The healthcare industry is constantly evolving, and the integration of artificial intelligence (AI) into clinical and operational workflows has the potential to bring significant improvements in care delivery, especially for frontline health workers such as community health workers (CHWs). Innovations in Healthcare recently conducted a study that focused on how three organizations - Audere, Simprints, and ThinkMD - are using AI to support CHWs in low- and middle-income countries (LMICs). The study findings were recently shared in a webinar hosted by Innovations in Healthcare. During the webinar, leaders from these social enterprises and their implementing partners, delved into this growing field, unraveling insightful perspectives and illuminating the potential of AI in revolutionizing healthcare accessibility and equity.
The panelists shared the insights below as part of their work leveraging AI for frontline health workers:
- Unlocking AI's Potential:Revolutionizing Frontline Care. AI is emerging as a powerful ally for CHWs, offering multifaceted support ranging from refining clinical assessments, to facilitating accurate patient identification, and even aiding in disease diagnosis, notably malaria. Yet, as Eliab Mwiseneza, who works closely with community health workers in Rwanda, notes, "when you are implementing this AI technology, the first thing that you have to consider is the context.” In order to successfully introduceAI tools, you must understand the needs of the communities, the challenges they face, and the contribution of the technology. Examples shared during the webinar include machine learning platforms bolstering clinical logic in Nigeria (ThinkMD), facial biometrics streamlining patient identification in Uganda (Simprints), and computer vision facilitating the interpretation of rapid malaria diagnostic tests in Rwanda (Audere). These applications underscore AI's versatility in augmenting CHWs' capabilities to deliver quality care effectively and efficiently.
- Navigating the AI Landscape:From Hype to Reality. Amidst the current passion surrounding AI, the webinar shed light on a crucial distinction – while the current AI discourse primarily revolves around generative AI and deep learning through the use of unsupervised algorithms, practical implementations with CHWs predominantly involve supervised machine learning. Elina Urli Hodges further emphasized the importance of human intervention noting, "humans play a really important role in making sure that whatever the computers are learning, is accurate [or] as accurate as possible.” Human-in-the-loop approaches are critical in patient care as well as playing a pivotnatal role in training, monitoring, tuning, and measuring the accuracy of algorithms before field deployment. According to WHO, "[In the] adoption of AI …autonomy requires that any extension of machine autonomy not undermine human autonomy. In the context of health care, this means that humans should remain in full control of health-care systems and medical decisions… they should be designed to assist humans, whether they be medical providers or patients, in making informed decisions.”
- Empowering CHWs: Recommendations for Ethical AI Integration. Central to the discussion were key recommendations aimed at empowering CHWs and ensuring the responsible integration of AI into healthcare delivery. Key recommendations include:
- Providing comprehensive training for CHWs;
- Fostering a deep understanding of the local ecosystem among AI developers through strategic partnerships;
- Designing contextualized solutions tailored to specific needs; and,
- Upholding stringent data governance practices guided by regulations such as GDPR, and responsible AI frameworks included in the Global Index on Responsible AI.
- Towards Responsible, Sustainable Innovation: Navigating Emerging Technologies. The panelists offered thoughtful perspectives on strategies for implementing emerging technologies like AI in LMIC health systems to promote care access and equity. According to Dr. Barry Finette from THINKMD, “when you look at machine learning and AI approaches, the one thing that you need to have...are accurate, clean data sets. Without that, AI and machine learning is really has no value.” Sarah Grieves from Simprints stressed the importance of"not starting with a technology and finding somewhere to use it, but starting with the problem you're trying to solve." Additionally, panelists emphasized the critical importance of long-term sustainability, underscored by robust government support and alignment with practical healthcare needs.“Local partnerships, local capacity strengthening and having an awareness of local challenges...and incorporating that understanding into the design and implementation as you move forward with any sort of AI solution with frontline health workforce [is imperative to sustainability],” according to Natalie Maricich from Audere.
The integration of AI into healthcare for frontline workers in low- and middle-income countries marks a pivotal moment in the evolution of medical care. By understanding the nuances, embracing responsible practices, and fostering collaborative efforts, we can harness the true potential of AI to revolutionize healthcare accessibility and equity on a global scale. Listen to the webinar here.
*Panelists:
- Natalie Maricich, Product Manager, Audere
-
Eliab Mwiseneza, Child Health and Malaria Technical Officer, Jhpiego Rwanda
- Sarah Grieves, Strategic Partnerships Lead, Simprints
- Dr. Barry Finette, President and Co-Founder, THINKMD
- Moderated by Elina Urli Hodges, Assistant Director, Programs at Innovations in Healthcare and the Duke Global Health Innovation Center
This research and event was supported by the Bayer Foundation.