Design
This feasibility study is a three-parallel group, randomised, waiting-list control trial comparing ABMT vs MT for people who are overweight or obese. The protocol is outlined in Figure 1 and details of the assessments timepoints are given in Table 1.
Setting
The study will be carried out at the Institute of Psychiatry, Psychology and Neuroscience (IoPPN) at King’s College London and South London and Maudsley NHS Foundation Trust (SLaM), UK.
Ethical approval and trial protocol
Ethical approval for the protocol of this study v.2.2 dated 18th of October of 2018 was obtained from London - City & East Research Ethics Committee (REC ref: 18/LO/1683). The study is registered on the International Standard Randomised Controlled Trial Number (ISRCTN) registry (registration number: ISRCTN15745838).
Participants and recruitment
Participants will be recruited via advertisement on the KCL circular mail and social media, and on posters with general information about the study will be placed on notice boards at various KCL sites and public places. Adverts will also be shown on different online platforms such as the Beat website (the UK’s national Eating Disorders Association), the research team's departmental website and the MQ mental health charity website. In addition, suitable patients attending the Eating Disorders Unit at SLaM will be invited to take part in the study. Lastly, participants who have taken part in previous research and have consented to be contacted for other studies will be approached.
Table 1. Assessment timepoints
|
Screen Visit
|
Day 1
Pre-assessment
|
Week 1-8
|
Post-assessment
|
Follow-up
|
Patient information and informed consent
|
X
|
|
|
|
|
Data sharing Acknowledgement for Headspace®
(when applicable)
|
|
X
|
|
|
|
Eating Disorder Diagnostic Scale (EDDS)
|
X
|
|
|
|
|
Structured clinical interview for DSM-V (SCID)
|
X
|
|
|
|
|
Additional eligibility assessment
|
X
|
|
|
|
|
Visual Analogue Scale (VAS) for hunger, craving, mood, stress and anxiety levels
|
|
X
|
X
|
X
|
|
Eating Disorders Examination Questionnaire (EDEQ)
|
|
X
|
|
X
|
|
Power of Food scale
|
|
X
|
|
X
|
X
|
Mindful Eating Questionnaire (MEQ)
|
|
X
|
|
X
|
X
|
Mindful Awareness and Attention Scale (MAAS)
|
|
X
|
|
X
|
|
Depression, Anxiety and Stress Scale (DASS-21)
|
|
X
|
|
X
|
|
State and Trait Anxiety Inventory (STAI)
|
|
X
|
|
X
|
|
Weight, height and body composition
|
|
X
|
|
X
|
|
Experimental tasks (Food-Choice task, Bogus taste test, Food-challenge task)
|
|
X
|
|
X
|
|
Food-ANT
|
|
X
|
|
X
|
|
Visual Probe Task
|
|
X
|
|
X
|
|
ABMT/MT
|
|
|
X
|
|
|
Acceptability questionnaire
|
|
|
|
X
|
|
Eating habits, self-report body weight
|
|
|
|
|
X
|
Snack Diary (when applicable)
|
|
|
X
|
|
|
Demographics
|
|
X
|
|
|
|
Drink and food adherence
|
|
X
|
X
|
X
|
|
Inclusion Criteria
Male and female participants will be eligible if (1) they are 18 years old or older; (2) have a BMI of 25kg/m2 or higher; (3) are fluent in English and; (4) give written informed consent.
Exclusion Criteria
Participants will be excluded if they (1) have a diagnosis of a current other major psychiatric disorder (e.g., major depression, major suicidality, substance dependence, psychosis) needing treatment in its own right; (2) have a past or present DSM-5 diagnosis of Anorexia Nervosa (AN), Bulimia Nervosa (BN) or Other Specified Feeding or Eating Disorder (OSFED); (3) have a diagnosis of Diabetes Mellitus; (4) have recently started psychotropic medication or increased the dose (i.e. within the previous 2 weeks); (5) take medication for weight loss; (6) are pregnant (either current or in the past 6 months); (7) have a regular, current or past, mindfulness meditation or yoga practice (defined as > 20 minutes, twice or more times per week during the past 2 months); (8) have visual impairments that cannot be corrected with contact lenses or glasses.
