How to use machine learning and fuzzy cognitive maps to test hypothetical scenarios in health behavior change interventions: a case study on fruit intake.

BACKGROUND: Intervention planners use logic models to design evidence-based health behavior interventions. Logic models that capture the complexity of health behavior necessitate additional computational techniques to inform decisions with respect to the design of interventions. OBJECTIVE: Using empirical data from a real intervention, the present paper demonstrates how machine learning can be used together with fuzzy cognitive maps to assist in designing health behavior change interventions. METHODS: A modified Real Coded Genetic algorithm was applied on longitudinal data from a real intervention study. The dataset contained information about 15 determinants of fruit intake among 257 adults in the Netherlands. Fuzzy cognitive maps were used to analyze the effect of two hypothetical intervention scenarios designed by domain experts. RESULTS: Simulations showed that the specified hypothetical interventions would have small impact on fruit intake. The results are consistent with the empirical evidence used in this paper. CONCLUSIONS: Machine learning together with fuzzy cognitive maps can assist in building health behavior interventions with complex logic models. The testing of hypothetical scenarios may help interventionists finetune the intervention components thus increasing their potential effectiveness.

Measuring Engagement with Mental Health and Behavior Change Interventions: an Integrative Review of Methods and Instruments.

BACKGROUND: Engagement is a complex construct consisting of behavioral, cognitive, and affective dimensions, making engagement a difficult construct to measure. This integrative review aims to (1) present a multidisciplinary overview of measurement methods that are currently used to measure engagement with adult mental health and behavior change interventions, delivered in-person, blended, or digitally, and (2) provide a set of recommendations and considerations for researchers wishing to study engagement. METHODS: We used an integrative approach and identified original studies and reviews on engagement with mental health or behavior change interventions that were delivered in-person, digitally, or blended. RESULTS: Forty articles were analyzed in this review. Common methods to assess engagement were through objective usage data, questionnaire-based data, and qualitative data, with objective usage data being used most frequently. Based on the synthesis of engagement measures, we advise researchers to (1) predefine the operationalization of engagement for their specific research context, (2) measure behavioral, cognitive, and affective dimensions of engagement in all cases, and (3) measure engagement over time. CONCLUSIONS: Current literature shows a bias towards behavioral measures of engagement in research, as most studies measured engagement exclusively through objective usage data, without including cognitive and affective measures of engagement. We hope that our recommendations will help to reduce this bias and to steer engagement research towards an integrated approach.

Adherence to the Mediterranean diet among adults in Mediterranean countries: a systematic literature review.

BACKGROUND AND AIM: While the Mediterranean diet (MD) is promoted in non-Mediterranean countries, inhabitants of Mediterranean countries seem to be shifting away from this healthy diet. The aim of this study is to provide an overview of MD adherence in the general adult population of Mediterranean countries. METHODS: A systematic review was conducted following the PRISMA 2020 (Preferred Reporting Items for Systematic Review and Meta-Analysis) guidelines and registered in the Prospero database (CRD42020189337). Literature was searched in PubMed, Web of Science and PsycINFO databases for studies published from 2010 up to and including 2021. The following inclusion criteria were used: age 18 years and older, sample size > 1000 participants, and using a validated MD adherence score. Studies that only included participants with nutrition-related or other severe chronic disorders, as well as studies that only included specific subpopulations (e.g., pregnant women), were excluded in order to focus on the general adult population. A quality analysis of the included studies was done using the NCCMT scale. RESULTS: A total of 50 studies were included. The number of participants in the included studies ranged between 1013 and 94,113. Most of the included studies pertained to the European Mediterranean countries, with fewer studies from the Middle Eastern and North African Mediterranean countries. The vast majority of the included studies reported low or moderate MD adherence, both based on the mean adherence as well as the low or moderate adherence category often being the most prevalent. There were no clear differences noted between sex and age groups. The quality assessment generally showed weak or moderate scores. CONCLUSIONS: Mediterranean populations have been showing moderate adherence to MD in the past 10 years, indicating room for improving adherence to the MD in countries of its origin.

The association of hyperglycaemia and insulin resistance with incident depressive symptoms over 4 years of follow-up: The Maastricht Study.

