0.96; 95% CI 0.19-0.59), and with a lower monthly cost (mean preference parameter –0.38, SD 0.33; 95% CI –0.44 to –0.32). App description did not influence preferences. The uptake estimate for the best hypothetical app was 93% and for the worst, 3%. Participants were willing to pay a single payment of up to an additional US $6.96 (UK £5.49) for 4.8-star ratings, US $3.58 (UK £2.82) for 4-star ratings, and US $2.61(UK £2.06) for an app developed by a trusted organization. Conclusions On average, social proof appeared to be the most influential factor in app uptake, followed by credible source, one perceived as most likely to provide evidence-based apps. These attributes may support the selection of evidence-based apps.

People are encouraged to respond swiftly to digital health invitations like vaccination prompts, but they are also encouraged to question the legitimacy of digital communications. This research focuses on an overlooked aspect of digital communication that can be leveraged to foster trust and user engagement: embedded weblinks. Drawing from digital communication theory and psychological science, we posited that fluent weblinks that are easy to read and transparently identify the healthcare provider will improve trust and user engagement. In three experiments (total N = 3,183), participants read a hypothetical email from the UK’s National Health Service (NHS) inviting them to book their COVID-19 vaccination by clicking a weblink. The control invitation included the weblink initially used by the UK's national COVID vaccination services. The control weblink is compared to a clear weblink (easy to read and easy to identify host) (Experiment 1 and 3) or to concealed weblinks: a text-embedded weblink and a shortened weblink (Experiment 2 and 3). Participants reported their trust in the invitation, their perceived readability of the weblink, who they thought the website host was, and their intention to book an appointment by clicking on the weblink. Clear weblinks and text-embedded weblinks increased the identification of the host organization and were perceived as easier to read, which was associated with increased trust perception and booking intention. Shortened weblinks aided host identification but had a detrimental impact on perceived readability and did not increase trust perception or booking intention. Health-related digital communications should include readable weblinks with an identifiable host to foster trust and engagement. Where changing the weblink is impractical, concealing it via a text-embedded weblink is an effective alternative.

OBJECTIVE:We aim to identify vaccination invitations that foster trust and improve vaccination uptake overall, especially among ethnic minority groups who are more at risk from coronavirus disease (COVID-19) and less likely to be vaccinated. METHOD:In a preregistered 4 × 4 mixed-design experiment, we manipulated how much risk-benefit information the message included within-subjects and the message source between-subjects (N = 4,038 U.K. and U.S. participants, 50% ethnic minority). Participants read four vaccine invitations that varied in vaccination risk-benefit information (randomized order): control (no information), benefits only, risk and benefit, and risk and benefit that mentions vulnerable groups. The messages were sent by one of four sources (random allocation): control (health institution), medical professional (unnamed), warm and competent medical professional (unnamed), and named warm and competent medical professional (Sanjay/Lamar). Participants assessed how much they trusted the message and how likely they would be to book their vaccination appointment. RESULTS:Information about vaccination benefits and risks increased trust, especially among ethnic minority groups-for whom the effect replicated within each group. Trust also increased when the message was sent by a warm and competent medical professional relative to a health institution, but the importance of the source mattered less when more information was shared. CONCLUSIONS:Our research demonstrates the positive impact of outlining the benefits and disclosing the risks of COVID vaccines in vaccination invitation messages. Having a warm and competent medical professional source can also increase trust, especially where the message is limited in scope. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Objectives The UK government’s approach to the pandemic relies on a test, trace and isolate strategy, mainly implemented via the digital NHS Test & Trace Service. Feedback on user experience is central to the successful development of public-facing services. As the situation dynamically changes and data accumulate, interpretation of feedback by humans becomes time-consuming and unreliable. The specific objectives were to 1) evaluate a human-in-the-loop machine learning technique based on structural topic modelling in terms of its serviceability in the analysis of vast volumes of free-text data, 2) generate actionable themes that can be used to increase user satisfaction of the Service. Methods We evaluated an unsupervised Topic Modelling approach, testing models with 5-40 topics and differing covariates. Two human coders conducted thematic analysis to interpret the topics. We identified a Structural Topic Model with 25 topics and metadata as covariates as the most appropriate for acquiring insights. Results Results from analysis of feedback by 37,914 users from May 2020 to March 2021 highlighted issues with the Service falling within three major themes: multiple contacts and incompatible contact method and incompatible contact method, confusion around isolation dates and tracing delays, complex and rigid system. Conclusions Structural Topic Modelling coupled with thematic analysis was found to be an effective technique to rapidly acquire user insights. Topic modelling can be a quick and cost-effective method to provide high quality, actionable insights from free-text feedback to optimize public health services.

