#InTheSpotlight | Beyond the hype, what might AI actually mean for healthcare in SA?
With varying degrees of success, artificial intelligence has begun to play the role of research assistant, radiologist, health educator, and even therapist. In this Spotlight special briefing, Jesse Copelyn tries to see past the hype and pin-points the most immediate implications of these new technologies for healthcare in South Africa.
Artificial intelligence (AI) can perform a range of tasks that were previously the sole purview of qualified doctors, according to a growing body of research. For instance, several recent studies show that more tumours can be found during routine breast x-rays and colonoscopies (a procedure in which a camera is inserted into a patient’s colon) when AI detection software assists in scanning images.
Other research has looked at whether large language models (along the lines of ChatGPT) can correctly diagnose patients after being given information like their symptoms and medical history. Researchers at Google tested this, and found that an AI model that they developed could provide a list of possible diagnoses for real-world patients that were significantly more accurate than those offered by doctors who were provided with the exact same information. This is just one study and by no means the last word, but even so, it seems likely that doctors will increasingly turn to such models to help them get diagnoses right.
And new developments aren’t just confined to diagnostics. For instance, various health services in the US now deploy chatbots that do everything from answering patients’ medical questions to booking their appointments. In other cases, chatbots have even been developed to act as therapists, talking to patients via phone apps and providing them with techniques to overcome harmful thoughts. There are some signs that this kind of thing might be effective, but most studies in this area have been small so whether it works is far from settled.
Meanwhile, a range of AI tools are being used for medical research. This includes Google DeepMind’s AlphaFold, which can predict the structure of protein molecules. An Oxford research team that’s working to develop a more effective malaria vaccine used this software to model a crucial protein found in malaria parasites. This means researchers have more information about how to target the protein, which the parasite needs to reproduce.
The jury is still out on what exactly these various types of AI will mean for healthcare services and whether they will live up to the hype. For example, chatbots still occasionally generate false information, often with perfect confidence – such hallucinations can have devastating consequences in a healthcare context. It is unclear if, or when, this problem will be solved.
Chances are some applications will become a day-to-day part of 21st century healthcare, while others will fall by the wayside.
It is also still unclear what impact AI could make in developing country health systems. A technology that works well at a private hospital in the United States may not always be appropriate for a rural Eastern Cape clinic with unreliable power supply. Some AI products will inevitably cross the divide and flourish – in fact, some already have. This Spotlight special briefing takes a look at some examples.
AI is already helping tackle TB
The area where AI appears to be making the biggest impact in South Africa is in screening for tuberculosis (TB) – by some measures the country’s leading cause of death. The bacterial disease is usually tested by analysing a person’s sputum (saliva and mucus) using a molecular test, typically conducted in a lab. But because people who have TB in South Africa are often asymptomatic, many simply never get tested.
To get around this, screening initiatives have been developed in which health workers take vans containing mobile x-ray units to communities where the disease is prevalent. Residents go to the van-turned-clinic to get a free chest x-ray, which can reveal if they have abnormal looking lungs (even if they have no TB symptoms). If a person’s chest x-ray is irregular, they are sent for a sputum test to confirm the diagnosis.
But while conducting the x-rays is simple enough, the next step is often where projects face a bottleneck. Dr Emily Wong, an infectious disease scientist at the African Health Research Institute, explains that “you have to interpret those chest x-rays, and usually that requires a doctor or even a radiologist – which is a medical specialist – and [they] are very scarce in South Africa”. Indeed, the number of radiologists serving the public sector is currently a quarter of what’s required.
It’s here that AI comes in.
In 2021, a study showed that a series of AI-based software applications were not only able to detect TB in x-ray images as accurately as experienced radiologists, but were actually outperforming them. When Wong and her colleagues tested one of these AI products in a TB screening project in northern Kwazulu-Natal, they found that the software was roughly as accurate as doctors.
These AI-based tools – collectively known as computer-aided detection (CAD) – typically work as follows: A digital x-ray image is captured on a computer. The CAD software analyses it and gives it an abnormality score ranging from zero to 100, where a higher number indicates an increased chance of TB. Just like in the case of the radiologist, these scores aren’t definitive; a person with a low score might still have TB – there’s just a smaller chance that they do. Health workers decide on an appropriate threshold value (for instance 50) and anyone above that number is sent for a sputum test.
