AI in healthcare: from fighting epidemics to family doctors

Since May, teams in the Democratic Republic of Congo have been battling the Ebola virus. Health workers visit homes, assess cases and refer patients to hospitals. Their records, detailing entire communities, have been converted into digital data by Anthropic, an AI company. Using Anthropic's Claude model, researchers evaluated answers and reports, aiming to understand who was infected and their contacts and whereabouts before diagnosis.

This AI-assisted analysis now generates situation reports within an hour, compared to previous full-day efforts.

Dr Marie Roseline Darnycka Belizaire, Director of Emergency Preparedness and Response at WHO's African office, highlighted how AI can rapidly assess epidemic spread and prioritize medical aid. AI identifies potential outbreak hotspots and aids resource allocation, crucial for timely emergency response.

AI's role extends beyond epidemics. The technology could enhance diagnoses, alleviate administrative burdens on medical staff, and strategically deploy limited resources, according to experts. However, concerns remain about AI reliability, decision-making transparency, and patient benefit.

With growing global demand for healthcare and shortages of staff and resources, AI offers a potential solution. Dr Thomas Zapata, Director of the WHO Academy, sees AI as a tool to bridge the gap between healthcare professionals and rising service demand. However, he cautions against viewing AI as an end in itself, but rather as an enabler within existing health systems.

AI can reshape care organization and assist in training health workers to utilize new capabilities. At the World Health Summit, Dr Cecilia Mbaka, Senior Technical Adviser for Health and Emergencies at CARE, emphasized three key questions: AI's ability to support clinical decision-making, its capacity to handle repetitive tasks, and its potential to tailor care to individual patient needs. If AI excels in these areas, medical professionals can focus more on patient care, fostering trust within communities.

In the DRC, around 50% of deaths occur outside medical facilities, underscoring the need for trust in local health systems. While AI can aid in care organization, it cannot alone ensure patient trust. Transparency issues persist in AI diagnostics, as the decision-making process remains opaque. Nonetheless, AI shows promise in improving healthcare delivery and patient outcomes.