Antimicrobial Resistance (AMR) is a global health emergency that threatens modern medicine and incurs great cost to human health. The World Health Organization as part of a quadripartite joint initiative with the Food and Agriculture Organization of the United Nations, World Organisation for Animal Health, and United Nations Environment Programme, has recently published a One Health Priority Research Agenda for AMR. In this article we present a multidisciplinary approach, proposed by behavioural science experts, One Health experts and AMR experts to support the implementation of the Priority Research Agenda. We review, using specific examples of complex interventions designed to tackle AMR in which behavioural science has been embedded, five main steps: Define – what behaviours are a priority in each context; Diagnose - What are the barriers and enablers to the behaviours prioritised? Design - what interventions exist and what new or enhanced interventions could work to tackle the barriers identified? and, Implement and Evaluate the intervention(s). The approach presented will be useful for funders and researchers who wish to incorporate methods, frameworks and insights from the behavioural sciences into research plans, proposals and protocols in relation to a multisectoral One Health agenda and produce findings that are more relevant to policymakers.

Background The UK is rolling out a national childhood influenza immunisation programme for children, delivered through primary care and schools. Behaviourally-informed letters and reminders have been successful at increasing uptake of other public health interventions. Therefore, we investigated the effects of a behaviourally-informed letter on uptake of the vaccine at GP practices, and of a letter and a reminder (SMS/ email) on uptake at schools. Methods and results Study 1 was a cluster-randomised parallel trial of 21,786 two- and three-year olds in 250 GP practices, conducted during flu season (September to January inclusive) 2016/7. The intervention was a centrally-sent behaviourally-informed invitation letter, control was usual care. The proportion of two- and three-year olds in each practice who received a vaccination by 31st January 2017 was 23.4% in the control group compared to 37.1% in the intervention group (OR = 1.93; 95% CI = 1.82, 2.05, p < 0.001). Study 2 was a 2 (behavioural letter vs standard letter) × 2 (reminder vs no reminder) factorial trial of 1108 primary schools which included 3010 school years 1–3. Letters were sent to parents from providers, and reminders sent to parents from the schools. In the standard-letter-no-reminder arm, an average of 61.6% of eligible children in each school year were vaccinated, compared to 61.9% in the behavioural-letter-no-reminder arm, 63.5% in the standard-letter-plus-reminder arm, and 62.9% in the behavioural-letter-plus reminder condition, F (3, 2990) = 2.68, p = 0.046. In a multi-level model, with demographic variables as fixed effects, the proportion of eligible students in the school year who were vaccinated increased with the reminder, β = 0.086 (0.041), p < 0.036, but there was no effect of the letter nor any interaction effect. Conclusion Sending a behaviourally informed invitation letter can increase uptake of childhood influenza vaccines at the GP surgery compared to usual practice. A reminder SMS or email can increase uptake of the influenza vaccine in schools, but the effect size was minimal. Trial registration Study 1: Trial registration: ClinicalTrials.gov Identifier: NCT02921633. Study 2: Trial registration: ClinicalTrials.gov Identifier: NCT02883972.

