Context aware interventions for eating disorder prevention and improved body image

Project supervisors

Dr Roisin McNaney, Faculty of IT (Main Supervisor)
Dr Pari Delir Haghighi, Faculty of IT
Dr Gemma Sharp, Faculty of Medicine, Nursing and Health Sciences
Dr Jue Xie, Faculty of IT

PhD project abstract

Over 1 million Australians are living with an eating disorder yet less than a quarter of these will receive treatment or support. Eating disorders have the highest mortality rate of any mental health illness. Eating disorders are unique among mental health disorders in that they manifest in physical health complications, which can lead to serious and life-threatening illnesses such as diabetes, cancer, organ failure and even death if not treated. At least in 1 in 4 young people have serious body image concerns. Notably higher proportions of females (42.8% compared with 14.5% of males) were extremely or very concerned about body image. Moreover, Aboriginal and Torres Strait Islander females reported body image as the second most concerning personal issue, ahead of mental health. Furthermore, the rise of social media has seen the emergence of new concerns, e.g. orthorexia (a fixation on so-called ‘clean eating’). Since the majority of those experiencing disordered eating behaviours may not be underweight, it can be difficult to differentiate a desire to become ‘healthier’ from the emergence of harmful behaviours.

Research on treatments for eating disorders indicates that early identification and treatment improves the speed of recovery, reduces symptoms to a greater extent and improves the likelihood of staying free of the illness. It is important to acknowledge that developing an eating disorder is not a conscious choice. People suffering from eating disorders often do not understand the severity of their illness and are thus reluctant to seek help. It is critical to pursue early intervention strategies to prevent chronic malnutrition, long-term health complications and death. In other words, detecting and treating eating disorders as soon as possible has the potential to save lives.

Mobile applications often promote healthier living (e.g. calorie and activity trackers, healthy eating information), but can also become a source of obsessive negative behaviour, particularly for young people living in an age of ‘fitspo’ (fitness inspiration) perpetuated by social media. Mobile engagement itself poses significant challenges, with vulnerable populations freely accessing content without the extent of their engagement becoming visible to others. Furthermore, such applications are designed to promote continuous engagement, potentially fostering compulsive behaviours. This project will explore the development of context aware interventions for disordered eating behaviours and improved body image using a mobile sensing approach. It will investigate how negative app use behaviors (e.g. excessive social media use; photo filtering app use; healthy eating and fitness app use) might be mitigated through carefully co-designed ‘just-in-time’ interventions, delivered through the mobile. It combines the expertise of the supervisory team (sitting across the Faculty of IT and the Faculty of Medicine, Nursing and Health Sciences) in health co-design (Dr McNaney), context aware computing (Dr Delir Haghighi), software engineering (Dr Xie) and clinical eating disorders (Dr Sharp).

Areas of research

Mobile sensing, Context-aware computing, Co-design, Clinical eating disorders, Body image

Project description

This proposed work is a vital first step towards tackling the challenges related to negative body image and its potential development into eating disorders. By implementing early context-aware interventions to negative app use behaviours (e.g. obsessive logging of calories and fitness data; constant browsing on picture based social media platforms which can provide an unrealistic representation of the diversity of body shapes and sizes) it is hoped that we can counterbalance negative views of body image that young people hold, leading to better general mental health and wellbeing.

The PhD project will build upon an existing piece of work: Detecting Mental Health behaviours using Mobile Interactions (DEMMI). This project was led by the primary supervisor (Dr McNaney), with analytics phases led by Dr Xie, and explored whether we could indicate possible trigger points for binge eating or bulimic episodes, self reported by the user- through mobile sensor data (e.g. location, call frequency, apps use behaviours). We have existing code for a data collection app which employs the AWARE framework, an open source platform that facilitates mobile sensing data collection and analysis, and have an analytics pipeline for processing and interpreting the data.

The PhD project will extend this work by narrowing down to specifically focus on sensor features relevant to app use behaviors (e.g. how often, and for how long, users are engaging with certain social media, photo editing or healthy eating and fitness apps) and their impact on body image. Ecological Momentary Assessment (EMA)- repeated sampling of behaviours and experiences in real time- of mood, emotional response to app content and body image will be collected and used to frame the analysis phase. A period of data mining and analysis will be undertaken by the student to identify specific app use behaviours over time that are correlated with negative body image. Identification of potentially harmful app use behaviours will then allow us to understand where to intervene.

The student will have the opportunity to engage project stakeholders (the target population, eating disorder professionals) in a series of co-design activities that will aim to generate insights into possible countermeasures that might have the potential to address or mitigate negative app use behaviours and facilitate the realistic representation of young people’s bodies and eating habits. Possible examples include:
The implementation of an alert and information provision service linked to consistent calorie logging
The implementation of a body-positive pop up tool which displays images tagged by body positive campaigners on Instagram linked to excessive amounts of time on social media

Developing these into working prototypes and conducting an evaluation may be outside of the scope for the PhD project, however we will use this to form the basis of a funding proposal to develop the work further.

