Reducing Fast Food Consumption in Students Using a Parent-Teacher Participation-Based Intervention: An Experimental Approach

Background: Fast food consumption among students is increasing dramatically. This study aims to evaluate the effect of an intervention based on the Theory of Planned Behavior (TPB) in reducing fast food consumption among high school students. Method: 160 high school students from Iran were randomly recruited and assigned to experiment or control groups. The intervention was conducted over three consecutive weeks, consisting of four, 45-minute teaching sessions. Parameters were assessed on three occasions: pretest, posttest, and follow-up. In these stages, participants responded to a scale on fast food consumption which measures the beliefs and behaviors toward fast food. Results: ndings revealed a statistically signicant difference in the posttest between experiment and control groups in the major components of fast food consumption including behavioral beliefs (t = 5.1, p < 0001), evaluation of behavioral outcomes (t = 5.3, p < 0001), normative beliefs (t = 2.3, p < 05), motivation to comply (t = 5.5, p < 0001), control beliefs (t = 4.4, p < 0001), perceived power (t = 3.3, p < 0001), and behavioral intention (t = .68, p < 0001). Similar results were obtained in the follow-up stage. Conclusion: The ndings suggest that the parent-teacher participation based intervention can be used to reduce fast food consumption amongst high school students both cognitively and behaviorally. Moreover, this intervention can be further customized to increase healthy food consumption in school students and other age groups beyond the context of school.

that fast food consumption is remarkably growing among high school students (  . Obesity as the popular outcome of fast food consumption is affecting more than half a billion individuals around the world (Bhurosy & Jeewson, 2014). Global statistics show that over 340 million children and adolescents aged 5-19 were overweight or obese in 2016 (World Health Organization, 2018). A similar study amongst an Iranian sample ages 12 to 17 revealed that 51% of children consume fast foods in the form of prepackaged snacks, packaged juice and soda, weekly (Dehdari & Mergen, 2012). In addition, a similar study on an Iranian sample found that 20% of adolescents and 10% of adults consumed sandwiches as a fast food at least three times a week (Faghih & Anousheh, 2008). Students who compose a signi cant percentage of the population are highly vulnerable to physical, psychological and social detriments, and desirable nutrition is a prerequisite for maintaining their health (Ghaffari et al., 2015).
Young adulthood is a critical time for developing healthy behaviors, as habits formed during this time tend to persist throughout life (Hammar, 2017). It has been suggested that changing eating habits may be done through changing attitudes towards them (Ogdan, Karim, choudry, & brown, 2007). With further research needed for understanding the link between attitudes and eating behaviors, there is a reliance on theories to provide substantial guidance for creating healthy behaviors (Ghaffari, Esfahani, Rakhshanderou, Hosseini koukamari, 2018; Davis, Campbell, Hildon, Hobbs, Michie, 2015). The Theory of Planned Behavior (TPB) developed by Ajzen and Fishbein (1980) has been widely used to examine the attitudes and beliefs associated with food choices. TPB predicts the occurrence of a speci c behavior that an individual tends to engage in. In fact, TPB assumes that human behaviors are determined by behavioral intention to consume fast food affected by attitude, subjective norms, and perceived behavioral control. In relation to fast food use, the attitude would be the personal positive/negative feelings about the consumption of fast food. Subjective norms would be the degree to which an individual desires to respect and follow the opinions of individuals who are important to him or her. Perceived behavioral control would be an individuals perceived capabilities and beliefs in the use or disuse of fast food in a given environment ( Thus, we attempt to change the nutritional beliefs and behaviors of students, by reducing fast food consumption, through designing and testing an experimental intervention, in turn, preventing the prevalence of obesity and associated diseases. As such, the main hypotheses of this study are: Hypothsis 1. There will be a signi cant reduction in fast food consumption between intervention and control groups in posttest Hypothsis 2. There will be a signi cant reduction in fast food consumption of students between intervention and control groups in follow-up

