An artificial intelligence tool has reduced diagnostic errors by 16 per cent
in Kenyan clinics, according to a large-scale study involving 40,000 patient
visits.
The tool, known as “AI Consult,” was developed
by Easy Clinic and embedded into the Easy Clinic EMR platform. It offers
real-time clinical decision support and was tested across 15 primary care
clinics in partnership with Penda Health.
The study, published in Nature Medicine and highlighted by the OpenAI blog, also
reported a 13% reduction in treatment errors among clinicians using the tool,
compared to a control group that did not.
Clinicians using AI Consult also showed a 32 per
cent decrease in omissions during history-taking and triggered 10–15 per cent
fewer “red alerts” for unsafe treatment plans over time, indicating a training
effect even outside direct AI feedback.
The AI tool uses a traffic-light alert system
to flag issues based on severity, preserving clinician autonomy while offering
actionable insights at diagnosis and treatment points.
Girish Mohata, CEO of Easy Clinic, said the
findings were “extremely powerful,” citing 75% adoption by clinicians and
measurable improvements without added burden.
The tool’s effectiveness was assessed by
independent, blinded physicians who reviewed anonymized patient records.
The study received ethical approval from the AMREF Health Africa Ethical and
Scientific Review Committee and Kenya’s Ministry of Health.
“This is proof that high-quality, AI-driven
healthcare is achievable in any setting,” said Dr. Sarah Kiptinness, Medical
Director at Penda Health.
Experts say the study helps bridge the
long-standing gap between AI’s theoretical potential and its real-world
clinical impact, especially in resource-limited environments like Kenya.
Researchers emphasised that successful implementation depends not only on advanced models but also on deep alignment with clinical workflows and sustained trust from medical professionals.
