Purpose: Measuring food choices in health and nutrition research is challenging. Given the trend that more and more research is conducted online, there is a strong need for reliable and valid measures that can be efficiently used in online research. Therefore, we developed the Multiple Food Test as a tool for measuring food choices in online and laboratory settings.

Methods: The Multiple Food Test is a picture-based selection task consisting of 18 trails. In each trial, participants see four food items (ranging from 1 = unhealthy to 4 = very healthy) and are asked to indicate which of the depicted foods they would choose to eat. Across all 18 trails a mean is computed that informs about the healthiness of choices. In three studies we examined the validity and reliability of the Multiple Food Test. In Study 1 (N = 425), the relationship between choices in the Multiple Food Test and health- and eating-related variables was assessed. Study 2 (N = 201) tested the validity of the Multiple Food Test via assessing its relationship to existing scales to measure related constructs. In Study 3 (N = 40), it was tested whether choices in the Multiple Food Test can predict real choice behavior.

Results: Study 1 revealed that healthier choices in the Multiple Food Test are related to viewing health as a changeable construct (p < .05), eating healthily more often (p < .01), viewing a healthy diet as more important (p < .01) , and holding a stronger health versus taste motive (p < .001). In Study 2, positive relationships with self-control (p < .05) and habitual fruit (p < .001) and vegetable consumption (p < .05) were found. Study 3 showed that choices in the Multiple Food Test predict real food choice behaviour (p < .001).

Conclusions: The Multiple Food Test offers new opportunities to study variables or to evaluate interventions that influence food choices and eating behavior. Furthermore, it can be easily adjusted to cultural eating differences. Because the Multiple Food Test is nearly language free it can also be used to study children’s food choices.

Purpose: Poor dietary habits are established in childhood and often persist into adulthood. As higher levels of nutrition knowledge has been found to be related to healthier eating patterns, there is a pressing need to enhance children’s nutrition knowledge in Australia. However, interventions concerned with nutrition knowledge improvement need appropriate measurement tools to test their effectiveness. Unfortunately, few reliable and valid measurement tools exist and the aim of the current study was to develop and test a nutrition knowledge questionnaire for Australian children.

Methods: A team of nutrition and education experts developed a nutrition knowledge questionnaire for Australian children aged 8-12 years (the CNK-AU), based on an existing validated Belgian survey. The questions were translated and adapted to align with Australian culture and national dietary recommendations. The CNK-AU was test for content-, internal- and external validity among primary school children aged 8 to 12.

Results: The new tool consists of eight nutrition-related topics (i.e. healthy choices, AGHE serves, balanced meals, portion sizes, food safety, -groups, -functions and -sources). Year 5/6 students at a primary school in Newcastle, Australia were recruited and completed the nutrition knowledge survey twice; at baseline (T0: n=186, age=10.9, SD=0.76, 56% female) and after one week (T1: n=94, age=10.9, SD=0.76, 53% female). Teachers were asked not to teach any nutrition during the study. It was possible to pair 94 of the participants with their results on the T0 CNK-AU, as the remaining participants were sorted into the intervention arm of a simultaneous nutrition education program. Content was validated in a previous pre-test, item-analyses found varied item difficulty and discrimination, the internal reliability was low to moderate for all categories and the total score. A test-retest showed all (except ‘AGHE serves’) categories and total score of T0 and T1 to be significantly correlated (ΔM=0.646, r=0.756,p<0.001), confirming the test-retest validity of the survey.

Conclusions: Using the item analysis results, future studies should establish a shorter version of the CNK-AU and in addition test the CNK-AU on a younger population. The test-retest shows the CNK-AU is reliable, but indicated AGHE serves as a gap in the children’s knowledge.

Background: Social network theory offers explanations for how relationships among actors influence action. Effective policies to prevent obesity and achieve the 2013 World Health Organization goal to reduce non-communicable diet-related (NCD) mortality rates by 25% by 2025 are dependent on social networks composed actors able to enact power and positively influence the policy making process. However, there is limited evidence on the nature of the connections among actors involved in food and nutrition policies. The goal of this study is to explore the dimensions of power in the social networks of actors who influence the policy making process to restrict unhealthy food and beverage products to children in Mexico.