Participants
The present study follows an experimental longitudinal research design including intervention and control groups with pretest, posttest and follow-up stages. Participants consisted of 160 male students from four public high schools located in Khoy City, Iran. Primary recruitment was completed through the use of a multistage cluster sampling where 38 high schools were initially proposed and four were chosen at random selection. In the next step, two high schools were randomly selected for recruiting students for the intervention group and the other two were randomly selected for recruiting students for the control group. Next, from the rst two high schools 80 male students were randomly selected and assigned to the intervention group and from the second two high schools 80 students were randomly selected to be assigned to the control group. Figure 1 shows the steps in selecting schools and students, and assigning them randomly into intervention or control groups. Only students in the intervention group received the intervention. Students in the control group received no intervention.
The same method as the previous studies was used to collect data (e.g., Alizadeh, keshavarz, Jafari, & Ramezani, 2013), using sampling formula (α = 1.96 and β = 1.28), with an expectation of a 10% reduction in the sample, at least 80 participants were required for each group. Inclusion criteria consisted of: 1) participant must be a high school student (grade 11 in Iran), and 2) each participant must participate in this study voluntarily. Exclusion criteria included: 1) parents who refused their child to participate, and 2) students who refused to participate in all stages completely (pretest, posttest, and follow-up). According to the inclusion and exclusion criteria, 160 people were randomly assigned to experimental or control groups (80 students in each group). Before the intervention, both groups were compared in terms of confounding variables such as parent's education, parent's employment and socioeconomic status. As Table 1 shows, no signi cant differences were found between the groups and, therefore the two groups are comparable.

Procedure
This study was submitted to the ethics board of the Medical Science University of Shahid Beheshti in Iran and an ethics approval was granted. The consort checklist reporting the randomized trial can be seen in appendix 1. In accordance with this ethics approval, the research team contacted four randomly selected high schools in the city of Khoy, Iran to negotiate the possibility of conducting this study in these schools. All four high schools were informed and agreed to participate in this study subject to the voluntary nature as well as signed written consent from the parents of students before the students could participate in the study. Following informed consent, the research team distributed the self-report survey to the students in both the intervention and control groups. Students were informed that they could ask any questions during the study and had the right to withdraw from the study at any point in time. In all, 160 students in the intervention and control groups completed the self-report survey during the pretest stage. Immediately after conducting the intervention, the same number of students completed the self-report survey at the rst post-test stage. Two months after the intervention, the same students responded to the same survey to conclude data collection for the follow-up stage.
The experimental intervention was developed based on TPB propositions and consisted of 4 teaching sessions of 45 minutes each with the time interval of 5 days between sessions and lasted within 3 consecutive weeks. The rst training session was conducted in the staff room for the school manager, principals, and teachers. They were trained to talk about the harmful effects of fast food and the advantages of healthy food for the intervention group on a daily basis (5 days per week). They were also asked to clearly share with the students their positive attitudes and beliefs of healthy foods and their negative attitudes and beliefs of fast food. This session lasted for 45 minutes. The second training session was conducted for the parents of the intervention group. They were invited to the respective school to participate in a training session in which the research team brie y talked about the extent to which parent's attitudes towards fast food or healthy food can in uence their children's attitudes and beliefs. The parents were trained to openly share with their children the extent to which they disagree with the consumption of fast food and their favour of consuming healthy food. During the session there was a question and answer period. In the end, parents were asked to share their attitudes with their children on a daily basis. This session lasted for 45 minutes. Students were invited to the third training session which was comprised of the following: a roundtable question and answer period on the topic of fast food and healthy food, pamphlets about fast and healthy food were provided and video clips were shown depicting the advantages of healthy food and the disadvantages of fast food. The last training session was called "the fast food week", and all teachers, students and parents of students were informed about this week in advance. During this week the parent-school virtual communication system of the school sent out a total of eight text messages to the parents of the participating students. The messages were related to fast food consumption and contained topics such as the de nition of harmful foods and their negative impact on health, along with available solutions to reduce fast food consumption within families and schools. The messages aimed to in uence the current norms of students and to relay the message of parental dissatisfaction and overall negative attitudes toward fast food consumption of the students. Furthermore, during this week messages were delivered to students during morning rush [a routine morning program that students gather for 30 minutes, listen to the principal and teachers speeches and pray together] aimed at informing students that their teachers and principals do not have positive attitudes towards the consumption of fast food and they are encouraged to replace fast food consumption with consuming healthier foods. This was matched with the aim of the study to inform the students that both parents and teachers are in favor of consuming healthy foods and avoiding the consumption of fast food. At the end of this week, the students were asked to cook and bring in healthy foods and judges (three nutrition teachers) selected students who presented the best healthy foods in the competition. Winners of the competition were announced and received awards and rewards from the manager of the school. The overarching goal was to enhance the behavioral commitment of students to healthy foods versus fast foods.
Parameters assessed were measured on three occasions: before the intervention (pretest), immediately after intervention (posttest), and two months after the intervention (follow-up). In these stages, participants responded to a scale on fast food consumption which measures the beliefs and behaviors of individuals toward fast food consumption.

