At the peak of the Kala Patthar landmark, 5,500 metres above sea level in eastern Nepal, Jacob Cresswell, a team leader working for the United Nations Stop TB programme, lifts several layers of clothing to place a freezing laptop against his bare stomach.
Twenty minutes later, and with the help of a power bank, the computer is repowered and the expedition’s world record attempt for the highest deployment of artificial intelligence is back on track, which is being reviewed by Guinness World Records.
One of the Sherpas in the group grabs an item resembling a whiteboard and holds it to his chest, as another man aims an espresso-sized machine at his back, and starts beaming it with X-rays. The board collects the data and sends it to the now-thawed computer embedded with AI software. A message pops up half a minute later and the screening is complete. Without AI, the process would have to involve a radiologist and take several hours.
The group of health professionals views their expedition in May as a success. Their goal is to show that portable X-ray machines equipped with AI – each weighing about 25kg – could work in one of the world’s most remote locations. Already commercially viable worldwide, this technology is considered key to eradicating tuberculosis, or TB, which is a major health problem in many developing countries.
Mikkel Broholt, director of operations at Stop TB and a member of the expedition, says many people with the disease have gone undetected and unknowingly risk spreading the disease.
“To eradicate TB, we need to screen more people. If we’re able to carry an X-ray to a peak in the Himalayas, you can take it anywhere,” Broholt says.
AI has long held the promise of making healthcare more accessible, with chatbots in automated administration systems helping to free up precious time for doctors and nurses.
Google’s chief health officer, Karen DeSalvo, has likened AI to the discovery of penicillin.
It was the “most transformational opportunity of our lifetime to advance human health”, she said in June at the World Health Assembly in Geneva. AI was one of the hottest topics at the annual gathering of the global health elite.
DeSalvo described at the event a future “where geography, socioeconomic status, race, ethnicity and resource limitations no longer dictate health outcomes”.
One of the first market-ready AI applications for use in healthcare is computer-assisted diagnostics. Such tools are already being used for preliminary screenings in some developed countries as a triage.
The technology is also increasingly being deployed in developing countries where, due to a lack of radiologists, AI is expected to have a transformational impact on their healthcare services.
In a waiting room at a hospital in the Nepali town of Lukla, 63-year-old Pasang Yangji wakes up on a heavy wooden bench. Because she suffers from TB, it is not uncommon for her to fall asleep easily and for long hours.
“I come here every day for treatment,” she says in a barely audible voice.
Here in the high-altitude eastern province of Khumbu, coughing is such a common condition that it is called the “Khumbu cough”. Often, it is caused by dry and oxygen-lacking air, with some locals suffering from more severe respiratory issues. Adjusted for age, Nepal has one of the highest rates of chronic lung illness in the world, according to a study published in the peer-reviewed medical journal The BMJ in 2022.
Pasang Yangji contracted TB from her husband in 2017 and underwent two months of treatment, but she felt the daily hospital visits were preventing her from working. She stopped her hospital visits and persevered to work until three months ago, even as she was coughing all the time.
Then, her symptoms worsened and she restarted her treatment.
Stop TB’s Broholt recognises the pattern.
“The doctors we spoke with indicated that they didn’t receive a lot of TB cases. However, they acknowledged that this might not reflect reality, as they lack the means to screen and diagnose.”
The rugged landscape of Khumbu is only served by one doctor per 150,000 inhabitants, well below the global density of 17 physicians per 10,000 people, according to 2022 World Health Organization (WHO) data.
For screening, people need to go to Kathmandu – a trip that is costly and often takes so long that it risks patients losing their jobs. Even in the capital, resources are scarce. A mere 150 radiologists are available to analyse X-ray images in a nation of 29 million.
“With the help of new technologies such as ultraportable X-rays equipped with AI, we can quickly screen people and tell which cases are normal and which abnormal,” Broholt says. “It not only eliminates the need to travel to the big cities but also makes TB screening less dependent on highly trained medical personnel.”
Stop TB is working with some companies across the world to pilot field screenings. In Nepal, the organisation is partnering Qure.ai, a developer of healthcare AI. Qure.ai received a big boost during the pandemic, as its tools were used around the world to diagnose Covid-19.
“It was also a time when people started accepting and expecting to travel shorter [distance] for care. During Covid, we saw the hospitals started coming to us,” Qure.ai CEO Prashant Warier says. “Since then, our company has grown tenfold.”
Today, Qure.ai sells its tools to hospitals in several developed countries but Warier says the company’s goal has always been to help people living in remote and poorer regions.
“We started out in Mumbai, a city with some of the poorest and densest communities in the world. AI can solve pressing health challenges here and in a big part of the world.”
Early adoption of AI tools can enable emerging markets to scale up new technologies, Warier says. In future, AI can be as ubiquitous as mobile payments, he added.
“Today you can get by anywhere in India without cash, but in many other parts of the world you can’t,” Warier says.
Stop TB, a global partnership administered under the United Nations Office for Project Services, has been introducing portable X-rays equipped with AI in a handful of countries in the past few years. The technology has already been used to screen hundreds of thousands of people, in the process detecting several tens of thousands of abnormal chest X-rays.
Broholt is excited that the technology is soon ready to be rolled out on a bigger scale.
“TB is the largest infectious disease killer in the world, one that disproportionately hits the marginalised. AI represents an opportunity for investing in health equity.”
In Nepal, Broholt and his team have donated their X-ray machine to a local hospital, where healthcare professionals have been taught how to use it, mainly on chests for TB screening but also for other injuries.
“What matters now to bring the development further is a question of resources and political will.”
AI can unleash many potential breakthroughs for healthcare but hurdles remain. The WHO has been working for years to develop standards to guarantee the quality and safety of the new technology.
“We need to ensure that countries and regions can evaluate the various solutions being offered,” says Alain Labrique, WHO director for digital health and innovation.
This Week in Asia has spoken to healthcare professionals who expressed frustration at the slow implementation of AI tools. While Labrique can understand their sentiment, he cautions against hasty deployment.
“New areas such as digital health and AI, if implemented poorly or without due oversight, can lead to loss of public trust, resulting in even further delays in adoption and lost impact,” Labrique says.
The WHO has released a toolkit for the use of the new X-ray technology and has often singled it out to underscore the promise of AI in healthcare.
In the hospital waiting room in Lukla, Nepal, Pasang Yangji smiles meekly amid her discomfort and fatigue. She says: “I just look forward to living a normal life again.”