AIMS/HYPOTHESIS: Depression is twice as common in individuals with type 2 diabetes as in the general population. However, it remains unclear whether hyperglycaemia and insulin resistance are directly involved in the aetiology of depression. Therefore, we investigated the association of markers of hyperglycaemia and insulin resistance, measured as continuous variables, with incident depressive symptoms over 4 years of follow-up. METHODS: We used data from the longitudinal population-based Maastricht Study (n = 2848; mean age 59.9 ± 8.1 years, 48.8% women, 265 incident depression cases, 10,932 person-years of follow-up). We assessed hyperglycaemia by fasting and 2 h post-load OGTT glucose levels, HbA1c and skin autofluorescence (reflecting AGEs) at baseline. We used the Matsuda insulin sensitivity index and HOMA-IR to calculate insulin resistance at baseline. Depressive symptoms (nine-item Patient Health Questionnaire score ≥10) were assessed at baseline and annually over 4 years. We used Cox regression analyses, and adjusted for demographic, cardiovascular and lifestyle risk factors. RESULTS: Fasting plasma glucose, 2 h post-load glucose and HbA1c levels were associated with an increased risk for incident depressive symptoms after full adjustment (HR 1.20 [95% CI 1.08, 1.33]; HR 1.25 [1.08, 1.44]; and HR 1.22 [1.09, 1.37] per SD, respectively), while skin autofluorescence, insulin sensitivity index and HOMA-IR were not (HR 0.99 [0.86, 1.13]; HR 1.02 [0.85, 1.25]; and HR 0.93 [0.81, 1.08], per SD, respectively). CONCLUSIONS/INTERPRETATION: The observed temporal association between hyperglycaemia and incident depressive symptoms in this study supports the presence of a mechanistic link between hyperglycaemia and the development of depressive symptoms. Graphical abstract.

Predictors of Self-Determined Module Choice in a Web-Based Computer-Tailored Diet and Physical Activity Intervention: Secondary Analysis of Data From a Randomized Controlled Trial.

BACKGROUND: Tailoring an online intervention to participant preferences (eg, by giving participants a choice which modules to follow) may increase engagement in the intervention, motivation for behavioral change, and possibly intervention effects. So far, little is known about what characteristics predict these module choices. Filling this knowledge gap is useful for optimizing program engagement. OBJECTIVE: We investigated participant choice for a dietary and/or physical activity (PA) promotion module in our web-based computer-tailored intervention based on self-determination theory (SDT) and motivational interviewing (MI). Furthermore, we investigated which demographic characteristics, current behavior, psychosocial constructs and constructs from SDT and MI, and program-related variables such as advice on which module to follow were associated with these choices. METHODS: Observational data were used from the randomized controlled trial MyLifestyleCoach of participants who were randomized into the intervention condition, completed the baseline questionnaire, and made a module choice in the opening session of the intervention. Here, they received advice on their own dietary and PA behavior. At the session's end, they chose which lifestyle modules they would like to follow (both, diet, PA, or no module). Measurements included demographic information; self-reported diet and PA; and several psychosocial, SDT, and MI constructs. In total, data from 619 Dutch adults (59.6% women; mean age was 51.9 [SD 13.5] years) were analyzed. A stepwise multinomial logistic regression analysis was conducted to investigate which characteristics are related to module choice; the diet module served as reference category as almost everyone was advised to follow this module. RESULTS: Of this sample, 54.8% (339/619) chose to do both the diet and PA module, 25.4% (157/619) chose to follow the diet module, 17.8% (110/619) preferred to follow no module, and 2.1% (13/619) chose to do the PA module only. Furthermore, it was found that older people, those who consumed more fruit, and those who scored lower on importance to change their current diet were more likely to choose no module compared to the diet module. People who had more motivation to change their current PA and those who received strong advice compared with slight advice to follow the diet module were more likely to choose both modules compared with the diet module only. CONCLUSIONS: The results show that more than half of the sample was interested in following both the diet and PA module in this online lifestyle intervention. Several characteristics were found to be related to module choice. A future challenge is to examine how this knowledge can be used to improve future interventions, such as tailoring (messages or content) on specific groups or examining where and how MI could be used to motivate people to make a certain module choice. TRIAL REGISTRATION: Netherlands Trial Register NL7333; https://www.trialregister.nl/trial/7333.

Effects of Interactivity on Recall of Health Information: Experimental Study.

