Purpose: Despite effective antiretroviral therapy, persons living with HIV (PWH) continue to experience a high incidence of neurocognitive deficits and comorbid disorders such as cardiovascular disease. A personalized mobile health intervention (iSTEP) aims to promote a Mediterranean diet and increase physical activity (PA), thereby improving neurocognitive functioning and cardiovascular health among PWH. This article describes a qualitative study conducted to develop iSTEP for PWH, including assessment of diet habits and preferences for tracking PA.

Methods: Two focus groups, with seven and 13 PWH respectively, discussed barriers and facilitators of a healthy diet and provided feedback to refine iSTEP components, including the feasibility of using of a Fitbit and the content of text messages designed to encourage improved diet and PA. Study participants were predominantly male (80%), ethnically diverse (non-Hispanic white, 40%; black, 40%; Hispanic, 20%), reported concerns about their diet quality, and expressed interest in improving their eating habits.Focus group transcripts were independently coded by two investigators with high agreement (kappa = 0.85), based on emergent categories related to diet habits, barriers to a healthy diet, and facilitators of a healthy diet. Data analysis was conducted with MAXQDA Analytics Pro 12.3.1 qualitative data analysis software.

Results: Focus group discussions revealed several healthy diet barriers, including housing instability, time required for food preparation, cost of healthy food, depression, gastrointestinal issues, physical disability, and changes in appetite since HIV diagnosis. Participant-identified healthy diet facilitators included affordable price options for healthy food, a structured food plan, desire to modify appearance or weight, access to medical professionals, motivation for disease prevention, and social support. Participants endorsed wearing a Fitbit on the wrist and receiving text messages as useful methods to monitor and encourage a better diet and greater physical activity.

Conclusions: Conducting effective PA and diet interventions in PWH remains challenging due to physical and socioeconomic limitations in this population. These findings assisted the development of a low cost, feasible mHealth intervention supporting a Mediterranean diet and increased PA in order to improve neurocognitive and cardiovascular outcomes among PWH.

Background: Half of smartphone users have downloaded a health and fitness app—tracking physical activity (PA) is the most common reason for downloading. However, these apps typically fail to engage users long term to attain desirable outcomes. This study examined the prevalence of, and sociodemographic predictors associated with, duration of use and non-usage attrition of a commercial PA app over 12 months.  

Methods: We analyzed first year usage data from 40,447 Canadian users who enabled the Carrot Rewards App “Steps” program between June 13 and July 10, 2016. Sociodemographic information and app usage data were automatically recorded. We measured three usage indicators: 1) duration of app use is the total number of weeks that users had non-zero step count in the weekly aggregated data; 2) non-usage attrition is the first occurrence of having zero step count for a minimum of 2 weeks; 3) resumption of app use occurs when participants resumed using the app with at least 2 weeks of non-zero step count after non-usage attrition. Kaplan-Meier survival curves plotted non-usage attrition and resumption of app use. Linear regression and cox regression models examined predictors associated with usage indicators.  

Results: Twenty-nine percent of users used the app for 52 weeks, 17.4% for 41-51weeks, 17.5% for 24-40 weeks, 17.9% for 11-23 weeks, 16.3% for 1-10 weeks, and 1.8% were non-users. 50% of users experienced non-usage attrition after 18 weeks; 50% of users resumed app use at 15 weeks after non-usage attrition. Being male (vs. female) was associated with increased duration of use. Being male, living in an area with greater proportion of people with post-secondary education or higher PA levels at baseline was associated with reduced risk of non-usage attrition and increased probability of resuming app use after a lapse.

Conclusion: Users of the Steps program showed relatively high usage with 2/3 of users used ≥6 months over 12 months. Future studies that examine reasons for discontinuing and resuming app use after a lapse may inform future PA apps in designing strategies to reduce periods of non-use (e.g, chance-based reward) and encourage return after lapse (e.g, push notification).

Purpose: The effectiveness of a smartphone app (MINISTOP 1.0) targeting Swedish-speaking parents, with the aim to improve eating and physical activity behaviours in preschool aged children, was evaluated in 2015. Interest and requests for implementation of the app into primary child health care have since then been expressed. Prior to that, the app needs to be translated and tailored for priority populations such as families with a migrant background and/or low socioeconomic position to optimize accessibility. The aim of this study was to investigate user-requirements among parents and child health care nurses, in terms of content and technical features to tailor the app (MINISTOP 2.0) for priority populations.

Methods: Pilot studies over four months in the autumn of 2019, to pre-test the content of the MINISTOP 2.0 app were conducted at the primary child health care, Linköping, and included three semi-structured focus groups with Somali-, Arabic- and Swedish speaking parents (5 parents/group), and 14 one-on-one semi-structured interviews with nurses.

