Data collection

Data collection is by online survey hosted via the Qualtrics survey platform (https://www.qualtrics.com/) and administered via the Prolific online research platform (https://www.prolific.co/). Prolific has a database of over 150 000 participants internationally, from which a specific sample can be recruited based on over 100 demographic characteristics. The survey used in this study takes approximately 45 min to complete, consisting of 250 questions across seven sections. These include demographics (including the impact of COVID-19, body mass index (BMI)); weight management and health service usage; eating behaviours; personality; mental health; and health-related behaviours. Participants are paid £5.00 per survey at each time point.

Substudy component

At each time point there is an additional substudy, aiming to validate participants’ self-reported weight in the main survey through collecting images of the individuals’ weight captured on a set of scales. The substudy survey is hosted by the REDCap survey platform (https://www.project-redcap.org/) and participants are paid £1 for completion of the survey which takes less than 5 min. This is an optional extra component and, if choosing to participate, is intended to be completed within 7 days of completing the main survey to limit the time between the two sets of data being collected. For those who choose not to complete the substudy survey, they are invited to complete an additional brief survey (three questions) to determine their reasons for non-participation. Participants are paid £0.20 p for completion of this additional survey. The conduct and reporting of this work will adhere to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting cohort studies.32

Inclusion criteria

Inclusion criteria for the study were young adults, aged 18–34 years at baseline, with a body mass index (BMI) ≥20 kg/m², and residing in either the UK or Australia. Although the age range of young adults was defined as 18–35 years, participants aged 35 years at baseline were excluded as they would be outside of this age range by the 12-month follow-up stage. BMI ≥20 kg/m² was chosen based on excluding underweight individuals (BMI <18.5 kg/m²) and those on the lower end of the healthy weight range who may be at risk of becoming underweight. Participants residing in the UK and Australia were of interest, in line with current research and policy direction as outlined in the introduction. Exclusion criteria were being pregnant, trying to get pregnant, breast feeding and not fluent in English.

Recruitment

Participants were sourced via Prolific, with invitations sent by the researchers to potentially eligible participants based on the inclusion criteria and their demographic characteristics as collected by Prolific. Interested participants first completed an eligibility screening questionnaire to confirm they met the above inclusion criteria before proceeding to the main survey and the optional substudy survey.

Sample size calculation

A target of 500 participants at baseline was set. A total of 500 participants provides adequate statistical power to detect a 0.5 kg/m² difference over time in BMI, assuming an alpha of 0.05, power of 0.80, correlation of 0.60, a baseline mean of 23.2 kg/m² and SD of 4.5 based on previous research,33 and accounting for 20% attrition. A set proportion of the 500 were recruited within each BMI, sex and country category, to ensure adequate numbers for future analyses by these characteristics. The target was 100 participants per BMI category (20–24.9 kg/m²; 25–29.9 kg/m²; 30–34.9 kg/m²; 35–39.9 kg/m²; ≥40 kg/m²), with 50% male and female and 50% from the UK and Australia within each BMI category. However, due to more participants on the prolific platform in the UK compared with Australia the target was adjusted to a ratio of 4:1.

Survey testing

The survey was pretested by the members of the research team, as well as our patient and public involvement (PPI) members who represent young adults with a lived experience of the psychological impact of weight management (n=3), and adjustments made prior to baseline data collection. Adjustments included reordering some questions and adding/editing instructional text for clarity. Collected survey responses will also be checked for completeness and quality, and rejected if either incomplete, or incorrect responses are entered for more than one of the attention check questions.

Sociodemographics

Sociodemographic data collected included age, gender, sexual orientation, country of residence, household income, highest level of education completed, ethnic group and whether participants were currently enrolled at university/college (including current grades if yes). Income and education questions were tailored to whether participants resided in the UK or Australia in terms of currency and educational systems. Questions were sourced/adapted from the Australian and UK census questionnaires.34 35

Impacts of the COVID-19 pandemic

Ten questions were included regarding the COVID-19 pandemic, and included asking whether participants were in lockdown at the time of participating in the survey, and whether they had been in lockdown within the past 6 months. Participants were also asked to rate the impact of the COVID-19 pandemic on their eating habits, weight, mental health, physical activity, sleep, alcohol intake, tobacco smoking and non-prescribed drug use over the past 6 months on a 5-point Likert Scale (eg, from Eating much less healthy to Eating a lot more healthy). The questions for alcohol intake, tobacco smoking and non-prescribed drug use also had the option of not applicable in the case of participants not engaging in these behaviours.