BACKGROUND: Information provided in an interactive way is believed to be engaging because users can actively explore the information. Yet empirical findings often contradict this assumption. Consequently, there is still little known about whether and how interactivity affects communication outcomes such as recall. OBJECTIVE: The aim of this study was to investigate mechanisms through which interactivity affects recall of online health information. We tested whether and how cognitive involvement, perceived active control, and cognitive load mediate the effects of interactivity on recall. In addition, we examined need for cognition and health literacy as potential moderators of the mediation effects. Given the increasing popularity of dietary supplement use, our health website focused on this topic. METHODS: In an online between-subjects experiment (n=983), participants were randomly assigned to control condition (no interactive features), moderate interactivity (dropdown menus), and high interactivity (dropdown menus and responsive infographics). Two weeks before the experiment, background characteristics and moderating variables were measured. During website visit, data on users' online behavior were collected. Recall was measured postexposure. RESULTS: Participants recalled significantly less information in the moderate (mean 3.48 [SD 2.71]) and high (mean 3.52 [SD 2.64]) interactivity conditions compared with the control condition (mean 5.63 [SD 2.18]). In the mediation analysis, we found direct, negative effects of moderate (b=-2.25, 95% CI -2.59 to -1.90) and high (b=-2.16, 95% CI -2.51 to -1.81) levels of interactivity on recall as well. In the relationship between interactivity and recall, cognitive involvement had a partial negative mediation effect (moderate interactivity: b=-.20; 95% CI -0.31 to -0.10; high interactivity: b=-.21, 95% CI -0.33 to -0.10) and perceived active control had a partial positive mediation effect (moderate interactivity: b=.28, 95% CI 0.18 to 0.40; high interactivity: b=.27, 95% CI 0.16 to 0.40). CONCLUSIONS: Interactivity decreased recall. In addition, through interactivity participants were less involved with the content of the information, yet they felt they had more control over the information. These effects were stronger in the high need for cognition and high health literate groups compared with their counterparts.

How to foster informed decision making about food supplements: results from an international Delphi study.

Food supplement use can have beneficial and detrimental effects, making informed decisions about supplement use important. How these decisions are made and which communication strategies can stimulate informed decision making is unclear. This study identified the important characteristics of (i) informed decision making about food supplement use and (ii) important factors indicating how to communicate about food supplements to foster informed decision making. An online three-round Delphi study was conducted. International experts within the field of (risk) communication about food supplements or related fields were recruited via email. The participants' age ranged from 25 to 69 years, and sample sizes for the three rounds were 38, 89 and 51, respectively. Experts indicated that for making an informed decision about food supplement use one needs to have knowledge of their positive and negative effects, the ability to compare these effects, knowing alternatives besides supplements, feeling informed, and feeling able (self-efficacious) to make the decision and making the decision voluntarily. Important communication strategies mentioned were: provision of information about positive and negative effects and the nature of these effects including scientific evidence, ensuring information is easily accessible, well ordered, tailored and provided by a trustworthy, credible and independent source.

Exploring beliefs about dietary supplement use: focus group discussions with Dutch adults.

OBJECTIVE: Although dietary supplement use is increasing in Europe and the USA, little research involving adults' beliefs regarding dietary supplements has been conducted. Therefore, the present study aimed to explore and compare users' and non-users' beliefs towards dietary supplements. DESIGN: Thirteen focus group discussions were conducted of which seven groups were dietary supplement users and six groups were non-users. Based on the socio-cognitive factors of the Integrated Change Model, a semi-structured topic guide was set up. The discussions were audio-recorded and subjected to qualitative content analysis, applying the framework approach. SETTING: Data were collected in Maastricht, the Netherlands, in 2014 and 2015. SUBJECTS: In total fifty-six individuals participated in the study, of whom twenty-eight were dietary supplement users and twenty-eight non-users. The average age of participants was 42·9 years. RESULTS: Dietary supplement users' attitude beliefs were mainly related to mental and physical health enhancement, illness prevention and curative health benefits. Users were critical of the nutritional knowledge of health professionals and of the quality of food products. Non-users were convinced that the human body does not need any support and that regular food is enough to cover one's nutritional needs. Users and non-users held comparable beliefs regarding the definition and risks of dietary supplements, and perceived social influences. CONCLUSIONS: In their decision about dietary supplement use, both groups were guided by their own convictions to a great extent. Both groups would benefit from improved understanding of the health effects of dietary supplements to improve informed decision making.

Implementation of web-based interventions by Dutch occupational health centers.

The aim of this qualitative study was to identify barriers and facilitators to the adoption and particularly the implementation of a web-based computer-tailored obesity prevention intervention by occupational health centers. Participants were directors of Dutch occupational health centers who had adopted and implemented the intervention for the corresponding efficacy study (n = 8) as well as non-adopters (n = 12). Individual semi-structured interviews were carried out to study barriers and facilitators related to the intervention, the user, the organization, and the socio-political environment. All interviews were carried out by telephone, audio-recorded and transcribed verbatim. The transcripts were analyzed using a directed-content approach and coded by two persons. There were important differences in perceptions between adopters and non-adopters, particularly on barriers and facilitators related to the intervention and the personal beliefs of the implementer. The relative advantages of the intervention were considered to be most important. Participants also indicated that their personal attitudinal and self-efficacy beliefs influenced their implementation efforts. Regarding the organization, the possibilities to increase profits and integrate the intervention within the organization were considered to be important facilitators for the implementation. Participants mentioned few implementation barriers and facilitators related to the socio-political environment. Strategies to improve the implementation of web-based computer-tailored interventions by occupational health centers should be tailored to implementers' unique perceptions and particularly address the perceived advantages and disadvantages of the intervention, attitudinal and self-efficacy beliefs, and the potential to increase organizations' profits and competitiveness.