Results: Preliminary results can be summarized as follows. Parents requested strategies on how to handle difficult situations around eating and screen time. A common theme for all parent groups was the need for parental strategies in situations where a child is picky or refuses to eat. Nurses expressed that parents often are more worried about their child being underweight rather than overweight, especially in the Somali- and Arabic populations. Nurses also requested clear information on portion sizes and strategies to reduce parents’ concerns about their child not eating enough.

Conclusion: Parents requested information and strategies on how to handle challenging eating situations, especially picky eating. This was independently confirmed by the nurses that requested concrete information on portion sizes and eating behaviours to reduce parental concerns.

Background: Diabetes is a significant health challenge in many developing countries with increasing prevalence and healthcare costs. Mobile phone health programs have shown to improve health outcomes at low cost. However, there is currently a paucity of information on such interventions on self-management behaviour change in a developing setting. The aim of this study is to determine the effect of a mobile phone text messaging interventions on health behaviour change in type 2 diabetic patients in  Bangladesh. 

Methods: A prospective, parallel-group, randomized, controlled clinical trial was performed at a tertiary hosptial in Dhaka, Bangladesh. Patients with type 2 diabetes were recruited and randomized to a text messaging intervention plus standard care or standard care alone. The intervention group received daily text messaging for 6 months, with the content of the messages promoting treatment adherence and healthy lifestyle habits. Health behaviours (maintaining foot hygiene; walking/exercising; measuring blood pressure; and monitoring weight) were self-reported and assessed via questionnaires taken at 6 months. Logistic regression models adjusted for age, sex, education and occupation were performed to determine the effects of text messaging on health behaviour change.

Results: A total of 236 patients were randomized to either the intervention (n=118) or standard-care (n=118). At baseline, groups did not differ in health behavioural outcomes. Individuals that received the text messaging intervention were more likely to report caring for their feet within the last week (Odds Ratio 2.2, 95% CI 1.21–3.87). No statistically significant differences were observed for the other health behaviour outcomes, as well as the total health behaviour measure.

Conclusions: Mobile phone text messaging was shown to be effective in promoting foot care among people with type 2 diabetics in our study. Larger studies with a longer longitudinal design need to be performed to obtain more reliable conclusions on health behaviour outcomes.

Purpose. A blended intervention was developed to support older adults in performing home-based exercises. A tablet and a personal coach were provided to facilitate the self-regulation of exercise behavior. The objectives of this study were a) to assess the long-term usability of the tablet and b) to explore how the intervention supported older adults in performing home-based exercises.

Methods. Mixed-methods were used. At baseline, older adults participating in the blended intervention were asked to fill out a questionnaire about their general experience with ICT devices and rate their own skill level. After 6 months participants filled out the USE-questionnaire to assess the usefulness, satisfaction and ease of use of the tablet. With a random selection of participants in-depth interviews were held to explore how the tablet and coach supported the self-regulation. The interviews were double coded and analyzed with the Directed Content Analysis method.

Results. 65 participants filled out the baseline survey, 37 participants the USE questionnaire and with 7 participants follow-up interviews were held. Their average age was 71-73 years old. The baseline questionnaire showed that the large majority already had experience with a tablet. After 6 months of use, the participants rated the usefulness, satisfaction and ease of use of the tablet as 3.8, 4.2 and 4.1 on a 5-point scale. The interviews show that participants felt that the tablet supported action planning, behavior execution and self-monitoring. On the other hand, especially during the first few months, the personal coach had added value during the goal setting, behavior execution and evaluation phases of self-regulation.      

Conclusions. The process evaluation has shown that older adults are positive about a blended intervention that was designed to support them in performing home-based exercises. They reported that the tablet helped them to perform the exercises better, more frequent and safely. It supported them in various phases of self-regulation. The availability of personal coach was nevertheless crucial. To support physical activity in older adults a blended approach is promising.

Purpose: Machine learning (ML) can be used to recognize activity patterns resulting from cross-sectional wearable data. Daily or weekly physical activity (PA) levels are made easily available from wearable tracking devices but longitudinal recognition of patterns, i.e., physical activity for sustained or discrete periods, remains a challenge, despite being the most critical PA variable to attain health benefits.

This study aims to use an ML procedure to identify long-term (12 months) PA levels in participants involved in a large weight regain prevention trial. To further explain the results, motivational variables known to be associated with sustained PA, were analyzed.

Methods: Data was gathered in the NoHoW trial conducted in Denmark, UK, and Portugal. After data cleaning, 1139 participants (68.7% Female, 44,0±11.9 y) were included in the analysis.