This threshold score is significant. If it’s set very low – say at 5 – then virtually everyone with TB will be sent for testing, but so will many other people with relatively normal and healthy lungs. This means expensive sputum tests and scarce laboratory capacity will quickly get used up on people that didn’t need to get tested in the first place. But if the threshold value is set too high – say at 95 – then more people with TB will be missed, since at over 95, only the most extreme cases will result in testing.
In the 2021 study, it was found that at a threshold score of 60, the top performing CAD tool, called qXR, captured 90% of TB cases, while 74% of TB-negative people were correctly categorised as negative (the remaining 26% were incorrectly identified as abnormal and sent for further testing). By comparison, the human radiologists only captured about 89% of cases and classified 63% of the TB-negative people correctly (these values varied depending on the classifications used but they were always less accurate than the CAD).
Many of the CAD tools are based on deep learning, meaning that they identify patterns in large amounts of data. For instance, CAD software is trained on thousands of chest x-ray images, where each image is labelled as “indicating TB” or “healthy”. As it’s fed more labelled data, the algorithm identifies various features that are associated with TB – for example a more asymmetric chest x-ray image means a higher likelihood of the disease. It’s then tested on unlabelled data to see whether it can make accurate predictions.
Such models are quite different from large language models like ChatGPT – and while not perfect, they do not have the same problem with hallucinations.
How are these tools being used in South Africa?
In South Africa, CAD software is currently deployed in various mobile chest x-ray programmes sponsored by international aid groups. Dr Jody Boffa, a scientist working at the TB Think Tank, which advises the National Department of Health, explains: “Global Fund and USAID fund the machines, but then a variety of implementing agencies [typically NGOs] are actually taking them out into the field”. In turn, the health department “sets the rules” for how these programmes should operate.
Dr Elias Ramarumo, who works in the National Department of Health monitoring these projects, tells Spotlight that 38 CAD software products had been procured by Global Fund, while 8 were bought by USAID. Additionally, he notes that provincial health departments were “in a process of procuring digital x-ray units”, which will “come with CAD software”.
Currently, two CAD products are being purchased by funders, Boffa notes. The one is CAD4TB, which is owned by a Dutch company called Delft, while the other is qXR by the Indian venture, Qure.AI. These CAD products were the two top performers in the 2021 study, which tested five different tools. Ramarumo adds that both companies are working with South African partners: Delft is working with Lomaen Medical, while Qure.AI is working with Vertice MedTech.
According to Boffa, the products are being used in two kinds of screening programmes. In one, vans containing mobile x-ray devices are parked next to overburdened clinics, where people can have a CAD-assisted X-ray screening. In the second case, the vans are taken to “hotspot” communities, and are placed in areas “where people that are less likely to come to the clinic would be found”. Before the vans arrive, project staff or community leaders will “rally the area to let them know that they’ll be coming in”.
AI to help miners
While international funding agencies are currently the driving force behind AI-assisted screening in South Africa, the National Department of Health says that it’s also planning on using CAD software – not only for tackling TB but also for silicosis. This is a lung disease that is caused by breathing in silica, a mineral found in sand and stone, which miners are often exposed to. Unlike TB, there’s no laboratory test which confirms its presence – an analysis of a person’s chest x-ray is final.
While there isn’t a cure for the disease, miners who are confirmed to have silicosis are provided financial compensation by the Medical Bureau for Occupational Disease, run by the health department. It’s the job of the bureau to determine who deserves recompense. However, it hasn’t always been able to manage claims fast enough, according to Professor Rodney Ehrlich, an occupational medicine specialist at UCT. “By about a decade ago, the completion backlog [of unpaid compensation claims] was over a hundred thousand, and all these paper files were piled up in back offices,” he explains.
Getting through claims requires enormous staffing capacity, he says, because a panel of doctors is required to analyse each x-ray image – the opinion of one doctor isn’t considered good enough.
It’s thus no surprise that the bureau is turning to AI, which has shown promise in this field, much like in TB screening. A 2022 study co-authored by Ehrlich analysed the ability of CAD software to detect silicosis and TB in chest x-ray images of North West gold miners. Despite concerns that the AI would not be able to distinguish between the two diseases – which can have similar presentations in x-rays images – the CAD products were able to make similar classifications as those made by doctors.