Development of a computer-tailored nutrition and physical activity intervention for lower-educated women of Dutch, Turkish and Moroccan origin using content matching and ethnic identity tailoring.

BACKGROUND: Unhealthy dietary and physical activity (PA) patterns are highly prevalent in most Western countries, especially among lower-educated and ethnic minority groups. Therefore, interventions to promote healthy eating and physical activity that can reach large numbers of lower-educated people are needed. When developing interventions, the ethnic diversity of the lower-educated population may be taken into account to make intervention material more appealing to the target group. This article describes the development and evaluation of two computer-tailored nutrition and physical activity interventions for lower-educated Dutch, Turkish and Moroccan women. One version is tailored to sociocognitive variables (traditional tailoring), while the other is additionally tailored to ethnic identity (EI-tailoring). METHOD: Using intervention mapping, two evidence- and theory-based interventions were developed. In the traditional tailoring intervention, messages are tailored to health behavior, awareness of own behavior, attitude and self-efficacy. The behavior change techniques used to address these factors are: descriptive and evaluative feedback, arguments, modeling, goal setting, planning, barrier identification and advice on how to deal with barriers, stimulating resistance to social pressure, mobilization of social support (nontailored), active learning (nontailored) and iterative feedback. In the EI-tailoring intervention, the material is additionally tailored to ethnic identity (EI). This means that recipients who feel strongly attached to their ethnic background receive different intervention material than recipients with a weak attachment to their background. This includes, for instance, the use of more traditional colors, role models that match with their origin and advice messages that refer to their ethnicity of origin. DISCUSSION: Developing an intervention that matches the needs of this specific target population was challenging due to the little evidence regarding the determinants of their health behavior, as well as the behavioral change techniques that have not been tested among Turkish and Moroccan women in the Netherlands before. Based on previous research among this and other target populations we hypothesize, however, that the determinants and strategies we use will be suitable. A randomized controlled trial will show whether the interventions are effective among our specific target group and whether EI-tailoring is beneficial. TRIAL REGISTRATION: Dutch Trial Registry NTR4506 , registration date: 1(st) may 2014.

Similarities and differences in underlying beliefs of socio-cognitive factors related to diet and physical activity in lower-educated Dutch, Turkish, and Moroccan adults in the Netherlands: a focus group study.

BACKGROUND: Unhealthy eating patterns and a lack of physical activity (PA) are highly prevalent in most Western countries, especially among lower-educated people, including people of non-Western origin. The aim of this study was to investigate and compare the beliefs and barriers that underlie socio-cognitive and planning constructs related to healthy eating and PA among lower-educated Dutch, Turkish, and Moroccan adults. METHODS: Focus group interviews were conducted with 90 Dutch, Turkish, and Moroccan lower-educated adults between March and August 2012. Five semi-structured group interviews were conducted with Dutch participants, five with Turkish participants, and four with Moroccan participants. Men and women were interviewed separately. The question route was based on the Theory of Planned Behavior and self-regulation theories. The theoretical method used for the qualitative data analysis was content analysis. The interviews were recorded, transcribed, and analyzed by applying the framework approach. RESULTS: Some participants seemed to lack knowledge of healthy eating and PA, especially regarding the health consequences of an unhealthy lifestyle. Important attitude beliefs concerning healthy eating and PA were taste and health benefits. Participants suggested that social support can encourage the actual performance of healthy behavior. For instance, exercising with other people was perceived as being supportive. Perceived barriers to PA and cooking healthily were a lack of time and tiredness. These previously mentioned beliefs arose in all the ethnic groups. Differences were also found in beliefs between the ethnic groups, which were mainly related to religious and cultural issues. Turkish and Moroccan participants discussed, for example, that the Koran contains the recommendation to eat in moderation and to take care of one's body. Furthermore, they reported that refusing food when offered is difficult, as it can be perceived as an insult. Finally, men and women usually cannot exercise in the same location, which was perceived as a barrier. These factors did not emerge in the Dutch groups. CONCLUSIONS: The same cognitive beliefs were discussed in all three ethnic groups. The importance of cultural and religious factors appeared to be the most significant difference between the Turkish/Moroccan groups and the Dutch groups. Accordingly, interventions for all three ethnic groups should focus on socio-cognitive beliefs, whereas interventions for Turkish and Moroccan populations can additionally take religious and cultural rules into account.

Exploring individual cognitions, self-regulation skills, and environmental-level factors as mediating variables of two versions of a Web-based computer-tailored nutrition education intervention aimed at adults: A randomized controlled trial.

