Artificial Intelligence HIV Management Tool a First in Africa

BY: .

An HIV/AIDS management tool able to predict whether a patient is resistant to ARVs has been formulated by UKZN PhD graduate, Dr Yashik Singh.

Singh’s mathematical model is being developed to alter drugs to change a patient’s CD4 count, to predict CD4 counts below 200 as well as the onset of AIDS and to monitor a patient’s use of ARVs. The study is believed to be the first of its kind on the continent.

‘The PhD was a pilot project, so to speak. International experts have analysed the framework so that they could further develop, build and implement it,’ said Singh. ‘The idea was to build it to apply to almost anything, so it has to be made as generic as possible. Governments ask a lot of questions about how to authenticate telemedicine, which is why this process of testing takes so long to implement. We are currently trying to see whether we can develop accurate algorithms to predict drug resistance based on information we can use from patients every day.’

The aim of Singh’s study was to ‘develop a physician-administered, artificial intelligence based decision support tool that aids the management of patients on antiretroviral therapy. This would be facilitated through the creation of machine learning algorithms (MLA) to predict current and future CD4 cell count using genomic sequences, current CD4 cell count using standard of care data, report a single interpretation to a HIV resistance profile; and the development of a web-based portal design that integrates these tools together.’

Singh’s idea will allow tests to be done on a patient’s resistance to HIV based on information and standard of care data, which doctors acquire from their patients. Masters students are currently building the framework developed by Singh and adapting and applying the framework model so that it can be applied to any domain before taking it to hospitals to be tested on patients.

‘In South Africa, we need to take into account the merging of telemedicine and informatics. We needed to develop a more holistic system that takes into account the environment, whether in a rural area or a city, along with the doctors, patients and medication so that integration can be maximised to provide more data needed to help patients.’

Currently, Singh is trying to implement and further develop the model produced from his PhD. ‘The model aims to improve fertility and prevent resources from being wasted. An education campaign is also currently underway to increase awareness to doctors and patients of drug resistance tools which are available in South Africa.

A survey was conducted amongst final year students at the UKZN’s Nelson R Mandela School of Medicine and only 4% of students knew about drug resistance tools being available. Doctors need to be trained to know of these tools as well.’

Singh’s PhD, which he completed last year, was titled: “A Physician-Administered Artificial Intelligence-Based Decision Support System Tool that Facilitates the Management of Patients on Anti-Retroviral Therapy”.

Singh’s PhD was the first in medical informatics to be awarded by an African university and one of the very few internationally. ‘I guess the most important thing I learned from the PhD is gratitude. As one comes to the end of the PhD journey, one realises how little one actually knows, and you are enveloped with a feeling of immense gratitude.’

Singh’s main interests are in artificial intelligence applied to finding ways of creating mathematical models that explain the behaviour of data. ‘As such, most of my current work deals with mining data. I am working on various issues that range from predicting fertility, models that predict CD4 cell count and viral loads, predicting diagnosis of silicosis.

‘I am also working on needs assessment, design and implementation of electronic medical records in all its forms ranging from small practice computer based records systems to national electronic hospital records systems. I also am working on a few projects dealing with education and trying to improve access and knowledge to medical informatics tools.’

 Singh is a Lecturer and the Academic Co-ordinator of the Medical Informatics programmes in the School of Nursing and Public Health’s Telemedicine discipline. Singh supervises Masters and PhD candidates, as well as lectures many Medical Informatics modules. In his free time, he participates in local outreach programmes for his community temple.       

-         Zakia Jeewa