Methods: The study uses the Gaventa’s powercube framework developed by the Institute of Development Studies. This research collects a social network data using semi-structured interviews based on a snowball sampling method to identify relevant actors in networks of interest. For network data, we conduct 2-mode analysis to examine which actors are most powerful and how correlated various power networks are with one another. To analyze we use measures of centrality (in-degree, betweenness, and eigenvector) and Quadratic Assignment Procedures (QAP). To visualize the networks structures and actor relationships graphs are generated using UCINET software package.

Results: The results map actors into power networks (e.g. levels, spaces and forms of power) and connectedness networks (e.g. relationships). This study highlights the dynamics of power in the policy making process between and within actors from Mexican government, academia, civil society, private foundations, international non-governmental organizations, media, and food and beverage firms or industry trade associations. This network analysis identifies interactions and power dynamics between the actors that may influence decision-makers and the policy process.

Conclusion: Social network analysis provides a powerful method that can be used in policy research to improve nutrition, physical activity, and sedentary behaviors among populations. This study provides insights to help policymakers and other actors to develop strategies to increase support to effectively implement supportive policies that protects children, prevent and manage obesity and NCDs .

Purpose: Several countries established weight control registries, with the aim of identifying the main sociodemographic, psychological and behavioral characteristics of individuals successful at weight loss maintenance in their contexts. The purpose of this study is to identify and systematically synthesize the correlates of the magnitude of weight loss maintenance (WLM) in successful weight loss maintainers. 

Methods: A comprehensive search of peer-reviewed articles was conducted in three electronic databases: PubMed, Web of Science and SCOPUS (all articles published until November 2018). Searches included various combinations of the following terms: weight loss maintenance, weight control registry, national weight control registry, Portuguese, Greek, Finnish, German. Studies were selected for this review if the sample included participants of weight control registries. The MOOSE guidelines were used to extract relevant information. The methodological quality of the studies was assessed using the Effective Public Health Practice Project – Quality Assessment Tool for Quantitative Studies. All associations identified through Pearson and Spearman correlations were considered. In some specific cases, linear and multiple regressions, odds-ratio/relative risk and X² (with post-hoc tests) were also considered. 

Results: The search yielded 159 potentially relevant studies after title and abstract screening. After full-text screening, 52 articles met the eligible criteria and were included. Thirty sociodemographic, behavioral and psychological correlates of the magnitude of WLM were identified across 16 studies. Age was the only sociodemographic variable showing an association (negative) with the magnitude of WLM. Physical activity was the most frequently studied correlate (k=5) with all of the studies reporting a positive association with the magnitude of WLM. Energy intake and fat intake were identified as negative correlates (k=3), contrary to protein intake (k=1). General (k=3) and internal eating disinhibition (k=2) were identified as psychological negative correlates. 

Conclusions: Behavioral and psychological factors seem to play a key role for maintaining weight loss, contrary to sociodemographic variables. Systematically identifying and summarizing the sociodemographic, behavioral and psychological correlates of the magnitude of WLM in successful weight loss maintainers can contribute to inform future obesity prevention and treatment initiatives towards better long-term outcomes.

Methods: A participatory approach with communities in Tanzania and Malawi to inform a new arts-based methodology (ABM) for exploring local socio-cultural contexts around NCDs/NCD risk factors. The ABM was piloted in four creative workshops (CWs: each N=7-9, men and women separately) involving three activities: verbalised senses, embodied images and performance. We also held four traditional focus group discussions (FGDs: each N=8-10, men and women separately).  Thematic analysis of researcher reflections, fieldnotes and written transcripts compared the two methods.

Results: Workshop participants chose to explore diabetes (Tanzania men and women), cancer (Malawi women), and hypertension (Malawi men) as NCDs of major concern locally. ABM activities allowed participants to depict local practices and beliefs around food and drink (including alcohol), physical (in)activity and obesity.  When compared to FGDs, the CWs appeared more inclusive, promoted engagement and supported community members to direct the enquiry. Through the ABM, participants expressed values, sentiments and fears in relation to NCDs (“Snake’s poison is similar to diabetes’ poison” CW_men_Tanzania) more freely than in FGDs. However, these (often negative) expressions may precipitate stigma and blame.

Conclusions: The ABM provides a promising approach to inform development of socio-culturally appropriate NCD prevention interventions, but further reseach is needed and careful attention required to minimize risk of unintended consequences.

Purpose  A higher prevalence of obesity in men compared to women, particularly amongst those 18 to 40 years of age, has been observed in Singapore. We hypothesized that poor adaptation during the transition from obligatory military training to civilian life in our young men may contribute to weight gain and sought to understand the underlying drivers.