IntroductionMachine-assisted topic analysis (MATA) uses artificial intelligence methods to help qualitative researchers analyze large datasets. This is useful for researchers to rapidly update healthcare interventions during changing healthcare contexts, such as a pandemic. We examined the potential to support healthcare interventions by comparing MATA with “human-only” thematic analysis techniques on the same dataset (1,472 user responses from a COVID-19 behavioral intervention).MethodsIn MATA, an unsupervised topic-modeling approach identified latent topics in the text, from which researchers identified broad themes. In human-only codebook analysis, researchers developed an initial codebook based on previous research that was applied to the dataset by the team, who met regularly to discuss and refine the codes. Formal triangulation using a “convergence coding matrix” compared findings between methods, categorizing them as “agreement”, “complementary”, “dissonant”, or “silent”.ResultsHuman analysis took much longer than MATA (147.5 vs. 40 h). Both methods identified key themes about what users found helpful and unhelpful. Formal triangulation showed both sets of findings were highly similar. The formal triangulation showed high similarity between the findings. All MATA codes were classified as in agreement or complementary to the human themes. When findings differed slightly, this was due to human researcher interpretations or nuance from human-only analysis.DiscussionResults produced by MATA were similar to human-only thematic analysis, with substantial time savings. For simple analyses that do not require an in-depth or subtle understanding of the data, MATA is a useful tool that can support qualitative researchers to interpret and analyze large datasets quickly. This approach can support intervention development and implementation, such as enabling rapid optimization during public health emergencies.

Background: The benefits of medication optimization are largely uncontroversial but difficult to achieve. Behavior change interventions aiming to optimize prescriber medication-related decisions, which do not forbid any option and that do not significantly change financial incentives, offer a promising way forward. These interventions are often referred to as nudges.Objective: The current systematic literature review characterizes published studies describing nudge interventions to optimize medication prescribing by the behavioral determinants they intend to influence and the techniques they apply.Methods: Four databases were searched (MEDLINE, Embase, PsychINFO, and CINAHL) to identify studies with nudge-type interventions aiming to optimize prescribing decisions. To describe the behavioral determinants that interventionists aimed to influence, data were extracted according to the Theoretical Domains Framework (TDF). To describe intervention techniques applied, data were extracted according to the Behavior Change Techniques (BCT) Taxonomy version 1 and MINDSPACE. Next, the recommended TDF-BCT mappings were used to appraise whether each intervention applied a sufficient array of techniques to influence all identified behavioral determinants.Results: The current review located 15 studies comprised of 20 interventions. Of the 20 interventions, 16 interventions (80%) were effective. The behavior change techniques most often applied involved prompts (n = 13). The MINDSPACE contextual influencer most often applied involved defaults (n = 10). According to the recommended TDF-BCT mappings, only two interventions applied a sufficient array of behavior change techniques to address the behavioral determinants the interventionists aimed to influence.Conclusion: The fact that so many interventions successfully changed prescriber behavior encourages the development of future behavior change interventions to optimize prescribing without mandates or financial incentives. The current review encourages interventionists to understand the behavioral determinants they are trying to affect, before the selection and application of techniques to change prescribing behaviors.Systematic Review Registration: [https://www.crd.york.ac.uk/prospero/], identifier [CRD42020168006].

BackgroundLarge-scale vaccination is fundamental to combatting COVID-19. In March 2021, the UK’s vaccination programme had delivered vaccines to large proportions of older and more vulnerable population groups; however, there was concern that uptake would be lower among young people. This research was designed to elicit the preferences of 18-29-year-olds with respect to key delivery characteristics.MethodsFrom 25 March - 2 April 2021, an online sample of 2,021 UK adults aged 18-29 years participated in a Discrete Choice Experiment. Participants made six choices, each between two SMS invitations to get vaccinated; each choice also had an opt-out. Each invitation had four attributes (1 x 5 levels, 3 x 3 levels): delivery mode, appointment timing, proximity, and SMS sender. These were systematically varied according to a d-optimal fractional factorial design. Order of presentation was randomised for each participant. Responses were analysed using a mixed logit model.ResultsThe logit model revealed a large alternative-specific constant (β = 1.385, SE = 0.067, p <0.001), indicating a strong preference for ‘opting in’ to appointment invitations. Pharmacies were dispreferred to the local vaccination centre (β = -0.256, SE = 0.072, p <0.001), appointments in locations that were 30-45 minutes travel time from one’s premises were dispreferred to locations that were less than 15 minutes away (β = -0.408, SE = 0.054, p <0.001), and, compared to invitations sent by the NHS, SMSs forwarded by ‘a friend’ were dispreferred (β = -0.615, SE = 0.056, p <0.001) but invitations from the General Practitioner were preferred (β = 0.105, SE = 0.048, p = 0.028).ConclusionsThe results indicated that the existing configuration of the UK’s mass vaccination programme was well-placed to deliver vaccines to 18-29-year-olds; however, some adjustments might enhance acceptance. Local pharmacies were not preferred; long travel times were a disincentive but close proximity (0-15 minutes from one’s premises) was not necessary; and either the ‘NHS’ or ‘Your GP’ would serve as adequate invitation sources. This research informed COVID-19 policy in the UK, and contributes to a wider body of Discrete Choice Experiment evidence on citizens’ preferences, requirements and predicted behaviours regarding COVID-19.