Measures
Fast Food Consumption Scale. This scale was developed by Ebadi, Rakhshanderou and Ghafari (2018) based on the propositions of Theory of Planned Behavior (Ajzen & Fishbein, 1980) and includes the six components as follows: demographic characteristics (6 items), actual behavior (1 item), behavioral intention (2 items), attitude towards fast food consumption (24 items), subjective norms (15 items), and perceived behavioral control (18 items). Content Validity Index (CVI) and Content Validity Ratio (CVR) values were higher than .80 and .70 respectively. Responses are recorded on a Likert response scale from 1 (completely disagree) to 5 (completely agree). The higher score indicates the higher tendency to fast food consumption in the total score and each of its components. This scale was used in pretest, rst posttest and second posttest.

Data analysis
The collected data was analyzed using IBM SPSS version 22 statistical package. Data analysis included two steps: 1) Descriptive statistics analyses which included frequencies, mean standard deviations, standard deviation and X 2 , and 2) Advanced statistics which included T-tests and ANOVA tests for comparing the pretest, posttest and follow-up stages. More speci cally, for descriptive analyses, absolute frequencies were calculated for categorical variables, and mean and SD for continuous variables. An ANOVA analysis was used to test for within-group differences from baseline to the end of the intervention. The type I error was set to 5% (two-sided). We used t and p estimates to con rm or refuse the research hypotheses. The method and procedure of this experiment was reviewed and con rmed by the Ethics Board of Shahid

Results
Demographic information of the parents of participants is presented in Table 1. A total of 160 high school students participated in this study. Table 2 shows the descriptive statistics of the scale across different components in control and experiment groups in pretest, primary posttest and secondary posttest.  Table 3 compares the changes between pretest and primary posttest. As the table shows the mean scores of behavioural beliefs (t = 1.12, p = 0.262) and evaluation of behavioral outcomes (t = 1.85, p = 0.066) had no signi cant difference between groups, while this difference appeared between the two groups immediately after the intervention for behavioral beliefs (t = 5.10, p < 0.0001) and evaluation of behavioral outcomes (t = 5.35, p < 0.0001). Moreover according to Table 4 this difference appeared two months after the intervention for behavioral beliefs (t = 4.27, p < 0.0001) and evaluation of behavioral outcomes (t = 4.62, p < 0.0001). Also, mean scores of normative beliefs (t = 0.73, p = 0.464) and motivation to comply (t = 1.46, p = 0.144) had no signi cant difference between groups before intervention; while this difference appeared between the two groups immediately after the intervention for normative beliefs (t = 2.32, p = 0.021), and for motivation to comply (t = 5.56, P < 0.0001). According to Table 4, this difference also was observed two months after the intervention for normative beliefs (t = 2.78, p = 0.006) and for motivation to comply (t = 5.81, p < 0.0001).   Table 3, mean scores of control beliefs (t = 1.50, p = 0.134) and perceived power (t = 0.29, p = 0.768) had no signi cant difference before the intervention, but this difference immediately appeared after the intervention for control beliefs (t = 4.5, p < 0.0001) and perceived power (t = 3.31, p < 0.0001). Moreover, as it can be seen in Table 4 these differences also appeared two months after the intervention for control beliefs (t = 6.85, p < 0.0001) and perceived power (t = 4.39, p < 0.0001). To assess behavioral intention before the intervention, mean scores had no signi cant difference between the two groups (t = 87, p = 0.383), while this difference immediately appeared after the intervention (t = 0.68, p = 0.001), and two months after the intervention (Table 4; t = 2.49, p = 0.014). Mean scores of participants in perceived behavioral control in pretest, had no signi cant difference with both post-test (t = 0.93, p = 0.354) and follow-up stages after intervention (t = 1.62, p = 0.105). Figure 2 illustrate the changes in pretest, posttest and follow up stages for each of the components of fast food consumption. As the gure shows there is a signi cant reduction in all ve major components of fast consumption between control and experiment groups before and after the intervention.