Methods  In-depth interviews were conducted with 26 healthy, normal weight young men, who completed military service the previous year. The interviews were guided by the health belief model to elicit perceived factors influencing weight change during the transition period, and barriers and motivators in weight management. Data was collected and cross-checked by two researchers who analysed using the thematic analysis approach.

Results   The participants generally perceived themselves to be less fit and healthy compared to being in army and felt that changes in their physical activity levels and eating behaviors were contributing to their perceived weight gain. They, however do not have immediate concerns about their weight changes, citing that effects were reversible or only observed later in life.

While most were aware of the benefits of regular physical activity, benefits of healthy eating were less mentioned.  Many do not view these lifestyle behaviors as current life priorities, citing reasons such as a lack of time, poor fit into current lifestyle or having little self-motivation. Other barriers included peer influence on food choices and a lack of common leisure time to exercise with peers. Motivators for these lifestyle behaviors were mostly intrinsic, including interest in a particular sport, making a personal choice and having personal fitness goals.  The need to compensate for the lack of physical activity or unhealthy food consumption through healthy eating was also mentioned. Extrinsic motivators were related to peer influence, accessibility to sports facilities and healthy food options and monetary incentives.

Conclusions There remains a need to educate our young men, particularly correcting their misconceptions about lifestyle, weight and health. A healthy environment and appropriate incentives would be important to facilitate lifestyle changes in these men, who are at risk of weight gain during this transition phase.

Purpose: UPnGO with ParticipACTION (UPnGO) is a 1-year Canadian workplace physical activity (PA) intervention to increase habitual PA (steps) at work. Core intervention components include (1) self-monitoring of steps and action planning behaviours using a web/mobile app with incentives and (2) organizational support, including senior management’s role modeling and program endorsement. After piloting, UPnGO was made available nationally to companies for a fee (~$100/employee).  The purpose of this study is to describe the process evaluation of the commercialization of UPnGO.

Methods: The process evaluation was guided by the RE-AIM framework. Platform analytics were used to determine registration, engagement, and PA through daily step counts. Off-platform surveys collected self-reported PA at baseline and 12 months of the program. Field notes were taken by the evaluation team during the program.

Results: Reach: In the first 2 years, 17 companies from 6 provinces paid to participate in UPnGO.  These companies varied in size (30-4100 employees) and structure including municipalities, health authorities, and corporate offices. Engagement: Across these companies, 1362 participants registered for UPnGO, representing 21% of 6393 total employees. At baseline, 69% of participants did not meet PA guidelines and 46% of respondents sat for ≥6 hours during the workday. Adoption: Significant declines were observed in daily step count at 12 months (Baseline: 7879±3802 steps, 12-Months: 7131±3615 steps; p=0.002).  Program engagement, measured by unique platform log-ins, averaged 49% of users over 12 months; however, engagement declined from 94% in Month 1 to 31% in Month 12.  Monthly step tracking rates ranged between 32%-43% of participants. Implementation: Evaluator observations revealed senior leadership support, identified as a key factor in the success of the pilot, was not tracked in the national implementation. Company expectations for a purchased product did not meet available resources for supporting implementation. Maintenance: Two organizations renewed their contract beyond the first-year commitment.

Conclusions:The commercialization of UPnGO has demonstrated limited success. Initial adoption and reach of the program were successful; however, engagement and steps significantly declined over 12 months. UPnGO is being modified to further explore the viability of commercializing workplace PA interventions.

Co-production is reliant on consensus between participants, but in-person meetings and workshops are near-impossible for time-constrained and hard-to-reach groups. An innovative new method of consensus-gathering was needed to further understand physical inactivity and how to intervene in a time-constrained group. A lay advisory group meeting resulted in the decision to hold a workshop over a secret Facebook group with participants to develop a theory of change and theory of action.

Purpose: To test the feasibility using a secret Facebook group to co-produce an intervention with a time-constrained group, women with previous gestational diabetes.

Methods: The researchers created a secret Facebook group with content that sequentially progressed to develop a programme theory. We evaluated the feasibility and acceptability of the group by analysing Facebook analytics and a post-workshop survey. The researcher posted 1-2 times per day for 14 days. Messages and content were typically posted between 15:30 and 17:30 GMT, as suggested by the lay advisor group as mothers will be returning from work and may have a short break before dinnertime.