BACKGROUND: This study explored whether the determinants that were targeted in two versions of a Web-based computer-tailored nutrition education intervention mediated the effects on fruit, high-energy snack, and saturated fat intake among adults who did not comply with dietary guidelines. METHOD: A RCT was conducted with a basic (tailored intervention targeting individual cognitions and self-regulation), plus (additionally targeting environmental-level factors), and control group (generic nutrition information). Participants were recruited from the general Dutch adult population and randomly assigned to one of the study groups. Online self-reported questionnaires assessed dietary intake and potential mediating variables (behavior-specific cognitions, action- and coping planning, environmental-level factors) at baseline and one (T1) and four (T2) months post-intervention (i.e. four and seven months after baseline). The joint-significance test was used to establish mediating variables at different time points (T1-mediating variables - T2-intake; T1-mediating variables - T1-intake; T2-mediating variables - T2-intake). Educational differences were examined by testing interaction terms. RESULTS: The effect of the plus version on fruit intake was mediated (T2-T2) by intention and fruit availability at home and for high-educated participants also by attitude. Among low/moderate-educated participants, high-energy snack availability at home mediated (T1-T1) the effect of the basic version on high-energy snack intake. Subjective norm mediated (T1-T1) the effect of the basic version on fat intake among high-educated participants. DISCUSSION: Only some of the targeted determinants mediated the effects of both intervention versions on fruit, high-energy snack, and saturated fat intake. A possible reason for not finding a more pronounced pattern of mediating variables is that the educational content was tailored to individual characteristics and that participants only received feedback for relevant and not for all assessed mediating variables. TRIAL REGISTRATION: Netherlands Trial Registry NTR3396.

A Mobile Phone App Intervention Targeting Fruit and Vegetable Consumption: The Efficacy of Textual and Auditory Tailored Health Information Tested in a Randomized Controlled Trial.

BACKGROUND: Mobile phone apps are increasingly used to deliver health interventions, which provide the opportunity to present health information via different communication modes. However, scientific evidence regarding the effects of such health apps is scarce. OBJECTIVE: In a randomized controlled trial, we tested the efficacy of a 6-month intervention delivered via a mobile phone app that communicated either textual or auditory tailored health information aimed at stimulating fruit and vegetable intake. A control condition in which no health information was given was added. Perceived own health and health literacy were included as moderators to assess for which groups the interventions could possibly lead to health behavior change. METHODS: After downloading the mobile phone app, respondents were exposed monthly to either text-based or audio-based tailored health information and feedback over a period of 6 months via the mobile phone app. In addition, respondents in the control condition only completed the baseline and posttest measures. Within a community sample (online recruitment), self-reported fruit and vegetable intake at 6-month follow-up was our primary outcome measure. RESULTS: In total, 146 respondents (ranging from 40 to 58 per condition) completed the study (attrition rate 55%). A significant main effect of condition was found on fruit intake (P=.049, partial η(2)=0.04). A higher fruit intake was found after exposure to the auditory information, especially in recipients with a poor perceived own health (P=.003, partial η(2)=0.08). In addition, health literacy moderated the effect of condition on vegetable intake 6 months later (P<.001, partial η(2)=.11). A higher vegetable intake was found for recipients with high health literacy after exposure to the textual or auditory intervention compared to the control condition (contrasts P=.07 and P=.004, respectively). In the case of relatively low health literacy, vegetable intake was the highest in the control condition (contrasts text control: P=.03; audio control: P=.04). CONCLUSIONS: This study provides evidence-based insight into the effects of a mobile health app. The app seems to have the potential to change fruit and vegetable intake up to 6 months later, at least for specific groups. We found different effects for fruit and vegetable intake, respectively, suggesting that different underlying psychological mechanisms are associated with these specific behaviors. Based on our results, it seems worthwhile to investigate additional ways to increase fruit and vegetable intake in recipients with low health literacy. CLINICALTRIAL: International Standard Randomized Controlled Trial Number (ISRCTN): 23466915; http://www.isrctn.com/ISRCTN23466915 (Archived by WebCite at http://www.webcitation.org/6hTtfSvaz).

Long term effects of self-determination theory and motivational interviewing in a web-based physical activity intervention: randomized controlled trial.

