I had to spend a couple of nights in hospital last year – the first time I’d been on a hospital ward in about fifteen years. Things have moved on since my last visit: being me, the difference I really noticed was the huge number of computers, often on wheely trolley devices so they could be pushed around the ward, and often only used for one task. There was one at A&E when I came in, used to check NHS numbers and notes; another for paramedics to do a temperature check (this was at the height of the Ebola scare). When my blood was taken for some tests, another mobile computer was hooked up to the vials of blood and the testing hardware right next to my bed, feeding back results to a database; one controlled my drip, another monitored my oxygen levels, breathing, heart rate and so on on the ward. PCs for logging and checking were everywhere. I’m sure the operating room was full of the things too, but I was a bit unconscious at that point, so had stopped counting. (I’m fine now, by the way. Thanks for worrying.)
The huge variety of specialised and generic compute in the hospital gave me something to think about other than myself (which was very, very welcome under the circumstances). Namely, how much all this was costing; and how you could use Raspberry Pis to take some of that cost out. Here’s a study from 2009 about some of the devices used on a ward. That’s a heck of a lot of machines. We know from long experience at Raspberry Pi that specialised embedded hardware is often very, very expensive; manufacturers can put a premium on devices used in specialised environments, and increasingly, people using those devices are swapping them out for something based on Raspberry Pi (about a third of our sales go into embedded compute in industry, for factory automation and similar purposes). And we know that the NHS is financially pressed.
This is a long-winded way of saying that we’re really, really pleased to see a Raspberry Pi being trialled in the NHS.
This is the MediPi. It’s a device for heart patients to use at home to measure health statistics, which means they don’t need daily visits from a medical professional. Telehealth devices like this are usually built on iPads using 3G and Bluetooth with specially commissioned custom software and custom peripherals, which is a really expensive way to do a few simple things.
MediPi is being trialled this year with heart failure patients in an NHS trust in the south of England. Richard Robinson, the developer, is a a technical integration specialist at the Health and Social Care Information Centre (HSCIC) who has a particular interest in Raspberry Pi. He was shocked to find studies suggesting that devices like this were costing the NHS at least £2,000 a year per patient, making telehealth devices just too expensive for many NHS trusts to be able to use in any numbers. MediPi is much cheaper. The whole kit – that is, the Pi the touchscreen, a blood pressure cuff, a finger oximeter and some diagnostic scales – comes in at £250 (the hope is that building devices like this in bulk will bring prices even lower). And it’s all built on open-source software.
MediPi issues on-screen instructions showing patients how to take and record their measurements. When they hit the “transmit” button MediPi compresses and encrypts the data, and sends it to their clinician. Doctors have asked to be able to send messages to patients using the device, and patients can reply to them. MediPi also includes a heart questionnaire which patients respond to daily using the touch screen.
Richard Robinson says:
We created a secure platform which can message using Spine messaging and also message using any securely enabled network. We have designed it to be patient-friendly, so it has a simple touch-tiled dashboard interface and various help screens, and it’s low cost.
Clinicians don’t want to be overwhelmed with enormous amounts of data so we have developed a concentrator that will take the data and allow clinicians certain views, such as alerts for ‘out of threshold’ values.
My aim for this is that we demonstrate that telehealth is affordable at scale.
We’re really excited about this trial, and we’ll be keeping an eye on how things pan out. We’d love to see more of this sort of cost-reducing innovation in the heath sector; the Raspberry Pi is stable enough and cheap enough to provide it.