Globally, around 300,000 women and three million babies die each year during childbirth or post-childbirth complications. A significant proportion of these deaths could be prevented with better monitoring and care.
An accurate reading of what stage a pregnancy is at is a “crucial” measure in preventing the ‘big three’ of maternal and neonatal death – premature birth, preeclampsia, and stillbirth, according to Dr Aris Papageorghio.
“The traditional way of estimating gestational age is asking when a woman’s last period was – but that’s generally a very inaccurate way of estimating development,” he said. “For example, in places where malnutrition is an issue, women may have absent or irregular periods”.
Whilst traditional ultrasound technology is expensive – in the region of £50-£70,000 per machine, Intelligent Ultrasound’s product, ScanNav Fetal Check, costs just £5,000. It is also much easier to transport.
But it is the use of AI that could be the game changer. On average, there is one doctor per 10,000 people in Africa – in the UK, by comparison, there are 31.
If the new machines work as is hoped, the healthcare workers operating them will not need the technical skills required to interpret a scan as the device will do it for them.
“[Normally] measuring a baby is difficult – you need to have someone really quite trained to move the probe to exactly the right position, manipulate the probe in the right position, freeze the image, transcribe the measurements, etc,” explained Dr Papageorghio.
The hope with the new devices is that you will just place them on the mothers’ abdomen and the AI will reveal the stage of development, he added.
AI healthcare revolution
If successfully scaled up, this technology has the potential to save hundreds of thousands of lives across the continent – and beyond – but will it work?
There have been many examples of the new generative AI “hallucinating” and producing false and inaccurate results.
Also, the WHO recently released guidance on AI in healthcare, warning that the tech can be prone to biases if trained on overly narrow datasets.
Only last week, for example, an independent review commissioned by the UK government said immediate action was needed to tackle the impact of ethnic and other biases in the use of medical devices – especially imaging devives.
It found pulse oximeter devices, which measure blood oxygen levels, were less accurate for people with darker skin tones, making it harder to spot dangerous falls in oxygen levels.
Many people with darker skin tones are thought to have missed out on vital care and died during the pandemic because of this bias.
On the other hand, AI is being successfully used in many areas of medicine. The BBC reported this week that an AI tool tested by an NHS hospital trust successfully identified tiny signs of breast cancer in 11 women which had been missed by human doctors.
Dr Papageorghio said that no system was without risk but that without innovation in Africa and other developing countries many more babies would die.
“Any system will get it wrong – sometimes a Tesla car will crash on auto drive – and the same is true with these systems,” he said.
“But the question isn’t ‘is it better than a trained physician with really good ultrasound equipment?’ because of course it’s not. The question is is it better than no ultrasound at all – which many of these settings don’t have.
“If we were to build the ultrasound infrastructure we have in high-income countries across Africa, it would take 15 to 20 years – and what do we do in the meantime?”
Protect yourself and your family by learning more about Global Health Security