The health department and its partners hosted a workshop in June which was designed to “forge a way towards harnessing AI” for silicosis and TB screening in “mining, peri mining and labour-sending communities of SA”, according to Ramarumo. The event included academics, CAD companies, the World Health Organization as well as health departments from the Southern African Development Community. Ramarumo says that it was agreed “the adoption of CAD systems for TB and silicosis in the mining sector is essential to enhance diagnostic accuracy, improve patient outcomes, enhance the compensation process and reduce the financial burden on the [mining] industry”.
Are we using the right data?
While researchers who spoke to Spotlight are excited about the capacity of AI to make our public health system more efficient, there are also a host of issues that need to be overcome with the tech. One is that the data used to train AI isn’t always appropriate for our current context.
For instance, these days screening programmes in the country often try to find TB-positive people who do not yet have TB symptoms – as these people are often already infectious and at risk of falling ill. But for a long time such ‘subclinical’ TB didn’t accord with our traditional understanding of the disease. Wong explains that “the original paradigm… is that someone with TB is highly symptomatic – they have fevers, they’ve been coughing for weeks, they’ve lost a lot of weight, they’re sick – they’ve now come to seek care and they’ve been diagnosed with TB”.
The result? When CAD is trained on banks of chest x-rays, the images that are labelled “TB” will be from people who were highly symptomatic, says Wong. As such, the CAD software only learns to associate the disease with more extreme cases.
This can have practical consequences. When Wong and her colleagues used CAD for a screening programme in KwaZulu-Natal, they began by sending anyone with a CAD abnormality score above 60 for a test. This was in line with what had worked elsewhere. Yet as the research went on, it became clear that lots of patients with subclinical TB were being missed because many had scores below that level. This forced them to use a much lower threshold to detect these cases, which as noted comes with tradeoffs as many healthy people then get sent for testing.
It is likely that as AI is rolled out in other healthcare contexts in South Africa, more of these knotty nuances will emerge.
New regulations needed?
A related issue is that the data that CAD is trained on is often proprietary. Companies that make the software aren’t obliged to share information about where their data comes from or how their algorithms change when new versions come out. In response, Wong and her colleagues released a statement last year which called for “regulation to require CAD-developing companies to communicate changes between software versions”.
And this isn’t the only area where regulation appears to be lagging.
In South Africa, medical devices are technically regulated by the South African Health Products Regulatory Authority (SAHPRA). But a 2022 journal article argued that existing legislation is outdated for reviewing AI-based technologies.
One issue in particular, the paper argues, is that the safety and efficacy of a device is supposed to be reviewed according to “predefined static specifications and standards”. For instance, a defibrillator might be assessed on how well it performs a specific function (to restore a person’s heartbeat), and reviewers know that a given model would work in the same way over time. However, the function of an AI-based chatbot is broader – it provides answers to different kinds of questions depending on what it’s asked – and its responses may change over time as it is fed more data. Assessing the technology thus becomes much more difficult.
Asked about this problem, SAHPRA’s communications officer, Nthabi Moloi, told Spotlight that the body “has not commenced with the registration of medical devices” so this is presumably not yet a problem (though SAHPRA does sometimes use backdoor routes to review devices).
The way forward
For all these issues, researchers say that these are exciting times, and it appears that both international funders and the South African government are taking significant strides to use AI to address some of the country’s most devastating diseases. Speaking about the recent workshop on the use of CAD tools for silicosis and TB screening, Zhi Zhen Qin, a digital health specialist at the United Nations Office for Project Service, says she was “impressed by the vision” of the conference organisers. Qin says that by aiming to use AI to screen for both TB and silicosis at the same time – instead of viewing them as separate problems – “the South African government has been filling a much needed leadership gap”.
In the coming years, the role of AI in the South African health system might of course stretch far beyond TB and silicosis. Judging by our messy health data landscape, we are not ready for the transition. While the Western Cape Department of Health and Wellness and the National Health Laboratory Service have done impressive work, our health data systems on the whole remain patchy and fragmented. This will make it more difficult to train and deploy locally appropriate AI solutions. Whether new digital infrastructure being developed under the banner of National Health Insurance will solve the problem is an open question.
As with electronic data systems, building and deploying AI capacity in the South African public health system will not be easy. Patient data will have to be kept secure, systems will have to be inter-operable, and rather than outsourcing everything to software vendors, the state will need to build at least some internal technical capacity – at the very least government needs people with the technical expertise to know whether we are purchasing the right products. In a country where hundreds of thousands of Rands have been wasted on very simple websites, this can unfortunately not be taken for granted.