You may have seen this headline in the news a little recently…
As a resarcher who specialises in the study of technologies for autism, I was obviously intrigued.
In case you’re not famliar with this blog and website, I should just quickly mention what work I’ve done which I think qualifies me to comment on this news story. Starting in 2010, I led aproject to develop a new evidence-based therapeutic app for children with autism, working with an interdisciplinary team of experts. The project also involved extensive, collaborative work with children with autism, their parents, practitioners in health and education and so on. The app was released in January 2013 and has been downloaded by more than 100,000 users. More importantly, we also ran a randomised controlled trial of the app to evaluate whether it had demonstrable benefits to pre-schoolers with autism. It didn’t, and perhaps because of this null result we’re still struggling to publish our findings (though I hope they will be ‘out’ soon).
As well as that specific project, I’ve had three further app ideas licensed and these are being released on iTunes this month. I’ve written a couple of review papers on autism and techology and conducted studies in a school and with parents in the UK and Spain to explore hwo technology is being used at home and in the classroom. I publish app reivews on this site and on twitter (not recently though – sorry!) and I lead a special interest group on the topic of autism and technology, which meets annually at the International Meeting for Autism Research.
All this is not (just) showing off – I want to demonstrate why I think I’m in a position to say what I have to say about this headline and the video which has been released by Samsung to market the app, called Look At Me.
First, the claims. Samsung state that the app improves eye-contact in “60% of the children tested“. They do this in a highly emotive video in which footage of a mother and her autistic son is set against mournful piano music. They do not touch – the son doesn’t look at his mother. She says “I feel like I’m a stranger to my son“. The video then goes on to explain how a team of experts have gathered to try to harness the known preference for children with autism to interact with digital devices. The music moves up a key. The app is designed to help children with autism “make eye-contact, read facial expressions and express emotions” We see the boy playing with the app and hear from kind and intelligent experts. Finally, we see the boy interacting naturally with his mother – showing her a picture he’s just taken on his Samsung device, giving her a hug.
So, what’s rght with this? Yes, many – probably most – kids with autism particularly like playing with digital devices. I agree that this makes technology an excellent way to deliver supports for autistic people. Eye-contact is also often not a natural part of an autistic person’s social repertoire – something which can make the rest of us feel uncomfortable during interactions. The experts involved are, I am sure, well qualified and have conducted a decent preliminary investigation of the potential of a new app. The app itself looks good in the video, though I couldn’t test it myself as it kept crashing (a common problem judging by reviews on Google Play). The mother and son look happy and I am glad they feel their time enrolled in the research project was well-spent. I am sure that this project was a success academically. I hope that the group will publish their results in full and continue to work in this promising field.
So, what’s wrong with this? My central problem is with the huge mis-match between the evidence that has been gathered and the claims being made. During the video, a researcher states that “60% of children of the children tested showed improvement in making eye-contact”. This is accompanied by footage of a child doing an eye-tracking task – a good way to measure fairly objectively whether attention to the eye-region of a face has increased. But the footnote in the corner says “results are based on a parent’s survey”. In addition, it seems likely that no control group was recruited as they only mention the 20 children who played the app during the study. These two factors together make it highly plausible that the improvement being described is actually a kind of placebo effect – parents report the change they wish they saw in their children. Or, more subtle than that, they are motivated by the presence of the app and the observers to try new things with their child. Just as you might eat more healthily if someone watched you eat, even if you weren’t consciously trying to. Also, what happened to the eye-tracking task we saw in the video? This objective measure, much less prone to bias, should also have shown changes, if these were actually present. Moreover, looking at the numbers, 60% sounds impressive but this is, in reality, 12 children, versus 8 who didn’t show a parent-reported improvement. Not great odds.
Another claim is that “they can also identify emotional expressions more easily”. I don’t know what evidence this is based on – there’s no footnote for this part of the video, and the work isn’t yet published. But I do know that there are a number of well-evidenced computer games to help practice emotion recognition so I question whether we need another app for this – see FaceSay, Mindreading, and Let’s Face It! – all of which have good quality trial evidence. Which brings me to the Samsung element. By sponsoring this app and making it only available on Samsung devices, this company has the potential to boost their sales. The use of a highly emotive video to report the ‘scientific’ results strikes me as a rather exploitative advert, targeting the concerns of parents of children with autism.
Finally, here’s another quote from the mother in the film: “I also worry about how he’ll cope out there in the world, making friends or just doing simple everyday things…” This brings me to one last issue about the relevance of the ‘skill’ targeted in the app. Eye-contact is typically part of social interactions and something which many autistic people find unnatural, uncomfortable or downright stressful. However ‘training’ eye-contact seems beside the point. A bit like putting someone with a fever in an ice bath: their temperature might go down but the real problem hasn’t changed – though of course I’m not comparing autism to a disease, like a fever. But if we modify the eye-contact of someone with autism, have we really done something helpful? Some autistic adults report having been successfully trained to make eye-contact, but feeling consistently uncomfortable with it – like a leftie learning to write with their right hand. There’s no real evidence that improving eye-contact leads to better friendships or real world adjustment, and this particular study doesn’t report any effects in those domains (though they are implied by the way the film is constructed). If eye-contact makes the neurotypical community feel more comfortable, maybe it is our responsibility instead to learn to adapt to the interactive style of autistic people. If we’re so “socially skilled”, then why are we so bad at doing this?
In conclusion what I think this study does show is that an app has been made which aims to teach some behaviours which might not come naturally to people with autism. Preliminary evidence so far could be used to build a case to fund further research on the app, or to develop others addressing more fundamental abilities. However, what the video implies is that you can have a happier, more ‘normal’ relationship with your autistic child if you buy a Samsung device and use this app.