Height and weight

Participants were asked to report their height in either metres or feet and inches, and to indicate whether their height was measured by themselves, a health professional or was an estimate. Participants were asked to report their weight in kilograms, pounds, or stones and pounds, and to indicate whether their weight was measured by themselves, a health professional or was an estimate. Participants were also asked to rate their level of confidence in the accuracy of their reported weight on a scale of 0/Not at all confident to 100/Entirely confident, and to indicate how long ago their weight was measured. The questions used to gather height and weight are standard questions in biomedical research.36

Health service usage

Participants were asked to report which health services they had accessed within the last 3 months from a predefined list (General Practitioner (GP); Psychologist; Psychiatrist; Dietitian; Medical Specialist; Exercise Physiologist; Physiotherapist; other, please specify; I have not used any healthcare services in the last 3 months).

Weight management

Participants were asked the number of times they had tried to lose weight in the last six months, the number of years they had been trying to lose weight, the number of times in their lifetime that they had lost more than 11 lbs (5 kg) by dieting, and whether they were currently trying to manage their weight. Those who indicated yes to currently trying to manage their weight were asked two follow-on questions. First, whether they were receiving support/treatment for their weight management from one of a predefined list of options (GP; Dietitian; Pharmacy; Exercise Physiologist; Physiotherapist; Personal trainer; Health coach; other, please specify; I am not currently receiving any support/treatment for weight management). Second, whether they were using any products/diets for weight management (supplements; meal replacements; meal delivery services; support group; smartphone app or online support; very low calorie diet; low calorie diet; low carbohydrate diet; fasting diet; other, please specify; I am not currently using products/diets for weight management). Participants were asked whether they were currently using any medications that may affect their weight (steroids; antidepressants; contraceptive pill; other, please specify; not currently taking any medications that affect my weight). Lastly, an optional open response question was asked to gather their views on support options that would be most helpful in managing weight or eating behaviours in future (What support would you find most helpful in managing your weight or eating behaviours in the future?).

Addictive eating

Addictive eating behaviours were assessed using the Modified Yale Food Addiction Scale 2.0 (mYFAS 2.0).37 The mYFAS 2.0 is a 13-question tool which assesses addictive eating behaviours, and associated clinically significant distress and impairment. From this, the number of food addiction symptoms can be calculated (0–11), and it can be determined whether an individual is food addicted (≥2 symptoms and endorsement of the criteria for clinical impairment or distress) or not, and the severity of food addiction (mild; 2–3 symptoms, moderate; 4–5 symptoms, severe; ≥6 symptoms). The mYFAS 2.0 usually assesses behaviour over the past 12 months; however in this study participants are asked to think about the past 6 months in order to align with the study data collection time points.

Grazing

Grazing patterns were assessed using the Short Inventory of Grazing, which includes two questions regarding the presence/frequency of grazing, and of grazing with a sense of loss of control over the past 3 months.38 Questions are answered on a 7-point Likert Scale from 0/none at all to 6/eight or more times per week, then summed and categorised into mild (1–3 episodes per week), moderate (4–7 episodes per week) or extreme/severe (≥8 episodes per week) grazing.

Disordered eating

The Eating Disorder Examination Questionnaire 6.0 was included, which is a 28-item tool for assessing the range and severity of behavioural features of eating disorders over the past 28 days.39 Twenty-two of the questions relate to severity and are scored on a 7-point Likert Scale (0–6), while the remaining six questions assess frequency of key behavioural features (eg, binge eating) and are not included in scoring. The questions relating to severity cover four subscales (restraint; eating concern; shape concern; weight concern) which can be scored individually, as well as calculating a global score. Subscale and global scores range from 0 to 6, calculated as the sum of the individual question scores, divided by the number of questions. Scores of 4 or higher are indicative of clinical range symptoms.

Emotional eating

The 19-item Positive-Negative Emotional Eating Scale was included as a measure of usual eating behaviours in response to positive and negative emotions.40 Seven and 12 questions relate to positive and negative emotional eating, respectively, all scored on a 5-point Likert Scale from 0/Never to 4/Very often. Higher positive/negative subscale scores indicate higher likelihood of eating in response to positive/negative emotions.

Reward based eating was assessed using the Rapid Assessment of Reward-Related Eating Drive (RED-X5).41 The RED-X5 consists of five questions to assess current behaviour regarding overconsumption and preoccupation with food (eg, I don’t get full easily), answered on a 5-point Likert Scale (0/Strongly disagree to 4/Strongly agree). Responses are summed to give a total score (0–20), where a higher score indicates stronger reward-based eating.

Impulsivity

The Short Urgency-Premeditation-Perseverance-Sensation-Seeking-Positive-Urgency Impulsive Behaviour Scale was included.42 This is a 20-question tool assessing impulsive behaviour across five domains (negative urgency, lack of perseverance, lack of premeditation, sensation seeking, positive urgency), via items answered using a 4-point Likert Scale (4/Disagree strongly to 1/Agree strongly’). Each four-item subscale is scored by either summing (or averaging) responses to produce scores from 4 to 16 (or 1–4), with higher scores indicating greater impulsivity.