BackgroundThere is a need to reduce antimicrobial uses in humans. Previous studies have found variations in antibiotic (AB) prescribing between practices in primary care. This study assessed variability of AB prescribing between clinicians.MethodsClinical Practice Research Datalink, which collects electronic health records in primary care, was used to select anonymised clinicians providing 500+ consultations during 2012–2017. Eight measures of AB prescribing were assessed, such as overall and incidental AB prescribing, repeat AB courses and extent of risk-based prescribing. Poisson regression models with random effect for clinicians were fitted.Results6111 clinicians from 466 general practices were included. Considerable variability between individual clinicians was found for most AB measures. For example, the rate of AB prescribing varied between 77.4 and 350.3 per 1000 consultations; percentage of repeat AB courses within 30 days ranged from 13.1% to 34.3%; predicted patient risk of hospital admission for infection-related complications in those prescribed AB ranged from 0.03% to 0.32% (5th and 95th percentiles). The adjusted relative rate between clinicians in rates of AB prescribing was 5.23. Weak correlation coefficients (<0.5) were found between most AB measures. There was considerable variability in case mix seen by clinicians. The largest potential impact to reduce AB prescribing could be around encouraging risk-based prescribing and addressing repeat issues of ABs. Reduction of repeat AB courses to prescribing habit of median clinician would save 21 813 AB prescriptions per 1000 clinicians per year.ConclusionsThe wide variation seen in all measures of AB prescribing and weak correlation between them suggests that a single AB measure, such as prescribing rate, is not sufficient to underpin the optimisation of AB prescribing.

Online supermarket platforms present an opportunity for encouraging healthier consumer purchases. A parallel, double-blind randomised controlled trial tested whether promoting healthier products (e.g. lower fat and lower calorie) on the Sainsbury's online supermarket platform would increase purchases of those products. Participants were Nectar loyalty membership scheme cardholders who shopped online with Sainsbury's between 20th September and 10th October 2017. Intervention arm customers saw advertisement banners and recipe ingredient lists containing healthier versions of the products presented in control arm banners and ingredient lists. The primary outcome measure was purchases of healthier products. Additional outcome measures were banner clicks, purchases of standard products, overall purchases and energy (kcal) purchased. Sample sizes were small due to customers navigating the website differently than expected. The intervention encouraged purchases of some promoted healthier products (spaghetti [B = 2.10, p < 0.001], spaghetti sauce [B = 2.06, p < 0.001], spaghetti cheese [B = 2.45, p = 0.001], sour cream [B = 2.52, p < 0.001], fajita wraps [B = 2.10, p < 0.001], fajita cheese [B = 1.19, p < 0.001], bakery aisle products (B = 3.05, p = 0.003) and cola aisle products [B = 0.97, p < 0.002]) but not others (spaghetti mince, or products in the yogurt and ice cream aisles). There was little evidence of effects on banner clicks and energy purchased. Small sample sizes may affect the robustness of these findings. We discuss the benefits of collaborating to share expertise and implement a trial in a live commercial environment, alongside key learnings for future collaborative research in similar contexts.

Behavioural change with societal transformation has been the key processes whereby hand and respiratory hygiene, social distancing and self-isolation that citizens across the world have been asked to implement to respond to the global COVID-19 pandemic. Is it possible to use such societal transformation approaches to change our behaviour for climate change adaptation? The European Commission (EC) funded research and innovation programmes that will be launched from 2021 will mobilise investment and EC's wide efforts to achieve measurable and time-bound goals on issues that affect citizens' daily lives. These programmes are based around five missions, one of which is the Mission on Adaptation to climate change including societal transformation. This will provide an opportunity to build evidence-informed assessment and design of interventions and should use a systems approach to determine and deploy the most cost-effective mix of public health behaviour change policy options according to the Nuffield Intervention Ladder and the Behaviour Change Wheel. This will maximise the likelihood of delivering societal transformation actions through ambitious but realistic research and innovation activities to help deliver planetary health programmes for Europe more widely.

