Applying machine-learning to rapidly analyze large qualitative text datasets to inform the COVID-19 pandemic response: comparing human and machine-assisted topic analysis techniques.

Introduction: Machine-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). Methods: In 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". Results: Human 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. Discussion: Results 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.

Barriers to and Facilitators of the Use of Digital Tools in Primary Care to Deliver Physical Activity Advice: Semistructured Interviews and Thematic Analysis.

BACKGROUND: Physical inactivity is a leading risk factor for many health conditions, including cardiovascular disease, diabetes, and cancer; therefore, increasing physical activity (PA) is a public health priority. Health care professionals (HCPs) in primary care are pivotal in addressing physical inactivity; however, few HCPs provide PA advice to patients. There can be obstacles to delivering PA advice, including lack of time, confidence, or knowledge. Digital technology has the potential to overcome obstacles and facilitate delivering PA advice. However, it is unknown if and how digital tools are used to deliver PA advice in primary care consultations and what factors influence their use. OBJECTIVE: We aimed to understand the use of digital tools to support primary care consultations and to identify the barriers to and facilitators of using these systems. METHODS: Overall, 25 semistructured interviews were conducted with primary care HCPs. Professionals were sampled based on profession (general practitioners, practice nurses, and health care assistants), prevalence of long-term conditions within their practice area, and rural-urban classification. The data were analyzed thematically to identify the influences on the use of digital tools. Themes were categorized using the COM-B (capability, opportunity, and motivation-behavior) model and the Theoretical Domains Framework to identify the barriers to and facilitators of using digital tools to support the delivery of PA advice in primary care consultations. RESULTS: The identified themes fell within 8 domains of the Theoretical Domains Framework. The most prominent influence (barrier or facilitator) within psychological capability was having the skills to use digital tools. Training in the use of digital tools was also mentioned several times. The most notable influences within physical opportunity were limited digital tools to prompt/support the provision of PA advice, time constraints, efficiency of digital tools, simplicity and ease of use of digital tools, and integration with existing systems. Other physical opportunity influences included lack of access to digital tools and technical support in the use of digital tools. Within social opportunity, a notable barrier was that digital tools reduce interpersonal communication with patients. Patient preference was also identified. Several important influences were within reflective motivation, including confidence to use digital tools, beliefs about the usefulness of digital tools, the belief that digital tools "are the way forward," beliefs related to data privacy and security concerns, and perceptions about patient capabilities. About automatic motivation, influences included familiarity and availability regarding digital tools and the fact that digital tools prompt behavior. CONCLUSIONS: A variety of influences were identified on the use of digital tools to support primary care consultations. These findings provide a foundation for designing a digital tool addressing barriers and leverages the facilitators to support PA advice provision within primary care to elicit patient behavior change and increase PA.

Interventions to change purchasing behaviour in supermarkets: a systematic review and intervention content analysis.

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.

Content and Mechanism of Action of National Antimicrobial Stewardship Interventions on Management of Respiratory Tract Infections in Primary and Community Care.

A major modifiable factor contributing to antimicrobial resistance (AMR) is inappropriate use and overuse of antimicrobials, such as antibiotics. This study aimed to describe the content and mechanism of action of antimicrobial stewardship (AMS) interventions to improve appropriate antibiotic use for respiratory tract infections (RTI) in primary and community care. This study also aimed to describe who these interventions were aimed at and the specific behaviors targeted for change. Evidence-based guidelines, peer-review publications, and infection experts were consulted to identify behaviors relevant to AMS for RTI in primary care and interventions to target these behaviors. Behavior change tools were used to describe the content of interventions. Theoretical frameworks were used to describe mechanisms of action. A total of 32 behaviors targeting six different groups were identified (patients; prescribers; community pharmacists; providers; commissioners; providers and commissioners). Thirty-nine interventions targeting the behaviors were identified (patients = 15, prescribers = 22, community pharmacy staff = 8, providers = 18, and commissioners = 18). Interventions targeted a mean of 5.8 behaviors (range 1-27). Influences on behavior most frequently targeted by interventions were psychological capability (knowledge and skills); reflective motivation (beliefs about consequences, intentions, social/professional role and identity); and physical opportunity (environmental context and resources). Interventions were most commonly characterized as achieving change by training, enabling, or educating and were delivered mainly through guidelines, service provision, and communications & marketing. Interventions included a mean of four Behavior Change Techniques (BCTs) (range 1-14). We identified little intervention content targeting automatic motivation and social opportunity influences on behavior. The majority of interventions focussed on education and training, which target knowledge and skills though the provision of instructions on how to perform a behavior and information about health consequences. Interventions could be refined with the inclusion of relevant BCTs, such as goal-setting and action planning (identified in only a few interventions), to translate instruction on how to perform a behavior into action. This study provides a platform to refine content and plan evaluation of antimicrobial stewardship interventions.