Sample size
The sample size of this feasibility study is based on standard suggestions considering 12 participants per arm as reasonable sample size for a feasibility trial mainly because estimates of the standard deviation for normally distributed variables tend to stabilise around this size (34). Considering a sample size of n=36 and assuming the attrition to follow-up rate is a = 0.25, we strive for a sample size of n=45, i.e. 15 participants per group [using an attrition correction factor of 1/(1-a)].
Randomisation
Participants will be allocated to either the ABMT, MT or waiting list condition. Randomisation will be done by minimisation to control for BMI (obese/overweight) and gender (male/female). An independent researcher (not connected with the study) will perform the allocation using a computer programme for the generation of the random component. Allocation for balance between groups will be done manually and will be communicated via email to the researcher for each participant. Participants allocated to the waiting list condition will be given the option to take the electronic version of the training of their choice (i.e. no randomisation process) once the waiting time is over.
Intervention conditions
Participants assigned to either of the two intervention conditions will be asked to attend the Institute of Psychiatry, Psychology and Neuroscience eight times during an 8-week period (i.e. once/week). In addition, participants will be instructed to complete an internet-based version of the allocated training every day for 8 weeks at home (except for days when they have their weekly lab-based session).
For every lab-based training session, participants from both conditions will be asked to choose between three high calorie snacks. The chosen snack will be available to them to activate craving and reward/self-regulation cognitions during training. Visual analogue scales (VAS) for hunger and craving for the chosen snack will be assessed before and after the trainings to measure any potential changes. In addition, to minimise variability in hunger and satiety, participants will be requested not to eat or drink anything (apart from water) two hours before each training. Prior to this window, they will be advised to have a light meal and food adherence for each training session will be recorded.
Adherence to home sessions will be monitored via self-report (i.e. a training log) and through an output generated electronically at the time of each training. Other than these data on compliance, no further data from daily trainings will be collected.
Attention Bias Modification Training (ABMT)
The aim of the ABMT is to relatively implicitly train participants to “look towards” low calorie food and “look away” from high calorie food using a modified version of the anti-saccade task (16), while eye movements will be recorded to assess participants´ training accuracy. For home sessions, participants will be given a tablet (Asus ZenPad 10) with a website version of the ABMT training to use for a daily 10 min session. Participants allocated to this condition will be asked to record their daily accuracy rate on the training.
No side effects or risks of this type of training have been reported. However, because of the intense level of concentration required, participants could report fatigue. To minimise this, participants will be given small breaks (e.g. two minutes) between each block of the training (each training is formed by 3 blocks). We do not anticipate any other unintended effects of this training.
ABMT training paradigm overview: This version of a modified anti-saccade task consists of 360 trials in total. Of these, 180 require participants to look towards low calorie foods, and 180 trials require participants to look away from high calorie foods. For all trial types, a black fixation point appears for 100 ms, followed by a red or blue fixation point for 500 ms. A blue point indicates that a pro-saccadic eye movement is required (i.e. looking towards the food picture appearing after the fixation point), whereas a red point requires an anti-saccadic eye movement (i.e. to direct the gaze to the opposite side of the screen to where the picture appears). A blank screen gap is inserted for 200 ms between fixation point and picture presentation to speed up subsequent reaction times (e.g.(35)) and then a pictorial stimulus (high or low caloric food picture) appears on either the left or the right side of the screen for 500 ms. Inter-trial interval is 1300 ms. Trials will be presented in 3 blocks of each 120 trials.
ABMT paradigm trial types: Low-calorie cues are always preceded by a blue dot (indicating that participants should perform a pro-saccade towards the stimulus) and high calorie food cues are always preceded by a red dot (indicating that participants should perform an anti-saccade away from the stimulus).
Stimuli: Participants will view 30 low calorie food and 30 visually matched high calorie pictures per block. The position of each picture is counterbalanced for the presentation on the screen side: each picture presented once on left and right side of the screen, resulting in the overall 360 training trials (30 food stimuli + 30 non-food stimuli × 2 positions × 3 blocks). The order of trials will be randomized individually.
