The present study protocol complies with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist (43). Additional file shows this is more detail [see Additional file 1].
Sample
The study will include adolescents between the ages of 13 and 16 from an educational center in the Valencian Community.
The inclusion criteria will include (1) ages between 13 and 16, (2) informed consent form signed by both parents or guardians and the adolescents, and (3) access to a smartphone with Internet connectivity.
The sample size was calculated using the free program G*Power (44). To determine the size of the effect of the predictor variables on the response, Cohen’s conventional tables will be used. To get a power of 80% and a medium effect size, 0.15 according to Cohen’s tables, the sample size required for our study is 68 individuals.
Study design
A randomized controlled trial will be conducted with adolescents who will be randomly assigned by gender to a control or experimental group.
The experimental group will perform cognitive training through the GGED-AD module; while the control group will use another module of the app, named GGNEUTRAL (Figure 1). The study will consist of a pre-assessment (T1) and then participants will have to use the corresponding application for a period of 14 consecutive days, for approximately 5 minutes a day. After completing all levels, a subsequent evaluation (T2) will be carried out, as well as a follow-up evaluation one month after having completed the app module (T3), in order to evaluate whether the possible changes are maintained over time. Figure 1 and Table 1 show the organizational chart of the study.
Figure 1
Organizational Chart of the Study
Table 1 Schedule of enrolment, allocation, intervention, and assessments within according to SPIRIT.
|
Enrolment
|
Allocation
|
Post-Allocation
|
Close-out
|
TIMEPOINT**
|
T(-1)
|
0
|
T1
|
T2
|
T3
|
ENROLMENT:
|
|
|
|
|
|
Eligibility screen
|
X
|
|
|
|
|
Informed consent
|
X
|
|
|
|
|
Allocation
|
|
X
|
|
|
|
INTERVENTIONS:
|
|
|
|
|
|
Experimental Group: GGED-AD module
|
|
|
|
|
|
Control Group: GGNEUTRAL module
|
|
|
|
|
|
ASSESSMENTS:
|
|
|
|
|
|
Primary Outcome:
|
|
|
|
|
|
EDBQ
|
|
|
X
|
X
|
X
|
Secondary Outcomes:
|
|
|
|
|
|
EDE-Q
|
|
|
X
|
X
|
X
|
BAS-2
|
|
|
X
|
X
|
X
|
PH-4
|
|
|
X
|
X
|
X
|
SISE
|
|
|
X
|
X
|
X
|
EDI-3
|
|
|
X
|
X
|
X
|
Note. T(-1) = enrolment; T0 = allocation; T1= baseline assessment; T2 = post assessment; T3 = follow-up assessment 1 month later; EDBQ = Eating Disorder Beliefs Questionnaire; EDE-Q = Eating Disorder Examination Questionnaire; BAS-2 = Body Appreciation Scale-2; PH-4 = The Patient Health Questionnaire; SISE = Single-Item Self-Esteem Scale; EDI-3 = Eating Disorder Inventory.
Intervention
The study groups will complete two different modules of the GGtude platform. The experimental group will complete the GGED module adapted to adolescents (GGED-AD) and the control group will complete the GGneutral module. Both modules are described below.
GG eating disorders – Adolescents (GGED-AD)
The GGED-AD is the adaptation of the GGED for adolescent, which is used to work on dysfunctional beliefs associated with eating disorders (45). The adaptation consisted of three phases. First, a group of researchers with expertise in eating disorders selected statements that may be difficult for teenagers to understand. Of the 708 statements that make up the application, a total of 131 statements were selected. Subsequently, an ad hoc questionnaire was created and administered to adolescents (n = 5, Mage = 12, SD = 0, 60% female). In this questionnaire, participants indicated their level of understanding of sentences on a visual scale with three response options (green: “I understand it perfectly”; yellow: “I don't quite understand it”; red: “I don’t understand it”). Subsequently, the teenagers had to propose an alternative for the sentences marked as “yellow”. Of the 131 statements selected in the first phase by the group of experts, 88 presented comprehension problems for the participants in this second phase. Finally, in the third phase, the appropriateness of the alternative statements proposed by the participants was evaluated by four psychology PhD researchers (Corberán et al., 2023).