BACKGROUND: Our main objective in the current study was to evaluate the long-term effectiveness (12 months from baseline) of I Move (a web-based computer tailored physical activity intervention, based on self-determination theory and motivational interviewing). To this end, we compared I Move to a web-based computer tailored physical activity intervention based on traditional health behavior theories (Active Plus), and to a no-intervention control group. As a secondary objective, the present study aimed to identify participant characteristics that moderate the long term effects of I Move and Active Plus. METHODS: A randomized controlled trial was conducted, comparing three research conditions: 1) the I Move condition, participants in this condition received I Move; 2) the Active Plus condition, participants in this condition received Active Plus; 3) the control condition; participants in this condition received no intervention and were placed on a waiting list. Main outcome measures were weekly minutes of moderate to vigorous physical activity and weekly days with minimal 30 min of physical activity. All measurements were taken by web-based questionnaires via the study website. Intervention effects were analyzed using multilevel linear regression analyses. RESULTS: At 12 months from baseline, I Move was found to be effective in increasing weekly minutes of moderate to vigorous physical activity (ES = .13), while Active Plus was not. In contrast, Active Plus was found to be effective in increasing weekly days with ≥ 30 min PA at 12 months (ES = .11), while I Move was not. No moderators of the effects of I Move were found. CONCLUSIONS: The results suggest that web-based computer tailored physical activity interventions might best include elements based on both self-determination theory/motivational interviewing and traditional health behavioral theories. To be more precise, it is arguable that the focus of the theoretical foundations, used in new web-based PA interventions should depend on the intended program outcome. In order to draw firm conclusions, however, more research on the effects of self-determination theory and motivational interviewing in web-based physical activity promotion is needed. TRIAL REGISTRATION: Dutch Trial Register NTR4129.

A longitudinal study on determinants of the intention to start smoking among Non-smoking boys and girls of high and low socioeconomic status.

BACKGROUND: This study identifies differences in socio-cognitive factors as they relate to the intention to smoke among boys and girls living in high socioeconomic status (HSES) and low socioeconomic status (LSES) neighborhoods. METHODS: A total of 1,643 children (aged 10-12 years) completed a web-based questionnaire assessing their intention, attitude, social influences, and self-efficacy toward smoking at baseline and at one year follow-up. Logistic regression analyses were conducted to examine the relations between intention and predictor variables (i.e. attitude, social influence, and self-efficacy). Three-way interaction terms were added to the first analysis to examine potential interactions of gender, socioeconomic status and predictor variables. A 3-way interaction effect was present, and therefore subgroup analyses for HSES and LSES boys and girls were warranted. RESULTS: The results indicated that positive attitudes toward smoking were related to the intention to smoke among HSES boys, whereas HSES girls had higher intentions to smoke if they perceived fewer disadvantages of smoking (OR: 0.42; 95 % CI: 0.22-0.82). The intention to smoke among LSES boys was predicted by perceived social norms (OR: 0.49; 95 % CI: 0.25-0.93); in LSES girls, the smoking behavior of people in their environment was most strongly related to their smoking intention (OR: 5.55; 95 % CI: 2.81-10.93). CONCLUSIONS: To prevent youth smoking, HSES boys and girls may benefit from interventions that address attitudes. Boys from an LSES neighborhood may profit from smoking prevention interventions that target social norms, while LSES girls may benefit from strategies aimed at resisting the influence of smokers in their environment. TRIAL REGISTRATION: The 'Fun without Smokes' study is approved by the Medical Ethics Committee of the Atrium-Orbis-Zuyd Hospital (NL32093.096.11/MEC 11-T-25) and registered in the Dutch Trial Register ( NTR3116 ).

Long-term efficacy of a Web-based computer-tailored nutrition education intervention for adults including cognitive and environmental feedback: a randomized controlled trial.

BACKGROUND: Unhealthy diets are prevalent in Western countries, especially among low-educated people. To have an effect on health, it is important that dietary changes are sustained over time. This study examines long-term effects of a cognitive and environmental-feedback version of a Web-based computer-tailored (CT) nutrition education intervention targeting fruit, vegetables, high-energy snacks and saturated fat. METHODS: A randomized controlled trial was conducted with a basic (tailored intervention targeting individual cognitions and self-regulation processes; n = 456), plus (additionally targeting environmental-level factors; n = 459) and control group (generic nutrition information; n = 434). Participants were recruited from the general population and randomly assigned to a study group. Online self-reported questionnaires assessed fruit, vegetable, high-energy snack and saturated fat intake, self-regulation, self-control, and Body Mass Index (BMI) at baseline and nine-months post-intervention. Linear mixed model analyses examined group differences in change over time. Educational differences were examined by 'group X time X education' interaction terms. Effects were examined in the total sample and among participants who did not comply with dietary- or BMI guidelines. RESULTS: The effects on vegetable intake in the total sample differed according to educational level (p = 02). Among low/moderate-educated participants, the basic version was significantly more effective in increasing vegetable intake than the control program (effect size (ES) = 0.32) and plus version (ES = 0.22). No effects were found for high-educated participants. Self-regulation change was significantly larger in the control group than in the basic (ES = 0.18) and plus (ES = 0.16) group. CONCLUSIONS: In general, both intervention versions did not result in long-term intervention effects. The exception was an effect of the basic version on self-reported vegetable intake among low/moderate-educated adults in the total sample. More research is needed on how targeting self-regulation processes and environmental-level factors in Web-based CT nutrition education interventions can increase long-term efficacy. TRIAL REGISTRATION: Netherlands Trial Registry NTR3396 .

