President William Ruto said something at the Kenya Health Summit last week that I completely agreed with.
The measure of reform, he said, is not the laws passed or the numbers registered, but the patient's experience. Elsewhere he put it more sharply: without data you operate on guesswork and are sure to fail.
I want to take him at his word. Not as a rhetorical trap, but because it is the right test, and applying it honestly tells us something the summit did not.
Let me start by conceding what is real, because a good deal of it is.
The Social Health Authority has registered 32.3 million Kenyans. Its chief executive says NHIF peaked at 16.9 million. Whatever one thinks of how the drive was run, that is a genuine change in the size of the pool. Some 7,414 UHC workers who spent years on contract moved to permanent and pensionable terms on July 1, backed by Sh8.94 billion. That corrects a real injustice.
Kemsa's order fill rate, by the Ministry's and Kemsa's own account, has climbed from roughly 40 per cent to above 90. Thirty-nine CT scanners, 67 theatres and 157 laboratories have gone into counties that did not have them. None of this should be waved away.
But every one of those figures measures what went in. Money collected, machines bought, people hired, cards issued. The summit also offered numbers that sounded like results: 1.2 million mothers supported to deliver safely, 500,000 surgeries covered, eight million people treated.
Listen carefully and these are volumes too. A surgery covered is a transaction. It does not tell you whether the patient walked out, or walked out sooner than four years ago, or walked out at all.
To be fair, the Cabinet Secretary did offer two outcome claims: full immunisation coverage up from 78 to 97 per cent, and new HIV infections down 37 per cent since 2022. Immunisation coverage is still a measure of service delivered rather than disease averted, though a welcome one.
The HIV figure is harder to square with the government's own data. The National Syndemic Diseases Control Council reported 19,991 new infections in 2024, up 19 per cent on 2023, while UNAIDS puts 2025 at under 14,000. Both cannot be right, and the HIV response has in any case been driven for two decades by donor programmes that predate these reforms. Two contested indicators out of a two-day summit is a thin outcomes case.
Here is the question the President's own test demands. Are Kenyans healthier than they were in 2022?
Nobody at KICC could answer that, and I do not say so to score a point. The instruments that would answer it have not been run. The 2022 Demographic and Health Survey, remarkably, omitted the maternal mortality module altogether.
The last dedicated household survey of what families pay out of pocket was 2018. The World Health Organization classifies Kenya's death registration data as unusable for policy evaluation. Four years into a reform built on the promise of digitisation, we still cannot say with confidence how many Kenyans died last year, or of what.
What we do know predates the reforms and is sobering. The most recent population estimate of maternal mortality, from the 2019 census, is 355 per 100,000 live births, barely moved from 362 in the 2014 survey. Neonatal mortality has drifted from 28 to 21 per 1,000 over roughly three decades, and was 22 in 2014 and 21 in 2022.
A Ministry of Health co-authored study published last year found that maternal deaths recorded in facilities fell steadily until 2018 and then stopped falling. UNFPA estimates that over 80 per cent of Kenya's maternal deaths are now attributable to poor quality of care rather than to women failing to reach a facility. Skilled attendance at birth has risen from 62 to about 70 per cent. Women are arriving. They are dying after they arrive.
That is a problem no enrolment figure addresses. You cannot register your way out of a postpartum haemorrhage recognised 20 minutes late.
As a surgeon, I would put it this way. If I presented my annual report to a hospital board and listed theatre hours, implants consumed and staff recruited, but could not state my complication rate or 30-day mortality, I would not be reporting results. I would be reporting effort. Effort matters. It is not the same thing.
The financing story belongs here, not as a separate indictment but because it is the mechanism by which effort fails to become outcome. The Economic Survey 2026 records that the contributory fund collected Sh57.7 billion in its first financial year against claims of Sh91.5 billion, of which Sh33.4 billion was still unpaid at year end.
The chair of the National Assembly Health committee, Dr James Nyikal, has said only about five million of the registered millions are actually contributing. The Rural and Urban Private Hospitals Association, whose survey figures are self-reported and should be read as such, found in 2025 that claims settlement averaged 34 per cent and that more than half of surgical claims were unprocessed after three months.
A facility carrying six months of unpaid surgical claims does not stop operating. It quietly starts choosing which operations it can afford to do. That is rationing, and rationing is an outcome nobody is counting.
One figure at the summit would, if it holds, be the best news of the week. The ministry said the share of households protected from catastrophic health spending has risen from six to 24 per cent. I would like to see the survey behind it.
Published analyses of the 2013 and 2018 household surveys put catastrophic spending incidence between roughly seven and 11 per cent depending on the threshold. If the ministry has new data showing a fourfold improvement in protection, it should publish the methodology, because that number alone would settle more than everything else presented across two days.
The Deputy President, to his credit, said the next phase of reform should focus on measurable outcomes, and listed waiting times, referral completion, patient safety and health outcomes among them. So here is a constructive proposal.
The facility data on maternal deaths and stillbirths already exists in the health information system, and the ministry's own researchers have shown it is good enough to track county trends. The household budget survey that finished fieldwork last month will carry health and mortality modules.
Let the 2027 summit open with six numbers, each compared to 2022: institutional maternal mortality and stillbirths by county; time from cancer diagnosis to first treatment; 30-day surgical mortality; under-five mortality; out-of-pocket spending from that new survey; and the median number of days SHA takes to settle a claim, published monthly.
If those numbers have moved the right way, the government will have earned the applause it sought this week, and I will give it. If they have not, we will at least know where to look. Either way, we will have stopped guessing. The President said that is the point. He is right.
Surgeon, writer and advocate of healthcare reform and leadership in Africa. [email protected]
