Obesity, a rapidly escalating public health issue worldwide, affects over 650 million adults. It leads to serious health consequences such as cardiovascular disease, hepatic steatosis, cancer and Obstructive Sleep Apnea syndrome (OSA). Obesity also has several social repercussions impacting quality of life. These multifaceted health issues emerge from an intricate interplay of biological, environmental and psychological factors.
Managing obese individuals in a primary care setting is difficult and time-consuming. It requires addressing related conditions such as perceived stress, sleep disorders as well as mental health issues including anxiety and depression. All of these conditions can influence a patient's ability to lose weight and maintain weight loss over time. Moreover, factors such as eating habits, compulsive eating and physical activity levels also play critical roles.
In response to these challenges, the investigators have developed an online platform (Avittam®), a digitally-enabled, person-centred online platform designed not only for obesity but also for various chronic conditions including Type 2 diabetes, sleep disorders, childhood obesity, bariatric surgery pathways and even rare diseases like tuberous sclerosis. The online platform (Aviitam®) streamlines healthcare consultations by providing a library of self-administered questionnaires covering an array of factors such as nutrition, physical activity, sleep, mental health, quality of life and medication adherence (Supplementary data eTable 1). Operational since 2014, the platform serves nearly 3000 healthcare professionals and around 18,000 patients across France.
This study focused exclusively on those who had engaged with the online platform's obesity management pathways. It aimed to identify distinct psychological and behavioural clusters among individuals seeking weight loss, thereby informing a more personalised, psychologically supportive treatment strategy.
METHODS Study design This study is an exploratory, observational and cross-sectional analysis of data collected from patients using the online platform. The primary aim was to identify patient clusters based on various self-reported and clinically validated variables. The study was approved by the Institutional Review Board (IRB) of the University Hospital of Montpellier, France (identification number: IRB-MTP\_2023\_02\_202301328). All participants in this study provided informed consent by accepting the online platform's General Terms and Conditions of Use (GTCU), which explicitly state that their data may be used for research purposes. The online platform has received approval from the French data protection authority, the "Commission Nationale de l'Informatique et des Libertés" (CNIL), ensuring GDPR compliance and the confidentiality and protection of participants' rights. To maintain privacy, all data used in this study were anonymised. Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of the research.
Setting The study uses an online platform (my.aviitam.com) which enables patients to provide personal information, respond to health questionnaires (validated or specific to the platform) and receive educational content about their pathology. An automatic summary is generated by the platform to facilitate patient care by the practitioner. The platform is accessible throughout France and is free of charge for both healthcare providers and patients. Furthermore, the online platform is a patient follow-up tool, with patients and healthcare professionals having access to the platform for patient management.
Participants the investigators included adult participants (Age ≥ 18) with a Body Mass Index (BMI) ≥ 25 kg/m² who agreed to the terms of use for their data to be included in the study. Exclusion criteria were the non-completion of mandatory questionnaires and a delay of more than 90 days between weight measurement and questionnaire completion.
Variables The primary outcomes of the study were the psychological health, physical health and eating behaviours of individuals, assessed through scores on various questionnaires. These include the Hospital Anxiety and Depression Scale (HADS), the Perceived Stress Scale (PSS), the Epworth Sleepiness Scale, Morin's Insomnia Scale, the Intuitive Eating Scale-2 (IES-2), the Binge Eating Scale (BES) and a Physical Activity questionnaire. A secondary outcome was Quality of Life, assessed using the EQ-5D questionnaire, which was added nine months after the initial launch of the obesity pathway on the online platform.
The study considered age and self-reported BMI which was subsequently validated by healthcare providers.
Predictors for the study were the scores on the mental, eating behaviour, physical health and quality of life questionnaires.
The following potential confounders were accounted for in the analysis: age, sex and time interval between weight measurement and questionnaire completion.
Effect modifiers were not explicitly investigated in this study. Inclusion criteria for participation in the study included: (i) age, (ii) BMI≥25 kg/m², (iii) completing all study questionnaires (except the EQ-5D) and (iv) agreeing to the use of their data for research purposes by accepting the platform's terms of use.
Data Sources/Measurement All data were initially self-reported via the online platform and later validated by healthcare providers during clinical consultations. Data were collected over a period of 67 months.
Bias The data were initially self-reported but then validated by healthcare providers during medical visits to mitigate self-reporting biases.
Study Size All users meeting the eligibility criteria and with valid data were included in our analysis. the investigators did not pre-calculate the sample size due to the exploratory nature of the study and the availability of a large patient population.
Quantitative Variables Age and BMI were treated as continuous variables. The questionnaires provided scores that were treated as continuous variables for analysis.
Data Sources and Variables The variables included in our analysis were age, BMI and overall scores from all questionnaires. The questionnaires were used to assess the participants' mental and physical health status, eating behaviours, physical activity and quality of life.
To ensure that the data were complete, the investigators included only participants who had filled in all questionnaires, in addition to their age, sex, weight and height. As a result, there were no missing data.
Statistical Analysis The statistical analysis was planned in advance of the data collection, with the aim of identifying phenotypes of obese patients using clustering methods. The analysis primarily used non-parametric methods, such as the Kruskal-Wallis and Chi-squared tests, for data interpretation.
The investigators categorised BMI according to the WHO classification: underweight, normal weight, overweight and varying grades of obesity .
Initial focus was on patients who completed all of the necessary questionnaires except the EQ-5D and who provided their weight, height and gender data. A subsequent sensitivity analysis was planned on a subset of patients who also completed the EQ-5D.
An effect size analysis was also conducted to highlight the main differences between the patient groups identified.
All statistical tests were deemed significant at a two-sided p-value of less than 0.05. Adjustments for multiple testing were made using the Bonferroni's correction.