Feedback: Response latencies are recorded during the task to monitor accuracy and provide participants with feedback. For each block, number correct responses are summed up and presented as percentage score of correct performance to the participant (e.g. 70% correct performance).
Mindfulness Training (MT)
The app-based MT (provided by Headspace®) is 15min long and guides participants through a combination of breathing exercises focusing on the present moment. Participants allocated to this condition will be requested to sign a data privacy acknowledgement to authorise the company to share information with the study researchers at KCL.
This type of intervention could trigger some anxiety in participants whilst they focus on the present moment. To minimise this, participants will be introduced to the basics of mindfulness before the intervention. Participants will be given the option to stop at any point during the training should they wish to do so due to any unintended effects of the training.
Procedure: Individuals allocated to the MT condition will be given free access to the Headspace® app for 8 weeks. Participants will be instructed to download the app to their own mobile phones/electronic tablets and sign in with their newly created account. As a feature of the app, participants will be able to choose a time of day when they wish to receive a notification as a reminder of their daily meditation.
The MT is divided into two stages: The first 4 weeks focus on mindful eating meditation followed by 4 weeks focusing on overcoming cravings. Both stages (i.e. mindful eating and overcoming cravings) use a meditation technique called “noting”, which is an attempt to recognise those instances when the mind has wandered and identify the source of distraction, let it go and return to the object of focus (e.g. the breath).
Every training session (both lab-based and home sessions) will consist of 15 min of guided meditation. Both stages of the training (i.e. mindful eating and overcoming cravings) are divided into three competency levels: weeks 1-2 are the “Learn” level; weeks 2-3 correspond to the “Practice” level and week 3-4-are for the “Master” level.
Waiting list
After baseline assessment, participants allocated to the waiting list condition will be asked to wait for 8 weeks and, once they have completed the post-assessment after their wait, they can opt to receive the online version (i.e. daily home sessions only) of either ABMT or MT for 8 weeks. After 8 weeks of online training, participants will be asked to complete an electronic assessment including training adherence, current weight and a measure of eating disorder symptoms, eating habits, and general mindfulness. No further data will be collected.
Measures
Screening measures
- BMI (kg/m2)
- Eating Disorder Diagnostic Scale (EDDS) (36): A 22-item questionnaire used to identify the presence of an eating disorder (i.e. Anorexia Nervosa, Bulimia Nervosa, Binge Eating Disorder).
- Structured clinical interview for DSM-V (SCID) (37): This measure will be used to assess for any major psychiatric condition.
- Mindfulness practice: Previous experience with mindfulness techniques (i.e. > 20 minutes, twice or more times per week during the past 2 months) will be assessed.
Within session measures
Before and after each training session, participants in both conditions will complete a Visual Analogue Scale (VAS) measuring hunger and craving for the specific snack used in each session. VAS scales consist of a 10cm line and participants are requested to indicate a degree or level of a specific emotion or behavioural urge ranging from “not at all” to “severe”. After every lab-based training, participants will be asked to report what they ate or drank (apart from water) in the 2h before the training.
Outcome measures
A range of outcome measures will be included with the aim of determining the most appropriate ones for a larger scale RCT (based on effect sizes). The metric, timing and method of aggregation specifications for all outcome measures are described in the “Analyses” section of this protocol. They are in accordance with Zarin et al. 2011 (38).
Clinical outcomes
- Eating behaviour-related measures: BMI will be calculated measuring body weight and height (kg/m2) and body composition (primarily body fat percentage) will be assessed using a bioelectrical impedance scale (InBody S10). The Eating Disorder Examination Questionnaire (EDE-Q) (39) assesses specific psychopathology of eating disorders using a 36-item self-report format. The Power of Food Scale (PFS) (40) evaluates the psychological impact of the modern "obesogenic" environment. Dietary Recall over 24h will be used to assess the quality of participants' diet. VAS scales will be used to measure hunger and craving. In addition, VAS scales for cue-elicited food cravings will be administered after the Food Challenge Task (41) (short films created from Marks & Spencer’s adverts showing palatable foods). The bogus taste test (42) will measure food consumption of highly palatable food (i.e. crisps, chocolates and soft sweets) after asking participants to rate the food items according to smell, taste and attractiveness for 10 minutes. The difference in grams before and after the taste test will indicate the total food intake.
