Digital Innovation

Areas of exploration

The KTU seeks to remain at the forefront of health systems strengthening.

We are committed to ensuring that the constant global developments in digital health and new technology offerings are adapted, assessed and deployed in a responsible, transparent and accessible manner.

Digital evidence repositories for our current PACK content

Transparent tracking of clinical updates and harmonisation of evidence

Mobile App development and related health technology assessments

Integration of clinical resources into more accessible clinical support for frontline workers

Integration of AI-leveraged approaches

Efficient evidence synthesis and interrogation techniques

What we offer

Clinical editorial and training team upskilled in AI

Integrate AI into research, content and innovative thinking

Unique skill set, bridging health and software development, with necessary in-house expertise

Build, deploy and implement clinical decision support applications

Research expertise

Focus on the validation of solutions through pragmatic research methodologies

Current Project

The Knowledge Translation Unit (KTU) is advancing the development of the AI-Supported Healthcare Assistant (AISHA) prototype. AISHA is designed to summarise clinical recommendations from diverse evidence sources, serving as a public health tool for clinicians, policymakers, and health educators.

The project will integrate a Large Language Model to monitor and analyze updates in healthcare guidelines, identifying inconsistencies across global, regional, and local levels. Additionally, the database of guidelines will be expanded to comprehensively address prevalent health conditions, beginning with those commonly managed in primary care settings in South Africa.

To ensure continuous improvement, KTU will establish programs that bring together local stakeholders with AI and healthcare experts from the global community. By enhancing the synthesis of healthcare evidence, AISHA aims to streamline and improve public health policy decision-making.

This project is supported by the South African Medical Research Council with funds received from National Department of Health the Bill and Melinda Gates Foundation.

Past Project

An exploratory project to assess if Large Language Models (LLM) will be able to:

Analyse text-based medical evidence in various formats,
Extract relevant clinical recommendations, and
Formulate clinical decision support algorithms (that have been validated by clinicians) in real-time for frontline workers.

This would improve speed, accuracy and inclusivity of decision-making. The project also includes validation research to assess the safety and robustness of LLM-generated outputs.