When Babylon, a UK digital health group, began operations in Rwanda in 2016, there was already a growing interest in using artificial intelligence to improve the country’s medical system.
Since then, Babylon’s telemedicine service has registered 2 million users nationwide in Africa and conducts 3,500 consultations daily. Although many providers are expanding the use of digital healthcare worldwide, its advancement highlights the limitations and debates surrounding the new technology.
Although digital devices can support more people cheaply and effectively in extended health systems, critics have raised concerns about unequal access and claims about digital tools such as artificial intelligence may be exaggerated and unproven.
Millions of patients in industrialized countries already use online medical services and applications, and companies expect further growth. Babylon is expanding its operations in Africa, Asia and Latin America, while its German-based rival Ada Health is expanding in Tanzania.
Ali Bursa, the founder of Babylon, noted the potential for prevention rather than costly treatment: “These nations are likely not to make the same mistakes that our (health) systems have made for two centuries. They can focus on keeping people healthy, rather than investing in disease.”
From his headquarters in the UK, contracted by the Babylon National Health Service, Bursa agreed to leave for Rwanda after a meeting with President Paul Kagame – with the renaming of Babylon. This led to a ten-year agreement with the government and the local health insurance system. “It has a small population (12.5 million) and a working executive,” he said. “We chose something we could handle,” he added.
In early 2018, Babel announced “the first complete digital healthcare service like this in East Africa using artificial intelligence”. The service has a chatbot to “tap the power of doctors’ brains and put that power on a mobile phone for medical advice and testing.”
In fact, the system is still the oldest form of telemedicine, with plans to test artificial intelligence in the coming months.
Opinions differ on the effectiveness of the Babylonian system in the UK. A recent study by Penn State University researchers found that online symptom screening tools “do not have the functionality needed to support the entire diagnostic process of offline medical attendance”, with a limited scope and often focusing on specific diseases.
In a review of digital symptom screening tools worldwide, academics at the University of Sheffield in the UK write that they are mainly used by younger, more educated people and that there is little evidence of how well medical advice is being followed.
Shiven Piamukhama, CEO of Babel, says some people in Rwanda own smartphones. “This service is designed for basic mobile phones and using text messages and voice calls.” Instead of using robots to detect symptoms, most people send a text message requesting phone calls. Nurses contact them and send them to doctors for advice. When needed, patients will receive a code to follow the prescription or lab tests.
The benefits include quicker and easier access to doctors, even in remote areas, reducing waiting hours at clinics and increasing privacy.
Babel, a consulting firm called Dalberg, concluded that in 2018, there is potential for cost-cutting, including the creation of highly efficient electronic health records. Currently, Bymugama says the company faces additional costs as it searches the size of the economy from its global systems, including the need to store all of its data on a local cloud server hosted in Rwanda.
Dolberg warned that using PayPal would “slightly increase the risk of fraud by impersonation” compared to face-to-face counseling. It also highlights the need to adapt sign screening mechanisms to suit ‘local forms of health and disease and language and communication practices’.
Hela Azatsoi, head of Ada’s Global Health Initiatives, says the use of local health information and local languages is important for the accuracy of the algorithm. “Local epidemiology is essential.”
Barcelona of Babylon claims that Babylon initially did not collect such data for Rwanda for its organization. He adds, “People often raise AI because they want to get funding. The truth is, in the beginning (…) AI will completely surpass our imagination in the years to come, and in the short term it will completely deceive us.”
It is funded by most analytical companies in the field and is not published in peer-reviewed journals. They are limited, have severe limitations in the clinical conditions being tested, and often do not compare with competing products or patient end results.
Hamish Fraser, a researcher at the Center for Biomedical Information at Brown University in the United States, recently co-authored a review of various diagnostic tools supported by the ADA. He also says that regular independent assessments and clear data requirements from medical regulators are needed.
“Coffee trailblazer. Social media ninja. Unapologetic web guru. Friendly music fan. Alcohol fanatic.”