Skip to content

Examining 105,000 cases decides the viability of intelligent diagnosis: higher accuracy and reduced doctors' burdens

Share
Examining 105,000 cases decides the viability of intelligent diagnosis: higher accuracy and reduced doctors' burdens

Listen to this article

Read by Anchor

Discussions about automation in the health sector have moved beyond theoretical promises when they collide with the realities of daily clinics and the pressure of examination rooms. The question is no longer about whether algorithms can read medical images independently of a doctor, but about their actual ability to reduce examination time and lower errors without disrupting the system with misleading alerts. This shift from software showcase to measuring clinical impact became the focus of the third conference of the Faculty of Clinical Sciences at the University of Medical Sciences in Ondo, Nigeria, where academics and experts gathered to rethink medical practice in the context of digital care.

The major clinical numbers settled a key part of the debate over the usefulness of algorithmic intervention in diagnostic rooms.In the keynote presented by Foulami Adeboli, CEO of AI Care, the discussion was based on a large Swedish study that included 105,934 women in breast cancer early-detection programs. The results of that experiment showed a 29 percent increase in case detection thanks to AI-supported screening, accompanied by a 44 percent reduction in the imaging workload for radiologists, and most importantly, this reduction was achieved without any rise in false-positive results. The evidence did not stop at advanced countries; another field study strengthened it by analysing 39,849 patient visits across 15 primary-care clinics in Nairobi, Kenya, documenting a tangible drop in errors related to recording medical history, diagnosis, and treatment protocol determination.

The operational value of these experiments stems from confronting the crisis of medical staff depletion and physician migration, a dilemma shared by many emerging health systems. In this context, technology solutions designed for low-resource settings emerged; among them is the Nigerian “Openwa” technology that uses newborn crying fingerprint analysis to detect birth-asphyxia, and the “Aspire” technology that combines a portable mesh camera with an Android app to screen for complications such as cerebral malaria, providing a comprehensive clinical assessment in about a minute even in centres with intermittent internet connectivity.

This applied pathway was accompanied by explicit warnings from university leadership; Vice-President Ibonoluwa Adijogbebi, through the Vice-Chancellor for Administrative and Clinical Affairs, confirmed that digital technologies are designed to complement clinical judgment, empathy, and human communication, not to undermine them or replace them. The discussions, with the participation of the dean of the faculty, O. H. Adijumo, and the organising committee chair, Foulami Ojuonedi, emphasised that expanding the adoption of these tools requires strict standards to protect patient data privacy, enhance cyber security, and hold algorithmic decisions accountable, while guiding medical students to use AI as a training partner for critical thinking and case simulation rather than a passive reliance.

What we see in this documentation redirects the compass of investment and regulatory decisions in Gulf hospitals, Egypt, and the Levant.In Gulf systems that see large capital flows toward hospital automation and health records, the real gain lies in reducing the routine audit burden on radiologists without imposing on insurance budgets the costs of confirmatory tests resulting from false alerts; this means that the purchase criterion must shift from the allure of interface displays to statistically proven clinical efficiency. In Egypt and the Levant, where equipment costs and staff shortages pressure primary-care centres, the lesson lies in adopting lightweight terminal software that runs on mobile devices with minimal network connectivity, enabling the general practitioner in peripheral areas to make a precise and rapid decision that preserves patient life and limits unnecessary referrals.

The bet in the next phase does not rely on amplifying model capabilities, but on starting with small, measurable applications within medical departments, supported by institutional guidelines that ensure outputs are compared against accepted clinical standards, so that the final judgement remains in the hands of the responsible physician.

Don't miss the next story

Subscribe for updates