Minute-by-minute steps, heart rate, and PA level were gathered for 12 months, 24/7. Baseline and 12-month questionnaires assessed PA's Motivational Regulations (BREQ-3) and Basic Psychological Need Satisfaction (BPNS). A data hub stored all the information.

ML involved PCA and K-Means clustering procedures, using an unsupervised learning algorithm. Python was used to develop the ML code. Group comparison statistics looked for differences in the motivational variables.

Results: The algorithm and PCA identified two clusters, 453 Highest PA level (HPAL 39.8%) and 686 Lowest PA Level (LPAL; 60.2%). As expected, HPAL presented higher values of steps, distance and caloric expenditure, and lower heart rate (all p<.001). Autonomous motivation was higher in HPAL at 0 and 12-months, and BPNS was higher in this group at 12-months. External motivation and amotivation were higher in LPAL at 0 and 12 months (all p<.05). Autonomous motivation increased from 0-12 months in HPAL. Interestingly, autonomy decreased in this group for the same period (all p<.05).

Discussion: Efficient and scalable identification of longitudinal PA patterns remains a challenge. ML procedures, such as the one proposed in this study, are an innovative solution that, however, needs further validation. In this study, the association with motivational variables provided evidence that the unsupervised algorithm correctly identified participants' 0-12 months PA level. Deep-learning procedures are the next steps needed to move this line of research forward.

Purpose: Interventions delivered via ehealth represent a promising approach to improving dietary intake among children and adults. However, to our knowledge, no studies have investigated the acceptability of the use of ehealth apps to support parent packing of healthy lunchboxes in childcare settings. The purpose of the study is to describe the parent reported acceptability of an intervention to decrease packing of discretionary foods in child lunchboxes, delivered via an existing childcare-parent communication app.

Methods: An eleven week intervention was delivered to seventeen childcare services (eight intervention, nine controls) in the Hunter New England region of NSW, Australia as part of a randomised controlled trial. Eleven brief messages were sent to parents from the intervention services (one message per week) via push notifications delivered through an existing childcare parent communication app. Messages aimed to address known parent barriers to packing healthy lunchbox foods and were developed using the Behaviour Change Wheel framework. Consenting parents from the intervention arm were invited to participate in a seventy-five item online survey, including questions to determine acceptability and self-reported use of the app.

Results: The survey response rate was 37% (n=70). Of these, 79% (n=55) had the intervention app on their phone. The majority of parents 76% (n= 42) recalled reading any of the messages. Self-reported reading of weekly messages decreased over time with 57% (n= 24) reporting reading message 1 and just 21% (n=9) reporting reading message 11, with the average reading rate for any message being 41%. Only 33% of parents (n= 14) reporting clicking through to further links for additional information. Parents reported they felt it was appropriate to receive lunchbox information via the existing app (80%, n= 40/50), found the information easy to use 88% (n=36/41) and 76% (n=31/41) found the information helpful. Parents found the frequency and timing of the messages acceptable (100% n= 41/41). 

Conclusion: Delivery of a lunchbox intervention using a series of weekly push notifications via an existing childcare communication app appeared to be highly acceptable to parents, however reasons behind less than expected parent viewing of weekly messages requires further investigation.

Purpose: Recruiting and engaging young adults with nutrition advice is a major challenge. We explore attributes related to successful reach and user-engagement from our co-designed healthy eating website (No Money No Time).

Methods: No Money No Time (https://nomoneynotime.com.au/) hosts free resources with components targeted to young adults. Cross-disciplinary input from professionals across health, computer science, digital marketing and design have contributed to novel approaches to reach and engage young adults. Recruitment strategies include: search engine optimisation (SEO), conversion rate optimisation, public relations, referral strategy from existing media sources, re-contacting users from existing mailing database and social media promotion. Engagement strategies include: heightened user-experience and design, personalised content updated weekly, email automations personalised to individual motivators and social media promotion of trending content. Objective data from Google analytics will detail acquisition and engagement with the site in the first three months since launch (July 17^th to October 17^th 2019)

Results: Recruitment: Over 3-months there were 9469 site visitors, with half (51%) 18-34 years and mostly female (76%). Smartphones were most commonly used to access the site (64%), with user acquisition predominantly (n=3271) from direct sources (typing URL of website into search bar). This demonstrates success of our public relation strategy. This was followed by organic search (i.e., entering information into search engine such as Google) (n=2795) which demonstrates effectiveness of our SEO strategy. The outreach and referral was also effective, acquiring 2519 users; with the most common referral source being nib (health insurers). Social media promotion was effective acquiring 1243 users, with most coming from Facebook (n=1047) after the site was promoted by a podcast resource. Engagement: a total of 61,497 page views, with average session duration of 3 min 56 seconds and 4.89 pages viewed per session. A third (32%) of users completed the embedded Healthy Eating Quiz (validated tool to assess diet quality). For content, the recipe page which included a filter option by ownership of specific kitchen equipment, motivators and dietary preferences had highest engagement (43,391 page views).

