ILLUMINATION-Curated | Apps | Diet & Behaviour
Is One Simple Mistake in Diet Apps Causing More Harm Than Good?
New expert guidance suggests app developers have a golden opportunity to improve lives.
The global weight-loss industry is worth $189.8 billion. Despite the vast sums of money pumped into the industry, “every single nation” has had increasing obesity rates over the past decade.
Analyses of global obesity rates are also likely to be underestimated.
From low-carb to the Mediterranean diet, numerous dietary approaches can help people lose weight. The challenge then, is supporting people with these behaviour changes so that they can reap the health improvements. As a registered dietitian, I’ve been helping people this way for over a decade.
Since I completed my PhD five years ago, technological advances in the digital space have led to novel behaviour change ‘interventions’. Today, smartphone diet apps are huge in the health landscape, with over a quarter of young people using them.
Thankfully, the quality of these apps have improved due to collaborative efforts between developers, health care professionals, and users. This ‘co-production’ method is now considered the gold standard.
Despite this partnership work, new expert guidance from the National Institute of Health and Clinical Excellence (NICE), in the UK, suggests most apps have been making a simple omission. And it’s at the cost of a vulnerable group of people.
How Diet Apps Apply Behavioural Science
I’m interested in UI (user interface) and UX (user experience) design, but I’m a dietitian and scientist, not a developer. My expertise focuses on the scientifically proven behaviour change techniques that we should integrate into programs.
According to a systematic scientific review, key behaviour change techniques for changing eating habits include:
- Self-Monitoring (such as food diaries, calorie counters, food barcode scanning)
- Goals and Planning (such as behavioural goals like cooking and structured meal planning)
- Feedback on an outcome of behaviour (such as, weight loss in pounds, kilograms, or bodyweight percentage)
- Implementing graded tasks, which is starting easier, and building more challenging, but achievable targets over time. A simple food example is to start with increasing fruit and vegetable intake by 1–2 portions for 1–2 weeks, then 3–4 portions in weeks 3–4.
- Adding objects to the ‘real-world’ environment (such as gift vouchers/codes for free or discounted foods). I’m particularly interested in how ‘augmented reality’ could enhance this behaviour change technique in future research.
- Person-centred and autonomy-supportive counselling: These traditional human-to-human interactions are beginning to have support from artificially intelligent chatbots and avatars.
Apps can quite easily integrate the majority of the above behaviour change techniques into the user experience. So, apps are particularly beneficial for health improvement, which explains their frequent use in the free-market.
Apps in a health care setting vs free-market
Health care professionals can advise on apps that take into account their patient’s needs. Including any behavioural and health risks someone might have. For example, ‘self-monitoring’ is one of the strongest behavioural predictors of weight loss success in people with obesity.
But for people with eating disorders, such close personal monitoring is a health risk. It can lead to excessive calorie counting and exercise, and trigger high-risk unhealthy behaviours such as binging, vomiting, and starving — not fasting!
The National Institute of Health and Clinical Excellence (NICE), in the UK, recently published guidance on ‘digital and mobile health interventions’ (including smartphone apps).
The document states that health care professionals should avoid recommending digital and mobile health interventions that include ‘self-monitoring’ components to anyone with an eating disorder history. The guidance extends the recommendation to people at ‘high-risk’ of eating disorders.
In a health care setting, medical history assessments and safeguards can prevent physical and emotional harm. In the free-market, such safeguards are often lacking.
Most individuals access smartphone apps straight through the App Store or Google Play. Aside from reading ‘user reviews,’ most people do not seek any advice about the safety or effectiveness of using a health app. This lack of reliable information leaves vulnerable groups, such as people with eating disorders, at quite significant risk of harm.
“I believe that if you show people the problems and you show them the solutions they will be moved to act.” — Bill Gates.
App onboarding: the golden opportunity to prevent harm
Onboarding is a process of introducing users to an app’s features and acquiring a user’s data to inform progression. Maitrik Kataria presents excellent onboarding examples that are worth a look.
Suppose you’re unfamiliar with the term onboarding. In that case, its purpose is to present an engaging content and a simple ‘sign-up’ method.
This process helps convince users an app is beneficial and worthy of their time. Apps could use onboarding to limit self-monitoring features to people with health risks. A quick ‘safety check’ alert could be implemented and trigger the simple question:
Have you ever had an eating disorder?
If the response is, “yes,” an app could close its self-monitoring features to the user and provide useful advice and guidance. That said, would such a simple approach be effective? Possibly not.
Such a question, in reality, is profoundly personal and confidential, which could incline people to lie. Especially given the mistrust people have with apps due to the limited data security and confidentiality assurances — see Owen Williams’ article. Also, such a question could disturb a user (with an eating disorder or not) and cause them to disengage from an app altogether.
You could argue this disengagement is a win regarding the health risks as a person will not access the app. But of course, someone could move onto a different app. One that does not include the intrusive question about eating disorders.
So, is there an indirect safety solution?
