Hacking biomedical science in 30 hours
On Sunday 10 December 2012 I had my first hackathon experience. A hackathon is essentially a computing marathon, in this case about 30 hours long, fuelled by pizza, sugar and caffeinated drinks with minimal time for sleep.
This first Wellcome Trust hackathon was a themed weekend around ‘open data’, exploring the vast array of biomedical data that is being collected about us all the time – in hospitals by clinicians but also by ourselves, online and through apps on our phones. These large datasets are potential goldmines of information for researchers and clinicians on everything from sleep patterns to exercise routines. Working with Wellcome Trust Public Engagement Fellow Kevin Fong and Rewired State, developers and designers were invited to develop web and mobile applications that would help make the scientific data interesting and understandable by the general public.
Some people worked alone, others formed teams – many formed ad hoc on the day. Some even did both! Similarly with project ideas, some came with an idea of what they want to work with, while others waited to see what emerged on the first day, piggybacking and collaborating on the fruits of the groups’ labour.
Despite losing some people to the winter sniffles, the developers managed to present 19 projects to our four judges: Tilly Blythe, Curator of computing and information at the Science Museum; Dallas Campbell of TV’s Bang Goes the Theory and Supersized Earth; Mark Thorley, Head of science information and data management co-ordinator at Research Councils UK; and our own Mun-Keat Looi, the Wellcome Trust’s Online Editor.
So after a long 48 hours of hardcore coding, on Sunday evening the teams mustered their last reserves of energy for their three minutes in the spotlight to present and (hopefully) demonstrate their apps.
From a phone app for hypochondriacs to a running app where the runner’s speed dictated how fast a film clip played on the individual’s portable device, the variety of ideas created from scratch was staggering. Picking winners was a tough task, but winners there must be for any competition. The competition was divided into three categories: Open Me — involving data we voluntarily provide (e.g. to apps); Open Research — hacks that use data from biomedical research, and; Open Data — hacks that use data collected by devices used by clinicians. There was also a special award for the hack deemed most useful to the public.
And without further ado, the winners were:
Open Me — winner: Boneplicator
3D printing is expanding. While the hardware is still expensive, there is an opportunity for developers to write open source software. This hack helps turn CT and MRI scans – normally the domain of doctors, radiographers and researchers – into a file that you can feed into a 3D printer to get a plastic model of your bones or brain. You can buy model skeletons today, but they are expensive and often representative of a ‘healthy’ body, not to mention impersonal. These tailored models could be useful as a teaching resource, as well as a way to engage the patient more with their data. It would allow the printing of damaged or diseased bones that training clinicians would rarely see. Furthermore, the technology could be adapted to print almost anything that will scan, such as soft tissue structures.
Open Research — winner: MEDIC (Monitoring and Early Detection for Intensive Care)
In an intensive care setting patients are often wired up to a vast array of machines with competing beeps and alarms. This hack brings together a variety of live data feeds into a single dashboard, allowing doctors and nurses to get a snapshot overview of all the patients currently on their ward. The MIT PhysioNet data, comprised of historical ICU data, forms the basis of the predictive algorithms that provide an early warning system using a series of thresholds, such as heart rate. This warning system should help intensive care units allocate resources effectively and decrease reaction times in a crisis.
Open Data — winner: HackBox
LifeBox is a non-invasive piece of medical equipment that checks the level of oxygen in a patient’s bloodstream. Healthy patients usually have an oxygen saturation level of at least 95 per cent. Oximeters like the Lifebox can detect changes as small as 1 per cent and sound an alarm as soon as an unsafe change is detected. Without one, clinicians often only notice signs of hypoxia (oxygen starvation) when a patient’s skin starts to turn blue, usually at an oxygen saturation around 85 per cent. However, clinical devices such as the LifeBox are often ‘closed’ machines, producing immediate readings but not recording the data, which could be useful if saved. The HackBox team reverse engineered the Lifebox to do this. They developed a system that both increases the frequency of readings and records then. Readings at more frequent time intervals would be of more use to hospital clinicians, while a logging function could produce large research datasets that could, for example, look at the effect of administering drugs at different time points during an operation.
Most useful to the public – winner: Slow Changes
According to a 2007 study, 88 per cent of those who set New Year resolutions fail within the first two months. This is mostly because many resolutions are too general with no measurable goal or because they’re trying to introduce too many changes at once. The ‘Slow Changes’ hack tries to help people make sustainable life changes by only allowing them to set one, modest, measurable goal at time. The user logs each time they achieve their goal, accumulating ‘ambition points’. Once a certain number of points have built up , the next goal can be set. This stepwise approach could help people break the cycle of failed resolutions and hopefully live a healthier, happier lifestyle.
Information, data and more about all of the projects is, of course, open source and can be found on the Rewired State website.
It was truly inspiring to see what the teams were able to produce in such a short space of time, Judge Dallas Campbell hit the nail on the head when he described the projects as ‘frighteningly brilliant work’. Hopefully the developers will continue to work on them so they’ll be available to you and me sometime soon.
Emma Rhule is a graduate trainee at the Wellcome Trust.