I have always been troubled by the “disconnect” between clinical practice and real life. Clinical evidence recommends a standard intervention, according to research, but it often flounders in the messy heterogeneity of the real world. This week I found myself worried about replicating this “disconnect” in the online world, especially with the addition of gamification, whilst strangely hopeful about the untested promise of natural language processing.
I recently had a discectomy. The rehabilitation process has been difficult and one of the strangest bits has been learning how to strengthen my core muscles. A lot of it seems to involve rocking one’s pelvis back and forth in a manner that I can only describe as suggestive.
For me, the desire to be back on the football pitch was enough to suppress my libido as my stupidly beautiful physio demonstrated how to strengthen my core. However, I know of an elderly Asian gentleman who had a similar operation, and while libido probably left him years ago the very idea that a young, female physio might show him how to move his pelvis is culturally ridiculous. Suffice to say his recovery has been limited.
I realise there are all sorts of issues at play here but the one that fascinates me is how unlikely it is that the recommended, standard intervention has any chance of leading to benefit across a heterogeneous population. It’s the well-known problem that one size really does not fit all.
We’re increasingly hearing more and more voices suggest services “go online”. The process seems to involve understanding what we do in the real world and then creating a digital version. But this approach clearly misses the opportunity to think about how things might be done better – more specifically how we might tailor services for different types of users.
As services go online there is the temptation to introduce gamification, the process by which competition can be used to engage users. The classic example of gamification is the website and mobile app foursquare in which users “check in” to places. When someone checks in somewhere often enough they become the “mayor” of that place, which confers benefits such as discounts on purchases. In essence, people get rewarded for their engagement.
Given the challenge of maintaining patient engagement, gamification is increasingly being seen as an attractive solution, perhaps even a panacea. This has troubled me for a while but I wasn’t always clear as to why until I saw this quote in an awesome presentation (slide 153):
“Things carry morality because they shape the way in which people experience their world and organise their existence, regardless of whether this is done consciously and intentionally or not. Designers…materialize morality.”
It made me think of that elderly Asian gentleman. The current real world service is simply the wrong “size” for him. However, if that service is transposed online andgamified, not only is it likely to remain the wrong size but by “failing” (and so not being rewarded) he’s ultimately being judged. Whether he’s conscious of it or not, this is likely to be demotivating.
Paradoxically, then, gamification has the potential to reduce engagement, the very opposite of what its proponents say.
I can’t help feeling the search for better engagement is a band aid. The real search needs to be to find ways to tailor off-the-shelf solutions to users. And that’s what makes me hopeful about natural language processing (NLP).
NLP is enabling computers to derive meaning from the words people use. I’ve been advising a few app developers of late and, as mentioned above, a lot of what they’re proposing to do is model real world practice and digitise it. However, if users could be encouraged to just say what they’re experiencing, say by leaving short comments each day, it could, through NLP, yield fascinating insights into patients’ experiences and hence how to tailor services to individuals.
This level of feedback and tailoring is the real promise of digital health. What’s most exciting, though, is that because NLP can work from what patients say in their own words it effectively empowers the patient without them having to learn medical jargon. It brings closer the promise of citizens as microexperts in health, as some have called it.
For me, it’s the opportunity to overcome the “disconnect”, to humanise care.
Competing interests: I do not have any competing interests in relation to this post. None of the app developers alluded to in this post will benefit from this post.
This post was first published on my original blog, Optimising Clinical Knowledge, and co-posted on BMJ Blogs.