Discussion
This study aimed to reduce fast food consumption using a parent-teacher interaction-based intervention program amongst high school students in Iran. Results of the current study indicate that there was a signi cant difference between the intervention and control groups in regards to behavioral beliefs and evaluation of behavioral outcomes, immediately and two months after the intervention. These changes indicate that the training intervention is effective for increasing students' negative beliefs and attitudes towards fast food consumption which is consistent with previous studies ( Subjective norms as the second effective factor on an individual's intention include the perceived social pressures an individual may experience to perform or not perform the desired behavior; in other words, it re ects the impact of social in uence on an individual's behavior (Glanz, Rimer & Viswanath, 2008). In this study, a training session was conducted for parents and teachers of students to in uence the student's subjective norms. Moreover, the parents received nutritional short messages on a daily-basis. Findings demonstrated that by involving parents and teachers in this intervention this could positively in uence the student's perceived subjective norms. This is consistent with previous ndings that parents play an important role in adolescent's food choices along with parental cooking styles, and expectations, in uencing the development of their children's eating habits ( Another concept of TPB is perceived behavioral control. That is the perception of an individual regarding how much he/she has control over his/her behavior (Glanz et al., 2008). As our ndings indicate, there were signi cant differences between the control and experimental groups in terms of control beliefs and perceived power immediately after intervention and 2 months later in the follow-up stage. This demonstrates that the intervention could improve student's self-e cacy to choose healthy foods as well as to show greater control over what they actually choose. This nding is consistent with previous studies that suggested perceived behavioral control can considerably shape behaviors associated with healthy foods (Armitage, 2005 We found a signi cant difference between the intervention and control groups for intention to consume fast food in posttest and follow-up stages. These ndings can be explained by the TPB theory. According to TPB, the intention is the most important determinant of human behavior. Recent researchers claim that intention could also be an important element of optional behaviors (Ebadi, Rakhshanderou, & Ghaffari, 2018). Therefore, any change in the intention for fast food consumption may directly in uence the actual behaviors associated with fast food consumption of individuals. This view is supported by previous studies ( fruits and vegetables were more likely to actually consume them. However, the link between intention and behavior is not always visible. This matches our nal nding that revealed there is no statistically signi cant difference between both groups at the behavioral level after the intervention. In agreement with this nding, there are studies that could not nd such an impact (Gao, Huang, Zhang, Liu, Feng, Liu, Sun, 2014; Anderson et al., 2005). Perhaps an explanation for the recent nding is that we should consider contextual factors (e.g. availability and accessibility to healthier food options in settings such as schools', shopping centers, restaurants, etc.) that could in uence the intention-behavior link before we expect to observe such a link.

Suggestions And Limitations
According to our ndings, several suggestions can be addressed. First, considering the crucial role of parents and teachers in shaping healthy food habits, future studies should always consider them as a key component of the training intervention. Second, parents and teachers should be trained on how to best transfer and express their knowledge to children and adolescents in regards to their disapproval of fast food consumption. As our ndings show, students can perceive this disapproval and change their own attitudes accordingly. Third, future studies are recommended to use multilevel interventions to prevent students from consuming fast food. For both students and parents, we suggest adding more professionally designed visual and graphical messages (eg. Infographics) or short video clips that could demonstrate the differences between fast food and healthy foods and their subsequent impact on the body. This study is not free of limitations. The use of self-report questionnaires to record the student's responses may increase the possibility of the social desirability effect. In other words, students may have responded to the questions in a way to satisfy members of the research team. The second limitation is related to sample size and type. Due to limitations we only sampled male students, which leaves out the female high school student population. Future studies are highly encouraged to include females as well. The third limitation is related to the socioeconomic pro le of the participants. We did not investigate the extent to which the social and economic status of the students may in uence the ndings. For example, wealthier students may consume more fast food because they have the nancial backing to do so. The fourth limitation may be related to the extent to which students show consistent avoidance behaviors to fast food consumption when outside of the school environment. Lastly, we believe that our study primarily focused on changes in attitudes and beliefs and future studies may consider the degree to which actual behaviors changed and which physical indicators, such as weight, may also in uenced by this intervention.

Conclusion
The proposed parent-teacher interaction-based intervention was found to be an applicable, timely and overall useful way to reduce fast food consumption among adolescent students. When students receive consistent and simultaneous messages from both parents at home and teachers at school regarding the negative impacts of fast food consumption on their health, they become more pliable to possess negative attitudes towards fast food and greater positive attitudes towards healthy foods. These ndings explicitly indicate that although high school students are mature enough to understand the difference between harmful and healthy foods, their actual attitudes and behaviors can still be in uenced and altered by parents and teachers.

Declarations
Ethics approval and consent to participate This study received an ethics approval from the ethics board of the Medical Science University of Shahid Beheshti located in Tehran, Iran (Ethic Code: 6727). Moreover, a written consent was obtained from the parents of each participant before the participant could take part in this study.

Consent for publication
All authors agree to publish this study in the journal of International Journal of Behavioral Nutrition and Physical Activity.

Availability of data and material
The data of this study will be available upon request.
LS and FD structured the manuscript, improved the English text and removed typos. MC and MG revised the rst draft. LS and FD improved the second version. MC, PH, and SR revised the second and third versions. All authors checked and con rmed the nal version.