Similar to the GGED module, the GGED-AD is made up of the topics that are normally worked on in CBT in relation to eating disorders (Fairburn, 2008). The final version of the GGED-AD includes the following topics: perfect physical appearance, fear of being judged, fear of body fat, body as an object, control of body weight, need for control, fear of complacency, self-esteem, fear of food-related thoughts, physical appearance in the media, fear of shame, negative internal image, and fear of abandonment. As in the adult's version, the sentences from these themes have to be "accepted" or "rejected" based on their adaptiveness. This module will be used by participants in the experimental group.
GG neutral module (GGNEUTRAL)
This a module belonging to the GGtude digital platform GGED designed exclusively for research purposes. The operation is similar, sentences appear in the center of the screen, but in this case, the person must accept those sentences that are true (e. g., "Madrid is the capital of Spain"), by dragging them to the top of the screen; and reject those that are false (e. g., "Tokyo is the capital of Spain"), by dragging them to the bottom of the screen. All the statements that compose it are totally independent of the target of the study, thus, there is no sentences related to weight or physical appearance. To use this module in the present study and facilitate its understanding, some sentences were modified to the Spanish context of the teenagers. For example, “Arizona is a state in the USA” was replaced by “Valencia is a city in Spain”. This module will be used by participants in the control group.
Outcome measures
Primary outcomes
Eating Disorder Beliefs Questionnaire (EDBQ; Cooper et al., 1997 (46)). It is composed of 32 items that examine core beliefs about weight, physical appearance and eating that are associated with eating disorders. Items are rated on an analog scale from 0 to 100, being 0 “I do not usually believe this at all” and 100 “I am usually completely convinced that this is true”. It contains four subscales: negative self-beliefs, weight and shape as a means of acceptance by others, weight and shape as a means of self-acceptance, and control over eating. In its original version, it presents Cronbach's α values between .94, .95, .90 and .89 respectively. Given that there is no validated Spanish version of this questionnaire, the present study will use the translation-retranslation method (47). First, a minimum of two native Spanish-speaking members of the research team will perform the forward translation of the original questionnaire into Spanish independently. Subsequently, a meeting will be held with the individuals who carried out the forward translation and the rest of the research team. During this meeting, a consensus on the translation of the questionnaire will be reached. This Spanish version will be sent to a native English speaker who, totally blind to the original version, will back-translate it into English. Again, the research team, composed by mental health professionals as well as psychometricians, will check that the back-translation and the original questionnaire match. In the present study, the psychometric properties of the instrument translated into Spanish will be calculated to check its reliability and validity.
Secondary outcomes
Eating Disorder Examination Questionnaire (EDE-Q; Fairburn & Cooper, 1993 (48); Spanish version: Peláez-Fernández et al., 2012 (49)). It is a 36-item self-report questionnaire that assesses the range, frequency and severity of behaviors associated with a diagnosis of an eating disorder. It comprises four subscales (restriction, food concern, weight concern and figure concern) which are composed of attitudinal and behavioral items. For the present study, we will only use 22 attitudinal items. Items are scored on a 7-point Likert-type scale, ranging from 0 (“No day/ No time/ Not at all”) to 6 (“Every day/ Always/ Completely”). Its original version and its validation in Spanish present an adequate internal consistency (Cronbach's α is .86, .75, .93 and .74 for the four scales).
Body Appreciation Scale-2 (BAS-2; Tylka & Wood-Barcalow, 2015 (50); Spanish version: Swami et al., 2017 (51)). It is a single-factor questionnaire that evaluates body satisfaction with 10 items on a 5-point Likert scale (from 1 = “Never” to 5 = “Always”). Both its original and Spanish versions present high internal consistency with Cronbach's α above .90.
The Patient Health Questionnaire (PHQ-4; Kroenke et al., 2009 (52); Spanish version: Lenz & Li, 2022 (53)). This is a 4-item measure of two factors: depression and anxiety. It assesses the symptoms experienced by participants during the 2-week period before taking the survey (two items are used for the main symptoms of anxiety and two for depression). The response format is a 4-point Likert-type scale (from 0 = “Never” to 3 = “Almost every day”), and its validation presents adequate psychometric qualities, with an internal consistency (Cronbach's α) of .86 (depression) and .91 (anxiety).