How design could improve lives — “safety features”
A research study of people daily self-weighing and using MyFitnessPal over 1-year, concluded that, “there was no adverse effect on disordered eating.” The caveat is the people in the study were of higher body weight, a body mass index (BMI) of 27kg/m² and over (classed overweight).
Many people accessing diet apps are “healthy” or underweight, not in the overweight and obesity categories. In these cases, excessive self-monitoring is the risky behaviour health care experts are keen to minimise. So, could a solution be to implement indirect app safeguards that monitor:
- Total daily screen time
- Number of walking / running steps each day
- The number of exercise sessions each day
- Frequency of inputting and reviewing calorie consumption
- Safe calorie intake over an extended period (weekly / monthly)
- Bodyweight and BMI. Muscle mass and body fluid — when an app pairs to a bioelectrical impedance scale. A technology which is more frequently in use now it’s becoming more affordable.
Deciding a tipping point that confirms ‘excessive behaviour’ is challenging, and frankly, requires research in the digital landscape.
An algorithm that integrates the combined monitoring data above (and possibly many others) could produce a series of safeguard messages. The features could include locking-out certain app self-monitoring components.
To my knowledge, such targeted safeguards do not exist, or if they do, there are no academic reports of them.
‘Built-in’ safety measures could minimise human error in the health care sector too. Expert guidance aside, the typical health care professional has a limited knowledge of apps, especially what behaviour change techniques they include. Automated app safety features have the potential to safely steer people at high-risk and with a history of an eating disorder.
Such features would also improve an app’s safety and security scores, leading to more frequent recommendations and listings by health agencies, such as the NHS Apps Library.
This current ‘mistake in diet apps’ presents a golden opportunity for app developers, especially if they can partner with users and health care professionals.
Blown out of proportion
I’ve heard prominent academics state that the negative consequences of obesity public health interventions on people with eating disorders is an exaggeration. I beg to differ.
The term ‘eating disorders’ includes the conditions:
- Anorexia nervosa
- Bulimia nervosa
- Binge-eating disorder (BED)
- Other specified feeding or eating disorder (OSFED)
Estimating eating disorder rates in a population is difficult because the stigma of the conditions often leads people to avoid disclosing their medical problem.
According to a US study, over 13% of girls were suffering from one of the above eating disorders by age 20. That’s more than 1 in 10 females!
For perspective, US obesity prevalence is about 2 in 10 by age 19.
Another study, this time in the UK, determined lower eating disorder rates. But they identified that people with eating disorders incur higher medical and financial costs. Specifically, there’s a need for long-term medical prescriptions and a higher likelihood of developing other conditions (comorbidities).
Vulnerable groups are usually a minority. The idea of purposely ignoring the needs of a vulnerable group because they are a minority is not a justification to do nothing. It is an example of discrimination.
“When we introduce discrimination of any kind, it’s anti-innovative.” — Professor David Isenberg.
A quick recap in 9 pertinent points
- Self-monitoring is an essential technique for changing eating behaviour in people with a higher body weight.
- Regular self-monitoring in people with a BMI ≥27kg/m² is not likely to be harmful unless they have a history of eating disorders.
- Excessive self-monitoring is a trigger for, and symptom of, an eating disorder.
- Healthcare professionals can avoid recommending apps with self-monitoring features to people with a high risk, or a history of an eating disorder.
- But most people access diet apps on the free market such as the App Store & Google Play — not while in a health care setting.
- Diet apps on the free market gives people with eating disorders open access to them. Misguided use of these apps could trigger excessive self-monitoring behaviours and lead to an eating disorder relapse.
- App developers are in a prime position to introduce safety features to steer people away from self-monitoring when needed.
- Algorithms could trigger safety features. The framing of questions and statements into the user experience requires research.
- The challenge of creating these safety features is a golden opportunity for app developers to improve the safety and effectiveness of diet apps.
Take-Home Message
Partnering of expertise between developers, health care professionals and users is essential to maximise health care outcomes and minimise safety concerns in apps and digital health interventions.
Diet and food apps are in the health domain. So they can pose risks as well as benefits.
Future research should explore if a direct question regarding eating disorder history during app onboarding is acceptable to users. Determining how such questions are framed and identifying users’ preferences could protect vulnerable people — and improve their lives by offering helpful guidance.
Users may find apps more acceptable if they have assurances that medical information is not shared. Transparency messages should include pledges that data is only in use to inform an app’s safety features.
When apps are a part of the free market, it’s our ethical duty to consider their effect on vulnerable groups. The new expert guidance by NICE is a stark reminder of potential harm to vulnerable groups. It is easy, but misguided to plan to help one group of people — such as people with obesity — without considering the consequences to others. This simple mistake could be causing more harm than good.
The weight-loss industry is worth $189.8 billion. So perhaps someone out there can fund some diet app improvements?
‘Built-in’ safety measures could minimise human error in the health care sector too. Expert guidance aside, the typical health care professional has a limited knowledge of apps, especially what behaviour change techniques they include. Automated app safety features have the potential to safely steer people at high-risk and with a history of an eating disorder.