Machines are not likely to replace doctors, at least not in any foreseeable future, but through Artificial Intelligence (AI), efficiency in diagnosing patients, and administering healthcare can be improved.
This is particularly important for Nigeria, which has an acute shortage of medical workers in a country of over 190 million people who require medical attention. That it involves ‘advanced technology’ is no suggestion the country will be unable to get access to AI platforms designed for healthcare. Rather, existing, tested and proven systems can simply be deployed for use locally.
Katherine Chou, head of product for Google AI healthcare, said during a presentation, “AI can help with improving the quality and efficiency of healthcare. AI can also help with access to care, and what we’ve found is that we can actually do that without increasing the cost of care.”
READ ALSO: Huawei launches Africa Cloud & AI Innovation Centre
As Chou explained, AI can help fill in the gaps that doctors are not able to reach right now in providing expert care. Also, the doctor often has limited attention and can only examine limited conditions at any point in time. There are also devices being developed that are more sophisticated and complex, producing multidimensional datasets in for instance MRI, CT Scans, and this is getting harder to interpret. These are areas where AI can help tremendously with challenges and assist doctors.
One of the areas the Google AI healthcare started work was Diabetic Retinopathy, which Chou described as the fastest growing cause of blindness in the world, even though it is preventable. It is caused by diabetes with 450 million cases worldwide, and one out of every three patients suffers from Diabetic Retinopathy. Nigeria is said to record over 100,000 cases of Diabetic Retinopathy every year and there is a chance that deploying AI could help in managing the disease.
“What we have done is to develop models that help predict Diabetic Retinopathy,” Chou said. The team took several retina images, and working with over 50 ophthalmologists, created almost a million diagnosis and those were the labels put into the model.
Chou noted in her presentation that, there were multiple opinions of the same case by doctors engaged. This showed a huge variability across the doctors, it was found that doctors were consistent with each other 60 percent of the time, and were consistent with themselves, 65 percent of the time. This was subsequently discussed with the doctors to understand differences in views as a way of ensuring only very high quality data is fed into the models.
Uvie Ugono, co-founder of Smaart Health, an AI healthcare platform had also discussed with BusinessDay, some of the potentials for AI in making healthcare accessible to more people in Nigeria.
“We don’t have anywhere near enough doctors to actually do the job, and meet the level of demand which also grows by the years as the population is growing. At the same time, we are actually losing doctors. So what we are trying to do is bridging that gap through significant efficiency gains.
“By combining what we call man plus machine, that is, by having artificial intelligence do the basic diagnostic work, this saves the doctor 90 percent of the time it would normally take to get to the point of diagnosis. Which means consultation time can now be shortened,” said Ugono.
CALEB OJEWALE
