CSIR boosts rural health with AI

Artificial intelligence (AI) is being used to develop cutting-edge technologies to improve the country’s rural health systems. 

To address the issue of limited diagnostic resources in rural areas, the CSIR is developing a machine learning-powered diagnostics system, a technology that combines cutting-edge machine learning algorithms to autonomously help medical professionals diagnose diseases with better accuracy and speed. Machine learning is a branch of AI technologies that aims to mitigate the potential errors made by newly appointed medical professionals. 

It seeks to expedite the diagnosis of diseases, which is often delayed because traditional treatment approaches are reliant on human involvement. By delivering precise and swift disease diagnoses, machine learning has the potential to reduce the spread of infectious diseases.

Last Wednesday, 28 June, young researchers from the  CSIR showcased these innovations, aimed at improving South Africa’s healthcare system in remote regions.

“The technology can be used in busy medical centres that handle many patient samples each day,” said PhD candidate Nkgaphe Tsebesebe. “With this technology, the diagnostic process can be accelerated, reducing patients’ waiting time. It can diagnose thousands or even millions of samples in just a few seconds, which is particularly helpful in preventing the spread of viral and infectious diseases.”

During the media briefing, Sipho Chauke, another PhD candidate, demonstrated optical-based biosensor technology for the detection of Mycobacterium tuberculosis (TB). A miniaturised point-of-care device utilises light to detect TB bacteria in samples containing nucleic acid. Its primary objective is to assist healthcare systems in remote areas, particularly rural regions, by facilitating the diagnosis of TB and streamlining the initiation and administration of treatment for patients. The technology aims to significantly reduce the diagnostic time required for TB cases, make TB diagnostic affordable and offer large-scale diagnostics of various diseases.

The World Health Organisation has an End TB strategy, which aims to eradicate TB by 2025. CSIR-developed optical-based biosensor technology for the detection of Mycobacterium TB contributes to this strategy by offering access to medical technologies that can be utilised in healthcare systems to enable the diagnosis of TB at little hassle to ordinary South Africans. By making TB diagnosis available to all through this technology, the aim of the End TB strategy can be achieved through the early detection of TB, which will result in early treatment initiation, prevention or control of the spread of TB and a reduction in the number of multidrug-resistant TB cases.

“Although molecular tests are available for detecting and diagnosing TB, they take several weeks to give a diagnosis and are often expensive to run, said Chauke. In addition, there are no point-of-care tests commercially available locally to ease the burden of using molecular tests and the costs associated with running them. 

“Furthermore, this technology will assist ordinary South Africans by improving early clinical prognosis and treatment initiation for TB, thereby decreasing the rate of transmission and spread of TB between people, especially in remote settings (i.e., rural areas) within South Africa.”

Another device addresses major changes in the virus genome of SARS-CoV-2 and HIV-1, characterised by new variants of concern and accumulative mutations resulting in drug resistance. This has fuelled the need for fast and reliable prediction of emerging mutations in managing the disease. 

The CSIR-developed Localised Surface Plasmon Resonance system, uses optical biosensors to analyse biological elements such as nucleic acids, protein, antibodies and cells without interfering with the molecules in the solution. Its low complexity optics and ability to excite unpolarised light make it ideal for point-of-care device development. In a point-of-care setting, this system eliminates the need for timeous laboratory testing for diagnostic purposes.

PhD candidate Phumlani Mcoyi said: “With a growing interest in laser-based techniques for point-of-care diagnostics, mutation detection will guide the development of the point-of-care diagnostic system, which will be of particular interest to the most disadvantaged South African communities. The availability of a simple, fast and reliable laser-driven diagnostic technique will reduce the time and costs involved in mutation detection in the health sector.” 

Machine learning-powered diagnostics systems and the optical-based biosensor technology for the detection of Mycobacterium TB utilise IoMT and AI to connect multiple machines, such as X-ray scanners, between different medical facilities and mobile clinics. This connectivity allows patients to be scanned and scanned images to be transmitted to a centralised database. These technologies use AI algorithms to perform diagnoses and send the results back to the facility or directly to the patient, using their preferred method of communication.

For more information, see https://youtu.be/vKvc1oVxAno