- Mindfulness and mindful eating-related outcomes: The Mindful Eating Questionnaire (MEQ) (43) is a 28-item questionnaire which will be used to assess awareness, distraction, disinhibition of eating and emotional and external eating. The Mindful Awareness and Attention Scale (MAAS) (44) is a 15-item scale to assess core characteristics of mindfulness.
- Other symptomatology: The Depression, Anxiety and Stress- Scale (DAAS-21) (45) is a 21-item self-report questionnaire which aims to evaluate mood, anxiety and stress levels over the previous week. The 20 items related to state-anxiety of the State and Trait Anxiety Inventory (STAI) (46) questionnaire will be used. VAS scales will also be used to measure current mood, stress and anxiety levels.
Neurocognitive outcomes
- The Food Attention Network Task (Food-ANT) will be used to assess three components of attention (i.e. orienting, alerting and executive function) using food (low and high calorie) vs non-food pictures (neutral items). This task was designed as reported by Hege et al. 2017 (47). Briefly, trials will include a fixation cross, a cue to indicate where the target would appear, and a picture (i.e. the target). In 70% of the trials, participants will be presented with a single arrow pointing either to the left, or the right side of the screen to indicate the position of the target. From these single arrow trials, 70% of the times the target will appear on the side where the arrow points (congruent) and 30% of the times the target will appear on the opposite side of the cue (incongruent).
On 15% of the trials, a double arrow will appear indicating equal likelihood of the target appearing on either side of the screen and in the other 15% of the trials, the target will be presented without a cue (i.e. no arrow).On double arrow and no arrow trials, the target will appear equally on each side of the screen.
Reaction times (RTs) will be recorded and the different components of attention will be measured by subtracting RTs as follows:
-Alerting: a subtraction of the RTs of double arrow (bidirectional) cue condition from the RT of the no cue condition.
-Orienting: Subtraction of RTs of the congruent directional cue from the RTs of the double arrow.
-Executive function: Subtraction of the RTs of incongruent directional cue from congruent directional cue condition.
- Visual Probe Task (VP) (22) will be used to assess visuo-spatial attention biases for food cues. Test-retest reliability of eye tracking indices (dwell time bias) and RT indices have recently been reported to be reliable measures of AB to food using this task (48). During the VP, two pictorial stimuli matched for visual characteristics (i.e. one food and one non-food item) will be presented simultaneously on both sides of a computer screen followed by a dot appearing on either the left, or the right side of the screen. Participants will be instructed to press the left or right arrow of the keyboard to indicate where the dot appeared. RTs will be recorded as an indirect measure of attention biases under the rationale that participants will be quicker to react to the dot appearing on the side of the picture they were attending to during the stimulus presentation. In addition, eye movements will be recorded as a direct measure of attention biases. The task will consist of two blocks of 60 trials each (120 trials in total). One block will present high-calorie food pictures against neutral items whilst the other will present low-calorie food pictures against neutral stimulus. The order of the blocks will be counterbalanced for each participant. The probability of the dot appearing behind the food picture or the neutral stimuli, as well as on the left or right side of the screen, will be equal (i.e. 50% on each side of the screen).
- A modified version of the Food-Choice Task (49) adapted for eating disorders (50) will be used to assess for food preference (i.e. low-calorie vs high-calorie). In the first two blocks of the task, participants will be presented with 43 food pictures (25 low fat and 18 high fat). The instructions will be to rate the pictures of food according to tastiness and healthiness respectively. Ratings will be measured using a 5-point scale: For the healthiness block, ratings will go from “unhealthy” to “healthy”, whereas in the tastiness block, the scale will go from “bad” to “good”. A food item rated as “neutral” on both scales will be used as a reference food for the third block. During the third block or the “choice” condition, participants will have to choose between their reference food and another food item imagining they would be presented with both options at the end of the study as a snack.