Delay discounting

The 27-item Monetary Choice Questionnaire was included to assess delay discounting.43 The questions ask participants to choose between a smaller, immediate monetary reward or a delayed, larger monetary reward X number of days later, with the amount of money and number of days varying for each question (eg, ‘Would you prefer £54 today, or £55 in 117 days?’). The amounts were expressed in pounds or dollars depending on whether the participant resided in the UK or Australia, respectively. The rate of discounting (k) is calculated, with a larger k indicative of a steeper discounting rate and greater level of impulsive choice.

Emotional regulation

The Sensitivity to Punishment and Sensitivity to Reward Questionnaire – Short Form was included,44 which consists of 24 yes/no response questions relating to usual reactions or feelings to each of the situations described. Responses are summed to produce scores for sensitivity to punishment (14-items; score range 0–14) and sensitivity to reward (10-items; score range 0–10), with higher scores indicating greater sensitivity.

Emotional regulation was assessed using the Difficulties in Emotion Regulation-18.45 This includes 18 questions, three relating to each of six subscales: lack of (1) Awareness of one’s emotions, (2) Clarity about the nature of one’s emotions, (3) Acceptance of one’s emotions, (4) Access to effective emotion regulation strategies, (5) Ability to engage in goal-directed activities during negative emotions, and (6) Ability to manage one’s impulses during negative emotions. Items are answered using 5-point Likert Scale responses (1/Almost never to 5/Almost always), which are then summed to produce subscale (score ranges 3–15) and total (score range 18–90) scores, where higher scores represent more difficulty in regulating emotions.

Stress

Stress was assessed using the Perceived Stress Scale 4 (PSS-4).46 The PSS-4 assesses an individual’s perception of stress in day-to-day life over the past month, using a 5-point Likert Scale (0/Never to 4/Very often). Scores are summed to give a total from 0 to 16, where a higher PSS-4 Score is indicative of higher perceived stress.

Anxiety

Anxiety was assessed using the Generalised Anxiety Disorder 7-item Scale (GAD-7).47 The GAD-7 assesses frequency of experiencing seven different symptoms of anxiety over the past 2 weeks, using a 4-point Likert Scale (0/Not at all to 3/Nearly every day). Scores are summed to give a total from 0 to 21, where a higher GAD-7 score is indicative of higher anxiety, and scores can also be categorised into minimal (0–4), mild (5-9), moderate (10-14) and severe (15-21) levels of anxiety.

Depression

Depression was assessed using the Centre for Epidemiological Studies Depression Scale 10-item screening questionnaire (CES-D-10).48 The CES-D-10 assesses frequency of depressive symptoms over the past week, using a 4-point Likert Scale (0/less than 1 day to 3/5–7 days). Scores are summed to give a total from 0 to 30, where a higher CES-D-10 score indicates greater depression symptoms.

Quality of life

Health-related quality of life was assessed using the EQ-5D.49 The EQ-5D is a six-question tool which assesses current health across five dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/depression) on a 5-point Likert Scale (1 /No problems to 5/Extreme problems), as well as asking respondents to rank overall health on a scale of 0 (worst health imaginable) to 100 (best health imaginable). An index value is generated from the five Likert Scale questions, by applying a formula and country-specific weights to each level within each dimension.

Dietary intake

Five short diet questions were included as key indicators of usual dietary intake. Questions were sourced/adapted from the Australian National Health Survey and New South Wales (Australia) Adult Population Health Survey.50 51 Questions assessed respondents' usual serves/portions of fruit per day, and vegetables per day (‘I don’t eat fruit (vegetables)’ to ‘six or more serves (portions)’), frequency of consuming a breakfast meal, and fast food takeaway meals or snacks, for example, burgers, pizza (1/never or rarely to 6/every day), and frequency of consuming sugar-sweetened drinks (0–7 days per week). Questions assessing fruit and vegetable intake specified serves or portions depending on whether the participant resided in Australia or the UK, respectively, in line with the relevant national dietary guidelines. These questions were accompanied by an explanation of what constitutes a serve/portion and picture examples.

Alcohol intake

Alcohol intake was assessed using the Alcohol Use Disorders Identification Tool-C (AUDIT-C).52 The AUDIT-C is a three-question tool assessing usual behaviour in terms of frequency of consuming alcohol, number of standard drinks consumed on drinking occasions, and frequency of consuming six or more standard drinks on a drinking occasion. Responses are scored and categorised into two levels based on the total score and participant sex. AUDIT-C Scores of ≥4 for men and ≥3 for women are categorised as hazardous drinking levels, or non-hazardous levels for scores below these thresholds. These questions were accompanied by an explanation of what constitutes a standard drink and picture examples.