The Single-Item Self-Esteem Scale (SISE; Robins et al., 2001 (54)) is a self-report measure in which the person has to indicate how the statement “I have high self-esteem” describes him/her on a 9-point scale (from 1 = "Not very true for me" to 9 = "Very true for me"). This scale presents high test-retest reliability and a similar criterion validity coefficient with the Rosenberg Self-Esteem Scale (Rosenberg, 1965 (55)) above .80.
Eating Disorder Inventory (EDI-3; Garner, 2004 (56); Spanish version: Elosua et al., 2010 (57)). It evaluates 18 scales with a total of 91 items, with a 6-point Likert scale, ranging from 0 = “Never” to 5 = “Always”. However, in this study only the interpersonal insecurity (7 items) and perfectionism (6 items) factors will be used. These factors present Cronbach's α values of .93 and .75 respectively at its Spanish validation (Elosua & López-Jáuregui 2012).
Study procedure
First, an interview will be held with the school leadership team and/or heads of the guidance department of different schools in the Valencian Community, to explain the research objectives, describe the evaluation instruments, request permission to apply them, and encourage their collaboration. The schools interested in participating will distribute an information sheet specifying the objectives and characteristics of the study among the students and their families. Parents of students interested in participating in the study will have to sign the informed consent form authorizing their children to participate in the research. Students themselves must also sign it. The participants that enroll in the study will be randomly assigned by gender to the experimental and control groups using the Random.org program.
On the first day of the intervention at the school center, there will be an informative session for the students to explain guidelines for their participation in the study. This will include the importance of being quiet and the option to stop participating at any time. Additionally, there will be assistance provided for the administration of the first evaluation (T1). T1 evaluation will include the evaluation instruments, and sociodemographic data, which will include information on gender, age, school, grade and academic group, height, weight. Once T1 is completed, participants will download the corresponding app to start completing the daily levels. The experimental group will use the GGED-AD app, and the control group will use the GGNEUTRAL app, named GGA and GGN respectively to avoid giving excessive information to participants. Specifically, both apps are divided into 14 blocks, and the participants will do one block each day. Each block has 4-5 levels, which is equivalent to approximately 4-5 minutes. During this time, members of the research group will be present in the classroom on school days, to solve possible doubts and to control the use of the application. During non-school days participants will have to complete the app at home. After completing the app, all participants will carry out the second evaluation (T2). This evaluation will be composed of the same set of questionnaires completed at T1. Once T2 is completed, the adolescents will receive instructions to not use the application until we contact them again to conduct the follow-up after one month (T3) to evaluate if the possible changes achieved are maintained over time. If someone has difficulties or feels uncomfortable during the completion of the evaluation of the module, they will be free to speak to members of the research group who, in case of it being necessary, will be referred to the school counsellor, and in addition will show them different services where they could find help for possible mental health problems.
Blinding
Adolescents will be randomly assigned to the intervention or control group and both them and their teachers will be blind to their assignment. Likewise, the information that will be provided to the parents and/or guardians of the adolescents will include information on the content and purpose of the study, but the guardians will also be blind to the group, intervention or control, their children have been assigned to.
Statistical analysis and power calculations
Statistical analyses will be performed using the free software R. To avoid overly optimistic estimates of the efficacy of the experiment, an intention to treat (ITT) approach will be used, using the method of the multivariate imputation by chained equations algorithm (MICE) (58).
Descriptive statistics will be used to report means, standard deviations and frequencies. In addition, tests will be performed to evaluate differences between the different variables considered in the study. To predict the different response variables under study, generalised linear mixed model (GLMM) will be conducted. Specifically, a multiple linear regression model will be used, in which the outcomes measures are dependant variables, and the independent variables are gender (male/female/other) and time (T1, T2, and T3).
Data management and data monitoring
All self-report questionnaires will be completed by participants online via Lime Survey, a secure server of the University of Valencia. This server meets the security requirements of the GDPR. Participants will use a unique personal identification number (ID), so the data storage will be anonymous. The ID will be used to match the data. In addition, two control questions (e.g., “Please, answer “always” in this question”) are included throughout the evaluations to try to avoid random answers.