EBOLA has a new enemy: an army of hackers. Alongside health workers and fast-tracked vaccines, software developers are now part of the campaign, putting together novel tools which could save lives.
Earlier this month, teams of physicians and graduate students from various disciplines spent a weekend huddled round laptops and drawing boards in the Saïd Business School at the University of Oxford. The aim of the Ebola Crisis Hackathon was to develop software and systems to help West African communities devastated by the worst-ever outbreak of the disease.
One such tool is a program which automatically detects people appearing in a video feed and measures their pulse using image analysis. “The idea is to be able to do contactless triage,” says Paul Brodersen, a researcher in systems biology, who showed off the system.
The technique is called Eulerian video magnification and it works even with a basic webcam. Brodersen’s software – based on freely available code developed at the Massachusetts Institute of Technology and Quanta Research in Boston – works by amplifying the red tones in a person’s face. The subtle variation in redness over time gives away the pulsing of blood beneath the skin.
When combined with handheld infrared thermometers already in use in the affected countries, the system allows people to be quickly assessed and then sent for the most appropriate treatment at health centres. That’s because there is a characteristic slowing of the pulse as the fever seen in Ebola reaches its peak.
Garfee Williams, a Liberian doctor who works for the charity Africare, has been involved in the fight against Ebola since the current outbreak began in March. He is impressed with this form of triage. “It would boost the confidence of our health workers, who would know they are working with a tool that means they don’t have to have lots of contact with patients,” he says.
For the system to be useful, however, there must be health workers in place. With this in mind, another team at the hackathon has come up with a real-time map displaying the locations of Ebola cases and medical staff. The idea is to show where suitably trained health professionals need to be stationed in order to contain an outbreak most effectively.
Slowing the spread
Williams says that even small changes in the placement of doctors can have a significant impact on Ebola’s spread. “By just increasing numbers of health workers at some outposts from two to four, we see results like fewer people becoming sick in that area,” he says.
One member of the mapping project is programmer Chris Rex, who previously worked with the New York Police Department to create a live map of police officers’ locations. “If we can improve transparency and let people know where the resources are, you can do a much better job of helping them make operational decisions,” he says.
But before you can map where doctors are, you need to know just who they are. This is where team leader and medical doctor Alexander Finlayson of the University of Oxford, comes in.
Finlayson recently helped establish the MedicineAfrica social network, which covers Somaliland. He’s hoping to extend the platform to compile a database of West African doctors in the midst of the Ebola crisis, information which can then be plugged into the live map.
What about situations when doctors and nurses just aren’t available at all? In Sierra Leone, for example, there are 45,000 people for every doctor. A system which helps members of the public assess their own symptoms could be of help.
The idea is to use a simple questionnaire sent by text message to evaluate symptoms and place people in one of four risk categories. Basic questions such as “do you have a fever?” are followed by questions that look at epidemiology, such as “have you taken part in a burial ritual recently?” Coming into contact with the bodies of Ebola victims during funerals has allowed the disease to spread, so a simple yes or no answer, also sent by SMS, would move the respondent along a flow chart of risk factors.
Advice by text
“We categorise you into a risk group and then we give you advice,” says student Mark Gilbert. “Based on that, you stay in our database so we can provide more automated services. The patient’s telephone number acts as a unique identifier.”
Like Brodersen, Gilbert has relied on freely available software, in this case RapidSMS, which lets coders build SMS-based services that they control via an online dashboard.
All of the teams are looking to gain the support of an aid agency and, in some cases, additional funding to take their projects forward. Gilbert and his teammates, for example, have already contacted aid organisations such as Oxfam and Médecins Sans Frontières to try to fast-track their idea.
Hilarie Cranmer, technical adviser on Ebola to the non-profit International Medical Corps (IMC), says hackathons are a great way to speed ideas to the frontline of a disease outbreak. “Any innovation towards record-keeping or communications, those are fantastic,” she says, adding that knowing how to implement such schemes is the hard part. The IMC is holding its own Ebola hackathon in Boston next month.
For Williams, having a better way of understanding the crisis unfolding all around him is vital. “It would let us be much more responsive in the communities,” he says.