Picture a woman in her 50s living in Makueni county with diabetes and high blood pressure.
Her sugar is checked at a dispensary near home, her kidneys tested at a private laboratory in town and her medicines bought from whichever chemist has them.
In March a doctor at the county referral hospital changes her prescription. In June a clinician at the dispensary, working from an older record or from none at all, renews the old one.
Each acted reasonably on what was in front of them. Yet she is now at risk, and the only complete file on her illness is a plastic bag of receipts and folded laboratory slips that she carries herself.
This is the kind of harm that World Patient Safety Day, marked on 17 September under the theme 'Safe care for noncommunicable diseases', asked us to confront.
The World Health Organisation estimates that around one in 10 patients is harmed while receiving care, and that about half of that harm is preventable.
We usually picture such harm inside the consulting room, in the wrong drug, the missed diagnosis or the infection caught on a ward. Those failures are real, but for people living with chronic illness, the more dangerous place is often the gap between one clinical decision and the next.
And chronic disease widens that gap, because it outlasts the attention of any single institution. In Kenya, the WHO country office's 2025 annual report puts their share at 41 per cent of annual deaths and half of all hospital admissions.
They are managed over decades, across public and private facilities, paid for partly by insurance and partly from the patient's pocket. The patient lives them as one journey.
The system records them as a string of unrelated transactions, and almost nobody is paid, measured or held to account for what happens in between.
Kenya, to its credit, has written the remedy into law. The Digital Health Act of 2023 commits the state to "a system of shareable and portable personal health records", promises health data portability and requires digital health platforms to be interoperable with the national system. The Ministry of Health is building a Health Information Exchange to carry it.
Yet digitisation is not the same thing as safety. A hospital can abolish paper and remain an island. A patient can hold three electronic records that never speak to one another. Poorly connected technology does not cure fragmentation. It simply digitises it.
Interoperability, the ability of one system to read and use another's data, should therefore be regulated as patient safety infrastructure as seriously as a reliable oxygen supply.
Still, two objections deserve a straight answer. The first is that Kenyan patients are harmed by more basic failures, too few nurses, empty pharmacy shelves and facilities waiting to be paid, and that talk of data exchange is a luxury. I disagree only in part.
A record that follows the patient does not replace a nurse, but it makes every scarce nurse safer, because an overstretched clinician cannot rebuild a medical history from memory and a bag of receipts.
The second objection is that I lead a company that builds health technology and has a stake in this argument. That is true, which is why the test I propose should be applied to firms like mine first.
My test is simple. Stop asking whether a platform is faster or processes more claims, and ask whether it reduces harm. Does a laboratory result reach the clinician who ordered it? Does a changed prescription appear at the next facility she visits? Is an unfinished referral flagged before the patient disappears from view?
Artificial intelligence can help by spotting the missed refill or the result nobody opened, but only if it sharpens a clinician's attention rather than burying it in alerts. Systems will fail. What matters is whether failure becomes visible and recoverable before it becomes harm.
But the same honesty is needed about patient engagement. WHO rightly insists that people with chronic illness be partners in safe care. Handing a patient an app is not empowerment, however, if it quietly transfers the system's job to her.
A 2025 study of the cancer patient navigation programme at Kenyatta National Hospital found that patients with more exposure to navigators understood their diagnosis and treatment better and were less likely to delay starting or continuing treatment.
Navigation is continuity delivered by human beings. Good technology should carry it to places where no navigator will ever be posted.
The next steps are concrete. Parliament should ask the Ministry of Health when the interoperability provisions of the Digital Health Act will be enforced, and against what timetable.
The Social Health Authority and private insurers should make data sharing a condition of contracting and measure whether referrals are completed, not merely whether claims are settled.
The Digital Health Agency should refuse to certify systems that cannot exchange records. Hospital owners should stop treating patient records as proprietary assets.
A patient experiences one illness, even when many institutions treat it. Our task is to make sure the gaps between those institutions stop becoming gaps in her safety.
Until they do, the most complete medical record in Kenya will remain a plastic bag of receipts, carried by the sickest person in the room.
Group managing director of Smart Applications International Limited