Prescriber Commitment Posters to Increase Prudent Antibiotic Prescribing in English General Practice: A Cluster Randomized Controlled Trial.

Unnecessary antibiotic prescribing contributes to Antimicrobial Resistance posing a major public health risk. Estimates suggest as many as half of antibiotics prescribed for respiratory infections may be unnecessary. We conducted a three-armed unblinded cluster randomized controlled trial (ISRCTN trial registry 83322985). Interventions were a commitment poster (CP) advocating safe antibiotic prescribing or a CP plus an antimicrobial stewardship message (AM) on telephone appointment booking lines, tested against a usual care control group. The primary outcome measure was antibiotic item dispensing rates per 1000 population adjusted for practice demographics. The outcome measures for post-hoc analysis were dispensing rates of antibiotics usually prescribed for upper respiratory tract infections and broad spectrum antibiotics. In total, 196 practice units were randomized to usual care (n = 60), CP (n = 66), and CP&AM (n = 70). There was no effect on the overall dispensing rates for either interventions compared to usual care (CP 5.673, 95%CI -9.768 to 21.113, p = 0.458; CP&AM, -12.575, 95%CI -30.726 to 5.576, p = 0.167). Secondary analysis, which included pooling the data into one model, showed a significant effect of the AM (-18.444, 95%CI -32.596 to -4.292, p = 0.012). Fewer penicillins and macrolides were prescribed in the CP&AM intervention compared to usual care (-12.996, 95% CI -34.585 to -4.913, p = 0.018). Commitment posters did not reduce antibiotic prescribing. An automated patient antimicrobial stewardship message showed effects and requires further testing.

Relationship between popularity and the likely efficacy: an observational study based on a random selection on top-ranked physical activity apps.

OBJECTIVES: To explore the relationship between popularity of mobile application (apps) for physical activity (PA) and their likely efficacy. The primary objective was to assess the association between app popularity (indicated by user ratings) and likely efficacy (indicated by the number of Behaviour Change Techniques (BCT) present). The secondary objective was to assess the relationship between user ratings and those BCTs that have been shown to be effective in increasing PA. DESIGN: Observational study. METHODS: 400 top-ranked free and paid apps from iTunes and Google Play stores were screened, and were included if the primary behaviour targeted was PA and they had stand-alone functionality. The outcome variable of user rating was dichotomised into high (4, 5 stars) or low (1, 2, 3 stars) rating. SETTING: iTunes and Google Play app stores. PARTICIPANTS: No individual participants but the study used user-led rating system in the app store. PRIMARY AND SECONDARY OUTCOME MEASURES: BCTs and user rating. RESULTS: Of 400 apps, 156 were eligible and 65 were randomly selected, downloaded and assessed by two reviewers. There was no relationship overall between star ratings and the number of BCTs present, nor between star ratings and the presence of BCTs known to be effective in increasing PA. App store was strongly associated with star ratings, with lower likelihood of finding 4 or 5 stars in iTunes compared with Google Play (OR 0.74, 95% CI 0.73 to 0.76, p<0.001). CONCLUSIONS: The findings of this study suggest that popularity does not necessarily imply the likelihood of effectiveness. Hence, public health impact is unlikely to be achieved by allowing market forces to 'prescribe' what is used by the public.

Implementation of the goal-setting components in popular physical activity apps: Review and content analysis.

OBJECTIVE: Goal setting is an effective strategy to promote physical activity. Commercial apps that tackle physical activity often include goal setting; however, it is unknown whether the implementation of the goal-setting components is congruent with the theory. This study evaluated the quality of goal setting in popular free and paid physical activity apps by assessing the presence of effective goal-setting components. METHODS: A six-item scale was developed based on the goal-setting literature and used for coding each app for the presence/absence of goal-setting components (i.e. specificity, difficulty, action planning, timeframe, goal evaluation and goal re-evaluation). Cohen's Kappa was used to evaluate inter-rater reliability for each scale item. The number of goal-setting components included in the 40 apps was calculated and the difference between free and paid apps was assessed. RESULTS: All scale items achieved satisfactory inter-rater reliability except 'goal evaluation'. The most frequently included goal-setting components in popular physical activity apps were 'goal specificity' (95% of the apps) and 'goal timeframe' (67.5%). Conversely, only 47.5% and 25% of the apps implemented 'action planning' and 'goal difficulty', respectively, and none included 'goal re-evaluation'. No differences emerged between free and paid apps. CONCLUSIONS: The quality of the goal-setting strategy in popular physical activity apps could be improved by introducing components scarcely implemented to date. In particular, tailoring the goal difficulty to the users' ability level and re-evaluating the goals based on achievements should be implemented to increase the quality of goal setting.