Conclusion: A cross-disciplinary approach using novel strategies demonstrates potential to heighten reach and engagement of young adults with nutrition advice.

Purpose: Despite recommendations, few early childhood education and care services in Australia, and internationally, provide foods consistent with sector dietary guidelines. A web-based menu planning program (feedAustralia), was designed as an innovative method to support services with implementing dietary guidelines. This process evaluation aimed to report service adoption and use of the web-based program, fidelity to the implementation strategies, and acceptability of the intervention.

Methods: Twenty-seven services randomised to the intervention arm of a parallel group randomised controlled trial conducted in NSW, Australia received access to a web-based menu-planning program (encompassing audit and feedback, educational materials, and reminders), in addition to a range of implementation strategies delivered by a health promotion officer (including educational outreach visit, ongoing consultation, technical assistance, change equipment, and involve executive boards and consumers). Process evaluation was measured at 12 months follow-up, including i) childcare service adoption and use of the web-based program assessed via google analytics; ii) fidelity to the implementation strategies assessed via internal records maintained by the project team; and iii) acceptability of the intervention assessed via computer assisted telephone interviews with childcare service staff.

Results: At approximately 12 months’ follow-up, all intervention services had adopted the web-based program, logged in an average of 40.4 (SD 31.8) times, and spent an average of 47.1 (SD 65.2) minutes in the program per login. Fidelity to the delivery of individual implementation strategies was relatively high (>78%) across intervention services. Acceptability of the intervention was high with over 90% of services staff reporting the web-based menu planning program to be useful, and over 95% of staff reported their menus improved as a result of using the web-based program.

Conclusions: Findings indicate the innovative web-based program was highly acceptable to childcare service staff. As engagement with web-based interventions is associated with a range of health behaviours, and the extent of use varied amongst intervention services, future research exploring perceived barriers and enablers to use of the program, and identification of strategies to best support end-user engagement with the web-based program is warranted.

Purpose:

Pregnancy presents a unique opportunity to positively influence exercise behaviours. Access to information, education and monitoring through m-health, including mobile applications (apps), may facilitate behaviour change. However, app quality and the presence and appropriateness of Behaviour Change Techniques (BCTs) used in apps is unknown. This study reviewed pregnancy apps to assess their quality and the prevalence and appropriateness of BCTs.

Methods:

A review of mobile apps was conducted (Apple AppStore and GooglePlay Store) using a systematic approach. Search terms (iterations of physical activity and pregnancy) resulted in 7,207 apps; 19 apps met criteria-based eligibility (app user rating of 4.5/5 or higher, English language and specific focus on exercise during pregnancy). Each app was then allocated to two independent reviewers for user-testing and assessment. App quality was assessed using the MARS (Mobile Application Rating Scale) tool (19 items). The inclusion of BCTs (26 items) were assessed using a taxonomy of BCTs, specifically those likely to influence the main determinants of exercise behaviours during pregnancy (intention formation; goal setting; review of behavioural goals; self-monitoring; feedback and social comparison).

Results:

Of the 19 apps reviewed, most were free to download (n=18) and accessed via Apple (n=6) and GooglePlay (n=13). MARS scores (out of 5) suggested a moderate level of quality (m=3.5±0.52). Functionality (m=4.2±0.5) and aesthetics (m=3.7±0.6) were the highest scoring domains whilst subjective app quality (m=2.5±0.6) and likelihood for behavioural impact (m=2.5±0.6) were the lowest. Overall, MARS quality ratings were positively correlated with the number of MARS theoretical background/strategies (rho=0.67, p<0.01). Apps included an average of four to five BCTs (range 2-10), with the most common being ‘provide instructions’ (95%) and ‘provide information on consequences’ (89%). Eleven apps included BCTs associated with change in determinants of exercise during pregnancy (range 1-3), with the most common being ‘provide opportunities for social comparison’ (n=8) and ‘prompt self-monitoring of behaivour’ (n=7).

Conclusions:

Popular pregnancy exercise apps were user-friendly with moderate overall quality. However, the incorporation and promotion of BCTs was scarce, especially those likey to influence the main determinants of exercise behaviour. Currently available apps are unlikely to inadequately support exercise behaviours among pregnant women.