Tobacco smoking and drug use

The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) was used to assess tobacco smoking and drug use.53 54 The ASSIST is a seven-question tool assessing lifetime and past 3-month use and consequences of use of tobacco, alcohol and other drugs including cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, opioids and ‘other drugs’. A risk score is calculated for each substance consumed (score range 0–36), and categorised as low (0–3), moderate (4–26) or high (27+) risk.53 54

Food security

Food security over the past 30 days was assessed using the US Department of Agriculture Food Security Survey Module six-item short form.55 This six-item tool focuses on economic access to food, including whether respondents: ran out of food, couldn’t afford to eat balanced meals, ever cut the size of meals or skipped meals (and how often), ate less than they felt they should, or went hungry, due to not having enough money for food. Affirmative responses are given a score of 1, and individual responses summed to give a total score between 0 and 6. A score of 0–1 indicates high/marginal food security, 2–4 low food security and 5–6 very low food security.

Physical activity

The International Physical Activity Questionnaire Short Form (IPAQ-SF) was used to assess physical activity.56 The IPAQ-SF assesses number of days and time (minutes and hours) per week spent in walking, moderate and vigorous activity, as well as the usual amount of time spent sitting on weekdays, over the past 7 days. Responses are summed to give MET (metabolic equivalents) minutes per week of physical activity, as well as to classify individuals as inactive, minimally active or highly active. Inactive is considered as engaging in less than the minimally active criteria. Minimally active includes meeting either of the following; three or more days of at least 20 min of vigorous activity per week; five or more days of at least 30 min of moderate activity per week; or five or more days of walking, moderate or vigorous activity and at least 600 MET-minutes per week. Highly active is considered as meeting either of the following; vigorous activity on three or more days and 1500 or more MET-minutes per week; or walking, moderate or vigorous activity on 7 days and 3000 or more MET-minutes per week.

Sleep

Sleep was assessed using the single-item Sleep Quality Scale.57 Respondents are asked to rate their overall sleep quality over the past 7 days on a scale of 0–10, considering hours of sleep, ease of falling asleep, waking during the night, wake time and how refreshing sleep was. Ratings relate to the following categorisations; terrible (0), poor (1–3), fair (4–6), good (7–9) and excellent (10) sleep quality.

Substudy survey

In the optional substudy survey, participants were asked to take an image of their weight recorded on a set of scales and upload this within the survey platform. Instructions were provided to participants on how to accurately measure weight (eg, place scales on a hard, flat surface), and how to take the image (eg, ensure the scale reading is visible), as well as a sample image as a visual aid. Participants were also asked to indicate the units of measurement of their scales (open response question). For those who chose not to participate and upload an image of their weight, they were asked to complete a brief survey consisting of three questions to determine their reasons for not participating. The first question asked whether they uploaded an image (yes/no/can’t remember) and two follow on multiple choice questions were asked for those that responded no or can’t remember, to determine their main reason, and any further reasons for not participating (eg, no access to scales, felt uncomfortable uploading an image of their weight, did not have the time).

Primary analyses

The primary research aims include:

1. To characterise subsets of young adults (18–35 years) of varying weight status by eating behaviours, mental health factors, personality, health-related behaviours and sociodemographics;

2. To investigate eating behaviours, mental health factors, personality, health-related behaviours and sociodemographics as contributors to weight change among young adults (18–35 years) over a 12-month period;

3. To compare health service usage of young adults (18–35 years) of varying characteristic subsets, for example, weight status, including changes over time.

The proposed analyses include creating static predictive models using baseline data (eg, using latent class analysis, factor analysis or similar), and mapping changes longitudinally (eg, using multivariate regressions). These analyses will enable changes in the study measures to be identified, as well as predictors and outcomes of change.

Secondary analyses

The secondary research aims and proposed analyses include:

1. To examine the shared and unique role of established substance use risk factors in food and substance addictions. The proposed analyses will involve a series of hierarchical linear and logistic regressions and/or multivariate analysis of variance, to explore predictors of food addiction, binge eating and substance use.

2. To determine the weight management support needs of young adults living with excess weight. The proposed analyses are mixed methods, with a thematic analysis of qualitative data and moderated logistic regressions exploring the relationship between weight status and service use and how this varies by gender, socioeconomic status and ethnicity.

3. To assess the impact of the COVID-19 pandemic on health-related behaviours in young adults. Proposed analysis will examine these factors using linear regression.

4. To validate self-reported weight against weight by image upload. The proposed analyses includes paired t-tests to evaluate differences between the two measures of weight, Pearson correlation to explore the strength of the linear relationship and Bland-Altman plots to evaluate the degree of agreement.

Additional exploratory analyses may be completed as indicated.