Acceptability and credibility
- A set of questions related to acceptability and credibility of the interventions will be provided. These questions will include items like “How credible did you find this training?” and “How useful did you find this training?”. Questions related to perceived benefits and drawbacks of the intervention will also be recorded to inform any unintended effects of the trainings.
Procedure
Screening
For participants recruited from general public spaces, study adverts will include the researcher's contact details for participants to contact them if they are interested in taking part.
Potential participants from a clinical setting will be identified by their clinical care team and ask for consent to pass their details to the research team. Alternatively, patients will be given an information sheet with the researcher's contact details and the option to contact them if they wish to.
Interested participants who either approach the researcher or are contacted by researcher, will be invited to have a phone call for the researcher to answer any queries they may have. To check for eligibility, participants will be asked questions using the screening measures described above.
Baseline assessment
If participants are eligible and wish to take part, they will be sent an information sheet and a consent form. Once the consent form is signed, participants will be asked to attend the Institute of Psychiatry, Psychology and Neuroscience to complete a baseline assessment including all the outcome measures described above in addition to demographic and clinical information including age, gender, ethnicity, education, history of previous and current treatments, diet, previous and current body weight. To control for hunger levels in the assessment session, participants will be asked to not eat or drink anything (apart from water) during the two hours before the session. Adherence to this request will be measured by asking participants to list what they ate/drank in the last 2 hours before the experiment.
Post-treatment assessment
The post-treatment assessment will be conducted after eight training sessions and will include all the measures used in the baseline assessment (apart from demographic information). In addition, to control for confounding variables, participants will be asked to answer a questionnaire about acceptability and credibility of the training as well as involvement in other weight-control intervention (including other research) during the eight weeks of training. A member of the research team will carry out assessment sessions with participants and will check outcome measures for completeness to ensure data quality.
Participants in the ABMT group will be asked to return the tablet after completion of the study.
Follow-up assessment
Four weeks after the last session (i.e. week 12), participants allocated to either ABMT or MT will be contacted via email and will be asked to complete an online version of the EDE-Q including a question on their current weight, the PFS and the MEQ.
For the waiting list group, the above questionnaires will be sent to participants at the end of the eight weeks of online training in addition to a question related to training adherence and involvement in other weight-control programmes. Participants will be compensated with £50 for their time and effort after completion of the study.
Withdrawal of subjects
One of the main outcome measures of this study is participant retention. We have a number of strategies to encourage continued participation of subjects (e.g. weekly email reminders and monetary compensation at the end of study). Should a patient decide to withdraw from the study, reason for withdrawal will be recorded and listed in the final report/associated publications. No further data will be collected from participants who decide to withdraw from the study.
Biological specimens
No biological specimens will be collected.
Blinding
Due to the heterogeneity of the study conditions, blinding of participants will not be possible. For resource reasons and given the feasibility nature of the study, we will also not be able to blind outcome assessors. It is our opinion that this is unlikely to introduce bias as main outcome measures are objective, biological or implicit measures, such as BMI, body composition and computerised tasks assessing e.g. reaction times. However, the data analyst will be blinded.
Analyses
Feasibility
Feasibility of the trial will be related to recruitment, as well as retention rates at both, post assessment and follow-up. Acceptability of both trainings and effect sizes of treatment outcomes will also inform the feasibility for a larger scale RCT.
Clinical and cognitive outcomes
Analyses will follow the intention to treat principle. Quality, completeness, and variability of the outcome measures will be determined by the use of descriptive statistical analyses and graphical methods. Group differences will be estimated using linear mixed effects regression models, controlling for the baseline level of the outcome and the strata variable used in the randomisation. The aim is not to determine significant group differences but to establish a suitably precise effect size for the primary outcome at the post treatment assessment. This estimate will be used to guide the sample size of a future efficacy trial.
The size of the treatment effect on each outcome measure will be the difference in mean scores between conditions at post treatment (8 weeks) and follow-up (12 weeks).