Abstract Background Sending a social norms feedback letter to general practitioners who are high prescribers of antibiotics has been shown to reduce antibiotic prescribing. The 2017-9 Quality Premium for primary care in England sets a target for broad-spectrum prescribing, which should be at or below 10% of total antibiotic prescribing. We tested a social norm feedback letter that targeted broad-spectrum prescribing and the addition of a chart to a text-only letter that targeted overall prescribing. Methods We conducted three 2-armed randomised controlled trials, on different groups of practices: Trial A compared a broad-spectrum message and chart to the standard-practice overall prescribing letter (practices whose percentage of broad-spectrum prescribing was above 10% and who had relatively high overall prescribing). Trial C compared a broad-spectrum message and a chart to a no-letter control (practices whose percentage of broad-spectrum prescribing was above 10% and who had relatively moderate overall prescribing). Trial B compared an overall-prescribing message with a chart to the standard practice overall letter (practices whose percentage of broad-spectrum prescribing was below 10% but who had relatively high overall prescribing). Letters were posted to general practitioners, timed to be received on 1 November 2018. The primary outcomes were practices’ percentage of broad-spectrum prescribing (trials A and C) and overall antibiotic prescribing (trial B) each month from November 2018 to April 2019 (all weighted by the number and characteristics of patients registered in the practice). Results We randomly assigned 1909 practices; 58 closed or merged during the trial, leaving 1851 practices: 385 in trial A, 674 in trial C, and 792 in trial B. AR(1) models showed that there were no statistically significant differences in our primary outcome measures: trial A β = − .199, p = .13; trial C β = .006, p = .95; trial B β = − .0021, p = .81. In all three trials, there were statistically significant time trends, showing that overall antibiotic prescribing and total broad-spectrum prescribing were decreasing. Conclusion Our broad-spectrum feedback letters had no effect on broad-spectrum prescribing; adding a bar chart to a text-only letter had no effect on overall antibiotic prescribing. Broad-spectrum and overall prescribing were both decreasing over time. Trial registration ClinicalTrials.gov NCT03862794. March 5, 2019.

This systematic review and intervention content analysis used behavioural science frameworks to characterise content and function of interventions targeting supermarket shoppers' purchasing behaviour, and explore if coherence between content and function was linked to intervention effectiveness. Study eligibility: in-store interventions (physical supermarkets) with control conditions, targeting objectively measured food and/or non-alcoholic drink purchases, published in English (no date restrictions). Eleven electronic databases were searched; reference lists of systematic reviews were hand-searched. Methodological quality was assessed using the GATE checklist. A content analysis was performed to characterise intervention content and function, and theoretical coherence between these, using the Behaviour Change Wheel, Behaviour Change Techniques Taxonomy, and Typology of Interventions in Proximal Physical Micro-Environments (TIPPME). Forty-six articles (49 interventions) met inclusion criteria; 26 articles (32 interventions) were included in the content analysis. Twenty behaviour change techniques (BCTs), and four TIPPME intervention types were identified; three BCTs ('Prompts/cues', 'Material incentive', and 'Material reward') were more common in effective interventions. Nineteen interventions solely employed theoretically appropriate BCTs. Theoretical coherence between BCTs and intervention functions was more common in effective interventions. Effective interventions included price promotions and/or in-store merchandising. Future research should explore the effect of specific BCTs using factorial study designs. PROSPERO Registration: CRD42017071065.

