As seen on ArsTechnica.com
By: Liat Clark
An Australian doctor is raising funds to launch an SMS service in West Africa that sends people to the right medical facilities based on key words used and crunches that data to look for the next outbreak spot.
“During my missions with Médecins Sans Frontières I have always noticed that no matter how distressed the populations we served, someone always had a mobile phone,” Mohamad-Ali Trad, who has a masters in public health and tropical medicine, tells WIRED.co.uk. “We did some research and actually found out that most areas traditionally considered under-resourced do have a mobile phone coverage.” As mobile penetration on the continent continues to rise, SMS money transfer services like M-Pesa are common in parts of East Africa, and Western Union is hoping to capitalise on penetration in western countries to launch its payment service with MTN.
It is also certainly not the first time an SMS service has been used during a period of emergency or outbreak. Even in April, as Ebola began to creep from Guinea to its neighbours, SMS messages were used to raise awareness about symptoms and protective measures. A similar system has been used in the past during cholera outbreaks, most recently in Mozambique in 2013.
For Trad—who has worked with MSF in under-resourced areas and conflict zones including Haiti, Sudan, Syria, and Somalia—and his friend Raja Jurdak, who works in the telecommunications sector, the initial idea was to use a similar system to reduce A&E waiting times in Australian hospitals. Patients text code words such as “fever” or “cough” to a specific number and immediately get an SMS directing them to the most appropriate service. It would save on the lengthy recorded messages many establishments in the NHS favor but also ensure patients go where there is a doctor or a bed free. It’s the secondary element of the system Trad is proposing, however, that is perhaps most interesting. This, he tells us, “records geographical patterns of those code words.”
“For example, we will be able to tell how many fever ‘code words’ have been sent from a particular area, and hence this could act as a warning system for the nearby healthcare centers.” This could be relevant in any situation—be it a local flu outbreak in the UK or the start of the malaria season elsewhere.
In targeting communities with limited medical resources, the system Trad is proposing would focus on “the five big killers,” including respiratory infections, malaria, injuries, diarrhoeal illness, and child and maternity issues. For now though, it could be tailored for Ebola.
“With the reports we are receiving from West Africa, we thought that this concept can still be applied in that setting—for example, if there are no beds in health facility 1, the data processor can direct the patient to health facility 2, and so on. Patients who have no facility to go to could be mapped, and a new facility could be opened in certain areas, or distribution of home kits could target those areas… at the moment it is very chaotic from the reports that are coming out. We hope that this service could help both sides: the receiver and the provider of healthcare.”
The service will use that geographical information influx from users to estimate their “orbits of movement” and, combined with their condition symptoms, “predict the geographic extent of the disease-spread risk.”
Tracking outbreaks and their source is already something researchers have attempted to do using tweets. In some cases it’s been done successfully—but usually after the event.
“Another large scale retrospective study was run on malaria spread in Kenya using call data records and was able to identify the sources and destinations for the spread of the disease,” Trad tells us. “Our system will take the lessons from existing studies to deliver real-time local directions for patients in need of healthcare and would capture population-level symptoms for disease risk mapping and prediction.”
Trad started off by floating the idea on ResearchGate, getting comments and suggestions from researchers in the US, Holland, and Zambia. Along with Jurdak and research scientist at University of Southern Queensland Rajib Rana, a proposal has now been delivered to the World Health Organization. In it, the team points out “there have been numerous reports where probable Ebola-infected patients had to be driven away from health care facilities due to lack of bed availability.”
Undoubtedly there is a problem here that requires all the help it can get. Of course, when we already know that Ebola victims in Dallas and Spain had been sent home by medical professionals several times before their complaints were heeded, it’s hard to see how mistakes will not be made using an SMS service. But at the very least, it could redirect patients to facilities they know will have the capacity to serve them—it’s not necessarily about diagnosis, but logistics.
If their proposal gains traction, the trio tells WIRED.co.uk it can get working on the service immediately.