Monday 24 November 2014

Ebola Crisis Hackathon

The Ebola Virus has killed almost 6000 people in West Africa since Dec 2013
Ebola has infected over 15000 people in West Africa since the start of the current outbreak in December 2013. The disease has an estimated case fatality rate of about 71%, and efforts to control the outbreak have been hampered by the political and economic situations in the countries affected (so far almost all cases have occurred in Liberia, Sierra Leone and Guinea). Problems have included, but are certainly not limited to, the extreme poverty in the region (including very limited infrastructure), a lack of trust in government officials, traditional death customs (e.g. physical contact with the dead) and the propensity of health workers to become infected with Ebola.

It was with these obstacles in mind that the Said Business School's Oxford Launchpad (which aims to support entrepreneurial endeavour in Oxford) organised and hosted the Ebola Crisis Hackathon, which aimed to "explore and solve 'pinch-points' in the response, care and management of the global response against Ebola". The hackathon (an exciting name for a workshop or study group, which has the consequence of labelling its participants as hackers, in this case making me sound much cooler and edgier than I am) was held on the weekend of the 8th-9th of November and brought together life-scientists, MBAs, software developers, physicians and charity workers.

Participants were encouraged to gather ideas over the week or so before the workshop and the event started on the Friday evening with brainstorming sessions. We worked in groups to generate and evaluate different ideas for solutions. The most popular ideas were then pitched to all participants, with groups forming around a few ideas.

While researching the problems around the diagnosis and treatment of Ebola, before the hackathon, I read about the potential for Ebola to be diagnosed using mobile apps. However I soon discovered that internet access was severely limited in West Africa (1.5% in Guinea, 1.3% in Sierra Leone and 3.8% in Liberia) which reduced the scope for using apps. However, mobile phone use is reasonably high (47% in Guinea, 52% in Liberia and 48% in Sierra Leone), and this encouraged me to think about developing similar systems using SMS text messages. I found that an open-source Django based package for writing SMS based systems had been developed and had been deployed in similar medical applications around Africa (RapidSMS).



Mobile phones are a primary communication method in West Africa


At the Hackathon, I teamed up with others who had had similar ideas. Given the severe lack of health workers in the affected region, we decided that a risk assessment tool would be particularly useful. Concerned individuals would text a number, and would receive a series of five questions on their symptoms and recent travel history. They would then  be categorised into several risk categories and would be given automated advice by the system and, in some cases, have their details passed onto agencies who would be able to help. We worked in three subgroups: one looking at the risk assessment questions, another looking at the obstacles to deployment and one implementing the system in RapidSMS.

We called our product SymptomSMS and were mentioned in an article on the Hackathon in last week's New Scientist.