Background Machine-assisted topic analysis (MATA) uses artificial intelligence methods to assist qualitative researchers to analyse large amounts of textual data. This could allow qualitative researchers to inform and update public health interventions ‘in real-time’, to ensure they remain acceptable and effective during rapidly changing contexts (such as a pandemic). In this novel study we aimed to understand the potential for such approaches to support intervention implementation, by directly comparing MATA and ‘human-only’ thematic analysis techniques when applied to the same dataset (1472 free-text responses from users of the COVID-19 infection control intervention ‘Germ Defence’). Methods In MATA, the analysis process included an unsupervised topic modelling approach to identify latent topics in the text. The human research team then described the topics and identified broad themes. In human-only codebook analysis, an initial codebook was developed by an experienced qualitative researcher and applied to the dataset by a well-trained research team, who met regularly to critique and refine the codes. To understand similarities and difference, formal triangulation using a ‘convergence coding matrix’ compared the findings from both methods, categorising them as ‘agreement’, ‘complementary’, ‘dissonant’, or ‘silent’. Results Human analysis took much longer (147.5 hours) than MATA (40 hours). Both human-only and MATA identified key themes about what users found helpful and unhelpful (e.g. Boosting confidence in how to perform the behaviours vs Lack of personally relevant content ). Formal triangulation of the codes created showed high similarity between the findings. All codes developed from the MATA were classified as in agreement or complementary to the human themes. Where the findings were classified as complementary, this was typically due to slightly differing interpretations or nuance present in the human-only analysis. Conclusions Overall, the quality of MATA was as high as the human-only thematic analysis, with substantial time savings. For simple analyses that do not require an in-depth or subtle understanding of the data, MATA is a useful tool that can support qualitative researchers to interpret and analyse large datasets quickly. These findings have practical implications for intervention development and implementation, such as enabling rapid optimisation during public health emergencies. Contributions to the literature ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement The study was funded by United Kingdom Research and Innovation Medical Research Council (UKRI MRC) Rapid Response Call: UKRI CV220-009. The Germ Defence intervention was hosted by the Lifeguide Team, supported by the NIHR Biomedical Research Centre, University of Southampton. LY is a National Institute for Health Research (NIHR) Senior Investigator and team lead for University of Southampton Biomedical Research Centre. LY is affiliated to the National Institute for Health Research Health Protection Research Unit (NIHR HPRU) in Behavioural Science and Evaluation of Interventions at the University of Bristol in partnership with Public Health England (PHE). The views expressed are those of the author(s) and not necessarily those of the NIHR, the Department of Health or PHE. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Ethics committee of University of Southampton gave ethical approval for this work I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable. Yes The datasets generated and/or analysed during the current study are available in the figshare repository, . * AI : Artificial intelligence IPA : Interpretative phenomenological analysis MATA : Machine-assisted topic analysis NLP : Natural language processing PBA : Person based approach STM : Structural topic model TA : Thematic analysis

Risk perceptions are important influences on health behaviours. We used descriptive statistics and multivariable logistic regression models to assess cross-sectionally risk perceptions for severe Covid-19 symptoms and their health behaviour correlates among 2206 UK adults from the HEBECO study. The great majority (89-99%) classified age 70+, having comorbidities, being a key worker, overweight, and from an ethnic minority as increasing the risk. People were less sure about alcohol drinking, vaping, and nicotine replacement therapy use (17.4-29.5% responding 'don't know'). Relative to those who did not, those who engaged in the following behaviours had higher odds of classifying these behaviours as (i) decreasing the risk: smoking cigarettes (adjusted odds ratios, aORs, 95% CI = 2.26, 1.39-3.37), and using e-cigarettes (aORs = 5.80, 3.25-10.34); (ii) having no impact: smoking cigarettes (1.98; 1.42-2.76), using e-cigarettes (aORs = 2.63, 1.96-3.50), drinking alcohol (aORs = 1.75, 1.31-2.33); and lower odds of classifying these as increasing the risk: smoking cigarettes (aORs: 0.43, 0.32-0.56), using e-cigarettes (aORs = 0.25, 0.18-0.35). Similarly, eating more fruit and vegetables was associated with classifying unhealthy diet as 'increasing risk' (aOR = 1.37, 1.12-1.69), and exercising more with classifying regular physical activity as 'decreasing risk' (aOR = 2.42, 1.75-3.34). Risk perceptions for severe Covid-19 among UK adults were lower for their own health behaviours, evidencing optimism bias. These risk perceptions may form barriers to changing people's own unhealthy behaviours, make them less responsive to interventions that refer to the risk of Covid-19 as a motivating factor, and exacerbate inequalities in health behaviours and outcomes.