The robots might be taking away some of our jobs. But they’re also helping us solve some of our most challenging and life altering problems.
That’s what Ubenwa, an Artificial Intelligence solutions startup, is proving with its work in healthcare. The health-tech venture’s flagship product is software that detects asphyxia in newborn babies by analyzing the sound of their cries.
It’s the sort of groundbreaking work that you’d more readily relate with first world STEM programs. But for Charles Onu, who founded the startup, it’s not just about showing of fancy new tech. It’s about tackling local issues with cutting edge solutions.
He came up with the idea of detecting asphyxia- a problem with babies’ breathing –when he found that a staggering number of babies who died after being birthed had been suffering from asphyxia. The stark reality of this hit him while he was at outreaches with an NGO in local communities in Nigeria.
“About one million babies die each year from asphyxia, and about one million others suffer severe life-long disabilities,” Onu explained in an interview. “One of the reasons for the high casualty rates is that in most resource poor settings, the fee for diagnosis is too high.”
This set him on the path to finding a bridge that would help less privileged mothers bypass the high cost of quality healthcare, and have the illness diagnosed by cheaper, non-human, and precise systems.
Onu’s journey has since been a winding course through the world of software engineering, medicine, entrepreneurship and cross-border collaborations. It’s what he’s had to give to ensure that deaths from asphyxia become a thing of the past. At the moment, the solution he and the team at Ubenwa have worked on is showing great promise.
Ubenwa’s AI application relies on a machine learning system which takes in and analyzes the cries of newborn babies and returns diagnoses in just 10 seconds. It’s able to do this by ‘checking’ for abnormalities in the frequency patterns of babies’ cries. The system is said to have made the correct diagnosis in 95% of cases it’s checked- a high success rate by any standard.
It does seem like an all-round improvement on the current method, which is a blood test with results requiring further analysis by a trained doctor. Ubenwa’s AI application doesn’t need blood samples, so it’s non-invasive. It’s also far cheaper, and doesn’t require any special skills to use.
There’s more work to be done at Ubenwa, as the startup continues to move towards a full scale entry into Nigeria’s challenging healthcare space. But it’s already proving that it’s capable of transforming healthcare in at least one important aspect.
Featured image source: Ubenwa