M-Health apps that don’t suck.

Pretty much everyone in the mobile business loves “m-health” these days. There are a couple of reasons for this – they imagine there’s money in it, it’s great CSR/public relations/lobby fodder, and it’s the sort of thing futurists chinstroke about. But I think we’re all barking up the wrong tree.

Here’s the problem. The typical use-case is as follows – it’s something integrated in a special-purpose medical device, that is sold, and whose function is monitoring. What it monitors, in particular, is the patient or as we know her, the user, or indeed you. It is of course telling that you’re not considered the user. Partly, what’s going on here is that we’ve just repackaged 1980s SCADA technology. Here’s a SIM, here’s a cellular radio, and let’s have the flood level gauge text headquarters every five minutes. You’ll also note that there is no user interface.

The stereotype application could be defined as “bugging granny”. We’re going to check some metrics at intervals, stick them into a control chart, and then badger you about it.

Who does this really benefit? Obviously, device vendors, and telcos. It also benefits doctors, both in an absolute sense, and in a relative sense – a lot of this stuff would otherwise have been carried out by nurses, who are icky, girls, and in the union. It also attracts pseudomedical systems like insurance companies.

The underlying mental model here can be characterised as the surveillance-compliance model. The basic idea is that patients’ problems are their own stupid fault, and if they remain ill, it is because they didn’t comply with instructions. Monitoring generates metrics. They can be judged on the metrics and badgered. You get some lip service to patient autonomy, but to some extent all this does is to demand that the user internalises the surveillance. It is very depressing the extent to which doctors talk about “managing” their patients these days.

Now, there is obviously some truth to this. Giving up smoking is a really good idea, as is taking your damn pills. But it is also highly problematic. For one thing, it assumes that the problem is non-compliance. In that sense, it transfers your problem from the domain of reality – a physical problem to be solved – into the domain of morality – a statement about good and bad. Rather than being poor, stressed, addicted, etc, the problem is that you are wrong and a bad person. As a rule, this is normatively evil, and of course it only works if the problem is not actually a real problem.

Another problem is that it assumes that the medics are always right. They aren’t. And replacing the front-line carers with SCADA technology can be see as a way of hiding from criticism. You can shout at the box but it doesn’t record your voice and report that back.

Also, there is the surveillant temptation. If the surveillance itself doesn’t work – and it won’t, for reasons we have seen – the answer is usually to enforce it. Perhaps your sick pay might be cut, or your premiums hiked. I’m sure we can all name politicians who would fucking jump at this.

So, Cap’n Swing he say LUDD SMASH. Or…Stafford Beer, he say redesign system. What would a user-driven m-health app look like? One that was pro-you? One whose design brief began “Don’t be evil”?

Well, I would start by thinking about how to manage your environment in hospital. This is notoriously vital and difficult, and the current staff-solution is to get a really tough, intelligent, and assertive woman to visit every day and whip the piss out of the medics. But having someone like my partner around is a rare privilege and one which is unavoidably class-skewed, because it really helps to be of similar or higher social status than the doctors and also to know their code-words.

At this point you might think we’d stumbled into a TV show called Operationalise This!

But let’s consider this from a systems perspective. Most of the time, what is it that you’re actually doing by monstering the doc? It’s usually either trying to extract information, or else to intervene in the ward’s operational routine on things like hydration, food, sanitation, and nursing care.

Someone close to me was recently in hospital and needed various post-surgical care, which they were only insufficiently getting because they were in the wrong ward and being taken care of by people outside their specialisation. As a result they were in serious pain, and their recovery was making no noticeable progress. The problem was to find out what instructions had been actually written down and compiled into the ward routine, make the specialist aware of this, and get them to insert correct ones into the system. These are not, in fact, tasks that can’t be operationalised.

This is rather like Doc Searles’ notion of Vendor Relations Management or VRM as the opposite of CRM. The problem is getting a look at the invisible infrastructure of instructions, records, and organisation that defines the hospital environment.

When another relative of mine was seriously ill not so long ago, I got into the habit of photographing all the paper that crossed his bedside, so as to have copies for reference and also so as to distribute them to other members of the family. (At one point, I half-inched the case file and perched on a bed to photograph it page by page, like Olga the beautiful spy with her microfilm camera.)

Medical systems are very papery, and they do have their reasons – if you have to evacuate the flooded building into ambulances with the power out at 3am, you can just take the file along with each patient and you’ll be able to read it at the far end, and a lot of information technology still can’t really say that. However, there’s a lot of interesting stuff you can do with paper and mobile devices.

The really depressing thing here is that had we gone a different way in 2002 we could have had an Open311 or FixMyStreet for hospitals. But we don’t have sensible data exchange formats, privacy/authorisation models, or URIs for any of this stuff because Tony Blair said so, and by the time Gordon Brown was listening to people like Tim Berners-Lee and Tom Watson, the NHS NPfIT (like the National Identity Register) had become a bureaucratic-industrial monster so big it took a government-killing nuclear bomb to stop it. And the Tories…well, you’ve never seen people who are more suspectible to Atos Origin lobbying them to make sick pay conditional on adequate five-a-day performance, have you? Really? No.

Now, the other problem from my point of view is that my customers, over on the other side of the fence, still believe that this time, it’s going to be like they thought it would with ISDN and IMS and so on and so forth. Everything is a network service. Delivered by us. But I don’t think it will happen. As an industry, we have a terrible record of pushing things that make sense from our point of view but just suck, so badly that they never happen.

9 comments

  1. SimonF

    The Economist doesn’t agree with you; they had a an article in the last Tech Quarterly full of praise for m-health.

    I tend to be with you. It has been one of the holy grails of the mobile industry since the early days of 3G business planning. Management consultants used to fall over themselves writing papers and businesses cases with all sorts of tele-medicine (as it was know then) services, none of which I have seen any sign of.

    • yorksranter

      The Economist is reliably full of shit, and this is exactly the stuff they fall for.

      I don’t think this stuff is useless. But I’m concerned that we’re converging on use cases that are basically harmful, and stupid.

      • John Styles

        “The Economist is reliably full of shit, and this is exactly the stuff they fall for.” I will be remembering this phrase, it will come in handy essentially every time someone quotes something from the Economist.

      • SimonF

        Me, a Conservative? That would amuse my friend who is deputy leader of our county council. However having the same name I do occasionally get his emails and forward them and he seems a nice guy.

        No, my claim to infamy is through LPUK,

  2. Pingback: Health care as a moral issue | Wis[s]e Words
  3. Pingback: Sensor data and the monitored life | Stats Chat

Post a comment

You may use the following HTML:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>