This PhD project is highly interdisciplinary in nature, sitting at the intersection between HCI, data science and clinical eating disorder prevention. The student will be able to draw from the interdisciplinary expertise of the supervisory team, who will support the development of their understanding of clinical presentations of eating disorders and body image issues (Dr Sharp), develop their capabilities in co-design (Dr McNaney), and extend existing technical skills that they bring to the project within data fusion (Dr Delir Haghighi) and software engineering (Dr Xie). The project fits within several ongoing programs of research that the supervisory team are working on. Aside from Dr McNaney and Dr Xie’s DEMMI project (described above), it also sits within a large body of work which is implementing mobile sensing with EMA  in student mental health within the Turner Institute (Dr McNaney consults on this project and has another PhD student working in this space). Dr Delir Haghighi’s work has been developing sensing technologies alongside EMA within clinical contexts and Dr Sharp has an ongoing program of work explicitly focusing on early digital interventions for eating disorder and body image issues through chatbot AI.

The pathways to impact for this PhD project are clear: 1) Early preventative interventions for body image and disordered eating and exercise behaviours can lead to better health outcomes and a lower risk of clinical eating disorders; 2) digital interventions have the potential to increase the reach of eating disorder treatment, with only 25% of people currently receiving support; 3) there are significant cost benefits associated with early and more optimal treatments (estimated $50 billion over the next 10 years); 4) this is a novel cross disciplinary research area with the potential to create a benchmark for future work, and place Monash at the forefront of this exciting space.

PhD student role description

Whilst psychology literature has been discussing the negative impacts of social media and calorie tracking applications for years, the role of technology in supporting people with eating disorders, and in particular how it might be leveraged to aid the prevention of eating disorders, is a novel area of research. The student within this role will have the opportunity to be at the forefront of this exciting research space, which has the potential to have real world impacts on the lives of young people. The student will have the chance to gain and develop skills in applied data science, co-design, and clinical intervention development, drawing on the expertise of a highly multidisciplinary team of supervisors and an extensive network of national and international experts.

The student will be responsible for implementing and building upon a current mobile sensing data collection and analysis pipeline which has already been successfully used to collect exploratory data with people experiencing binge eating and/or bulimic behaviours. With the support of the supervisory team, they will collect and analyse a novel dataset which focuses specifically on app use behaviours and their impact on body image. They will then use insights drawn from this dataset to frame a series of co-design activities, which will aim to develop unique design recommendations for implementable context-aware interventions to improve body image and disordered eating behaviours. The supervisory team will foster the student’s leadership and expertise development in this niche area of research.

The primary supervisor will meet with the student once a week and whole team meetings will occur at least once a month. The student will be able to reach out to any member of the supervisory team for additional meetings or email support depending on the stage of the project or expertise required. The student will be based at Action Lab in the Faculty of IT alongside Dr McNaney and Dr Xie, which is a highly collaborative interdisciplinary environment.

The student will be able to draw on the expertise of the supervisory team to learn and adapt state-of-the-art approaches and methods within data fusion and edge computing (Dr Delir Haghighi), software engineering and the AWARE framework implementation (Dr Xie), interpretation of complex mobile sensing data in relation to health outcomes (whole team), engaging people in co-design within a particularly sensitive topic are (Dr McNaney and Dr Sharp).

Improving mental health outcomes is a core priority for the Australian government so funding opportunities within this space are vast. This opens up lots of potential for a student wishing to pursue a future career in academia. If there is a preferred career pathway into industry within Digital Health, the student will be able to leverage the cross-disciplinary learning experiences gained from the PhD to enhance their opportunities for job prospects. Dr McNaney supported the development of the Digital Health Doctoral Training Centre in Bristol, UK and as part of this conducted numerous industry consultations (e.g. Philips, Microsoft, NHS Digital) to understand the skill sets desired for PhD graduates. Having cross-disciplinary experience and the ability to work and effectively communicate within interdisciplinary teams (e.g. clinicians, clients with lived experience, software engineers, designers) was one of the primary desired outcomes. The student will have ample training in this that will support their personal development in whichever career path they may choose.

Required skills and experience

The candidate needs to have previous experience with data collection and analysis, and programming skills, particularly with developing mobile applications. The student must be able to collect sensory data and analyse it on mobile devices using open source mobile sensing platforms such as AWARE, but will be supported in learning these skills through the supervisory team. An interest in mental health and Human Computer Interaction is key. Previous experience working on projects relating to mental health is desired, particularly if they have had a design component.