Results: Twenty-one participants took part. In total, 521 comments were provided in response to 18 posts of polls, video, text or photos (average = 28.9 comments per post). The total word count of participant comments was 21,142 words. The workshop was viewed positively, with 20 of 21 participants saying they liked the workshop somewhat or a great deal, and felt the group was a safe and open environment to share opinions. When asked if they would take part in something like this again, 15 of 21 said “Yes”. Participants mentioned the format was convenient to fit into their day; it allowed them to reflect on their own experiences; and how they liked helping research progress. The six participants who didn’t reply “yes” said it was still difficult finding time and it depended on what else was going on.

Conclusion: Using a secret Facebook group to develop a physical activity interventions proves to be a feasible and acceptable method. This holds significant potential for a diverse group of hard-to-reach groups who could benefit from this method of co-production.

Purpose: Physical activity is commonly measured as time spent in bouts of at least 10 minutes, however each minute of physical activity (PA) achieved per day is now considered to be beneficial for health by breaking up sitting time, changing posture and adding to daily PA minutes. This study compared the difference in the self-reported total time spent walking in a week to time spent walking in 10 minute bouts over the same 7 day period through a university population sample. 

Methods: A census-style, online survey was used to collect PA data from staff and students at The University of Sydney. One measure, from the validated Active Australia Survey (AAS) reports walking in bouts of 10min or more over seven-days, the second measure reports total estimated time spent walking over the seven day period. Data from n= 4582 was analysed and a paired t-test was performed to determine any difference between the mean time participants reported to walk in 10 min bouts and the mean total time they reported to walk over the same 7 day period. 

Results/findings: Our finding suggests very strong evidence that more walking was done in short bursts of less than 10 minute bouts. Over a seven day period, participants engaged in a mean of 352 minutes of walking in total compared with 235 minutes of walking in 10 minute bouts. When comparing the reported minutes of walking in either 10min bouts or total minutes, there was a significant 117min/week difference (p<0.001). 

Conclusion: Significantly more walking is done in short bursts across a week. Walking in this way contributed to an extra 16min/day of physical activity which is beneficial for health even when completed in bouts less than 10 minutes. Measuring total walking minutes rather than only in bouts of at least 10 minutes may provide a more useful estimate to monitor population trends and evaluate PA interventions.

Purpose: Current evidence for optimal activity durations is largely limited to studies exploring one activity in relation to one health outcome. But daily activities are intrinsically linked – one activity cannot change without other activities also changing to compensate. In addition, the best mix of daily activities may differ across health outcomes. This study aimed to develop novel analytical methods to optimise the mix of daily activities for multiple health outcomes. 

Methods: Participants (n=664, 11-12 y, 50% boys) were from the Child Health CheckPoint module of the Longitudinal Study of Australian Children. Daily activity behaviours (sleep, sedentary time, light physical activity [LPA] and moderate-to-vigorous physical activity [MVPA]) from 8-day, 24-h accelerometry were expressed as isometric log-ratios. Outcomes were composite z-scores for fitness (VO_2maxfrom cycle ergometry and standing broad jump) and fatness (measured waist-to-height ratio, body mass index and body composition [fat-to-fat free mass isometric log-ratio]). Compositional linear models regressed activity behaviour isometric log-ratios against fitness and fatness scores, adjusted for age, sex, puberty and socioeconomic status. The models were used to predict fitness and fatness for all possible empirical activity mixes. The best quartiles of activity mixes (optimal time-use zones) for fitness and fatness were plotted in quaternary tetrahedrons. The overlapping area of optimal fitness and fatness zones was described as the overall optimal time-use zone.

Results: The daily activity mix (isometric log-ratios) was associated with fitness and fatness (both p<0.001). The optimal time-use zones for both outcomes had relatively high MVPA and low sedentary time, but the optimal zone for fitness had higher LPA and lower sleep than the optimal zone for fatness. The overall optimal time-use zone for fitness and fatness was between: 9.0 and 10.7h sleep; 7.3 and 10.0h sedentary time; 2.5 and 4.8h LPA; 1.8 and 2.5h MVPA. 

Conclusions:  The analytical methods developed in this study offer a way to optimise the mix of daily activities for multiple valued health outcomes. Findings provide evidence to underpin 24-hour activity guidelines and lifestyle behaviour interventions. Future work will refine models using machine learning, non-linear response surfaces, higher-dimensional activity compositions and a wider range of health outcomes.