Pharmacological treatments for generalised anxiety disorder: a systematic review and network meta-analysis.

BACKGROUND: Generalised anxiety disorder is a disease that can be associated with substantial dysfunction. Pharmacological treatment is often the first choice for clinicians because of the cost and resource constraints of psychological alternatives, but there is a paucity of comparative information for the multiple available drug choices. METHODS: A systematic review and network meta-analysis was performed on randomised trials in adult outpatients with generalised anxiety disorder identified from MEDLINE, Web of Science, Cochrane Library, ClinicalTrials.gov, Chinese National Knowledge Infrastructure (CNKI), Wanfang data, Drugs@FDA and commercial pharmaceutical registries. Placebo and active control trials were included. Data were extracted from all manuscripts and reports. Primary outcomes were efficacy (mean difference [MD] in change in Hamilton Anxiety Scale Score) and acceptability (study discontinuations for any cause). We estimated summary mean treatment differences and odds ratios using network meta-analyses with random effects. This study is registered with PROSPERO, number CRD42018087106. FINDINGS: Studies were published between Jan 1, 1994 and Aug 1, 2017, in which 1992 potential studies were screened for inclusion. This analysis is based on 89 trials, which included 25 441 patients randomly assigned to 22 different active drugs or placebo. Duloxetine (MD -3·13, 95% credible interval [CrI] -4·13 to -2·13), pregabalin (MD -2·79, 95% CrI -3·69 to -1·91), venlafaxine (MD -2·69, 95% CrI -3·50 to -1·89), and escitalopram (MD -2·45, 95% CrI -3·27 to -1·63) were more efficacious than placebo with relatively good acceptability. Mirtazapine, sertraline, fluoxetine, buspirone, and agomelatine were also found to be efficacious and well tolerated but these findings were limited by small sample sizes. Quetiapine (MD -3·60 95% CrI -4·83 to -2·39) had the largest effect on HAM-A but it was poorly tolerated (odds ratio 1·44, 95% CrI 1·16-1·80) when compared with placebo. Likewise, paroxetine and benzodiazepines were effective but also poorly tolerated when compared with placebo. Risk of reporting bias was considered low, and when possible all completed studies were included to avoid publication bias. INTERPRETATION: To our knowledge, this is the largest contemporary review of pharmacological agents for the treatment of generalised anxiety disorder by use of network analysis. There are several effective treatment choices for generalised anxiety disorder across classes of medication. The failure of initial pharmacological therapy might not be a reason to abandon a pharmacological treatment strategy. FUNDING: No funding was received for this research.

Quality of Publicly Available Physical Activity Apps: Review and Content Analysis.

BACKGROUND: Within the new digital health landscape, the rise of health apps creates novel prospects for health promotion. The market is saturated with apps that aim to increase physical activity (PA). Despite the wide distribution and popularity of PA apps, there are limited data on their effectiveness, user experience, and safety of personal data. OBJECTIVE: The purpose of this review and content analysis was to evaluate the quality of the most popular PA apps on the market using health care quality indicators. METHODS: The top-ranked 400 free and paid apps from iTunes and Google Play stores were screened. Apps were included if the primary behavior targeted was PA, targeted users were adults, and the apps had stand-alone functionality. The apps were downloaded on mobile phones and assessed by 2 reviewers against the following quality assessment criteria: (1) users' data privacy and security, (2) presence of behavior change techniques (BCTs) and quality of the development and evaluation processes, and (3) user ratings and usability. RESULTS: Out of 400 apps, 156 met the inclusion criteria, of which 65 apps were randomly selected to be downloaded and assessed. Almost 30% apps (19/65) did not have privacy policy. Every app contained at least one BCT, with an average number of 7 and a maximum of 13 BCTs. All but one app had commercial affiliation, 12 consulted an expert, and none reported involving users in the app development. Only 12 of 65 apps had a peer-reviewed study connected to the app. User ratings were high, with only a quarter of the ratings falling below 4 stars. The median usability score was excellent-86.3 out of 100. CONCLUSIONS: Despite the popularity of PA apps available on the commercial market, there were substantial shortcomings in the areas of data safety and likelihood of effectiveness of the apps assessed. The limited quality of the apps may represent a missed opportunity for PA promotion.