Eating in moderation and the essential role of awareness. A Dutch longitudinal study identifying psychosocial predictors.

BACKGROUND: Eating in moderation, i.e. the attempt to monitor and limit the intake of energy-dense foods, is a promising strategy in the prevention of weight gain. The purpose of this study was to examine which psychosocial factors derived from the I-Change Model (ICM) were associated with eating in moderation, and whether these factors differed between adults with a correct (aware) or incorrect (unaware) perception of their dietary behaviour. METHODS: This study used a longitudinal design with measurements at baseline (N = 483) and six-month follow-up (N = 379). Eating in moderation was defined as the average daily energy intake from energy-dense food products and was measured by a validated food frequency questionnaire. Linear regression analyses were used to assess the associations between the ICM factors and eating in moderation. The moderating role of awareness was examined by including interactions between awareness and the ICM factors in the regression analyses using the pick-a-point approach to further examine the associations for aware and unaware participants. RESULTS: Participants who were aware of their dietary behaviour had a significantly lower average daily energy intake compared to those who were unaware. Eating in moderation was predicted by awareness, risk perception, social influence and intention. Among the aware participants, eating in moderation was predicted by risk perception, attitude, social influence and intention. Among the unaware participants, only risk perception and self-efficacy were significantly associated with eating in moderation. CONCLUSIONS: Our findings show that psychosocial factors may only predict eating in moderation when people are aware of their risk behaviour. Therefore, interventions aimed at promoting complex behaviours, such as eating in moderation, should first focus on improving individuals' awareness of their risk behaviour before targeting motivational factors.

Comparing a Video and Text Version of a Web-Based Computer-Tailored Intervention for Obesity Prevention: A Randomized Controlled Trial.

BACKGROUND: Web-based computer-tailored interventions often suffer from small effect sizes and high drop-out rates, particularly among people with a low level of education. Using videos as a delivery format can possibly improve the effects and attractiveness of these interventions OBJECTIVE: The main aim of this study was to examine the effects of a video and text version of a Web-based computer-tailored obesity prevention intervention on dietary intake, physical activity, and body mass index (BMI) among Dutch adults. A second study aim was to examine differences in appreciation between the video and text version. The final study aim was to examine possible differences in intervention effects and appreciation per educational level. METHODS: A three-armed randomized controlled trial was conducted with a baseline and 6 months follow-up measurement. The intervention consisted of six sessions, lasting about 15 minutes each. In the video version, the core tailored information was provided by means of videos. In the text version, the same tailored information was provided in text format. Outcome variables were self-reported and included BMI, physical activity, energy intake, and appreciation of the intervention. Multiple imputation was used to replace missing values. The effect analyses were carried out with multiple linear regression analyses and adjusted for confounders. The process evaluation data were analyzed with independent samples t tests. RESULTS: The baseline questionnaire was completed by 1419 participants and the 6 months follow-up measurement by 1015 participants (71.53%). No significant interaction effects of educational level were found on any of the outcome variables. Compared to the control condition, the video version resulted in lower BMI (B=-0.25, P=.049) and lower average daily energy intake from energy-dense food products (B=-175.58, P<.001), while the text version had an effect only on energy intake (B=-163.05, P=.001). No effects on physical activity were found. Moreover, the video version was rated significantly better than the text version on feelings of relatedness (P=.041), usefulness (P=.047), and grade given to the intervention (P=.018). CONCLUSIONS: The video version of the Web-based computer-tailored obesity prevention intervention was the most effective intervention and most appreciated. Future research needs to examine if the effects are maintained in the long term and how the intervention can be optimized. CLINICALTRIAL: Netherlands Trial Register: NTR3501; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3501 (Archived by WebCite at http://www.webcitation.org/6cBKIMaW1).

Use and Effectiveness of a Video- and Text-Driven Web-Based Computer-Tailored Intervention: Randomized Controlled Trial.

BACKGROUND: Many Web-based computer-tailored interventions are characterized by high dropout rates, which limit their potential impact. OBJECTIVE: This study had 4 aims: (1) examining if the use of a Web-based computer-tailored obesity prevention intervention can be increased by using videos as the delivery format, (2) examining if the delivery of intervention content via participants' preferred delivery format can increase intervention use, (3) examining if intervention effects are moderated by intervention use and matching or mismatching intervention delivery format preference, (4) and identifying which sociodemographic factors and intervention appreciation variables predict intervention use. METHODS: Data were used from a randomized controlled study into the efficacy of a video and text version of a Web-based computer-tailored obesity prevention intervention consisting of a baseline measurement and a 6-month follow-up measurement. The intervention consisted of 6 weekly sessions and could be used for 3 months. ANCOVAs were conducted to assess differences in use between the video and text version and between participants allocated to a matching and mismatching intervention delivery format. Potential moderation by intervention use and matching/mismatching delivery format on self-reported body mass index (BMI), physical activity, and energy intake was examined using regression analyses with interaction terms. Finally, regression analysis was performed to assess determinants of intervention use. RESULTS: In total, 1419 participants completed the baseline questionnaire (follow-up response=71.53%, 1015/1419). Intervention use declined rapidly over time; the first 2 intervention sessions were completed by approximately half of the participants and only 10.9% (104/956) of the study population completed all 6 sessions of the intervention. There were no significant differences in use between the video and text version. Intervention use was significantly higher among participants who were allocated to an intervention condition that matched their preferred intervention delivery format. There were no significant interaction terms for any of the outcome variables; a match and more intervention use did not result in better intervention effects. Participants with a high BMI and participants who felt involved and supported by the intervention were more likely to use the intervention more often. CONCLUSIONS: Video delivery of tailored feedback does not increase the use of Web-based computer-tailored interventions. However, intervention use can potentially be increased by delivering intervention content via participants' preferred intervention delivery format and creating feelings of relatedness. Because more intervention use was not associated with better intervention outcomes, more research is needed to examine the optimum number of intervention sessions in terms of maximizing use and effects. TRIAL REGISTRATION: Nederlands Trial Register: NTR3501; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3501 (Archived by WebCite at http://www.webcitation.org/6b2tsH8Pk).

A Web-based, computer-tailored smoking prevention program to prevent children from starting to smoke after transferring to secondary school: randomized controlled trial.

BACKGROUND: Smoking prevalence rates among Dutch children increase rapidly after they transit to secondary school, in particular among children with a low socioeconomic status (SES). Web-based, computer-tailored programs supplemented with prompt messages may be able to empower children to prevent them from starting to smoke when they transit to secondary school. OBJECTIVE: The main aim of this study is to evaluate whether computer-tailored feedback messages, with and without prompt messages, are effective in decreasing children's smoking intentions and smoking behavior after 12 and 25 months of follow-up. METHODS: Data were gathered at baseline (T0), and after 12 months (T1) and 25 months (T2) of follow-up of a smoking prevention intervention program called Fun without Smokes. A total of 162 schools were randomly allocated to a no-intervention control group, an intervention prompt group, or an intervention no-prompt group. A total of 3213 children aged 10 to 12 years old participated in the study and completed a Web-based questionnaire assessing their smoking intention, smoking behavior, and sociocognitive factors, such as attitude, social influence, and self-efficacy, related to smoking. After completion, children in the intervention groups received computer-tailored feedback messages in their own email inbox and those messages could be accessed on the intervention website. Children in the prompt group received prompt messages, via email and short message service (SMS) text messaging, to stimulate them to reuse the intervention website with nonsmoking content. Multilevel logistic regression analyses were performed using multiple imputations to assess the program effects on smoking intention and smoking behavior at T1 and T2. RESULTS: A total of 3213 children participated in the Fun without Smokes study at T0. Between T0 and T1 a total of 1067 children out of the original 3213 (33.21%) dropped out of the study. Between T0 and T2 the number of children that did not participate in the final measurement was 1730 out of the original 3213 (53.84%). No significant program effects were observed for any of the intervention groups compared to the control group at T1 for the intention to engage in smoking-prompt, OR 0.67 (95% CI 0.30-1.50), no-prompt, OR 0.76 (95% CI 0.34-1.67)-or for smoking behavior-prompt, OR 1.13 (95% CI 0.13-9.98), no-prompt, OR 0.50 (95% CI 0.04-5.59). Similar nonsignificant program effects were found at T2 for the intention to start smoking-prompt, OR 0.78 (95% CI 0.26-2.32), no-prompt, OR 1.31 (95% CI 0.45-3.82)-and smoking behavior-prompt, OR 0.53 (95% CI 0.12-2.47), no-prompt, OR 1.01 (95% CI 0.24-4.21). CONCLUSIONS: This study showed that the Web-based, computer-tailored feedback messages with and without prompt messages were not effective in modifying children's smoking intentions and smoking behavior as compared to no information. Future smoking prevention interventions are recommended to start closer to the age of actual smoking uptake. Furthermore, future studies on Web-based, computer-tailored smoking prevention programs should focus on assessing and controlling exposure to the educational content and the response to the prompt messages. TRIAL REGISTRATION: Netherlands Trial Register NTR3116; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3116 (Archived by WebCite at http://www.webcitation.org/6O0wQYuPI).