Kenya’s healthcare system is at a crossroads. A bold promise to expand health coverage to millions of informal workers has instead produced a crisis that hits the poorest hardest. The country’s AI-driven health reforms are under fire — and for good reason. A joint investigation by Africa Uncensored, Lighthouse Reports, and The Guardian has exposed deep flaws in the system that were ignored even before it launched.
What Is Kenya’s Social Health Authority?
In October 2024, Kenya replaced its National Hospital Insurance Fund (NHIF) with the Social Health Authority (SHA). President William Ruto introduced the change as his flagship healthcare reform. The stated goal was universal health coverage — a system that would finally bring affordable care to the 83% of Kenyan workers in the informal economy.
The SHA also created the Social Health Insurance Fund (SHIF) as its financing arm. Together, they require all Kenyans to contribute monthly premiums, regardless of income. That includes people who earn less than one dollar a day and those who cannot work at all.
How the Algorithm Works
At the core of SHA is a proxy means-testing (PMT) algorithm. This machine learning model estimates each household’s income — not by checking payslips or tax records, but by asking questions about physical assets. Trained volunteers visit homes and gather data on roofing materials, toilet types, livestock ownership, radios, family size, and other observable details. The responses feed into an algorithm, which then calculates a monthly premium for that household.
This approach treats a household’s visible attributes as a stand-in for income. However, it is an imperfect method — and the consequences of getting it wrong fall squarely on those with the least.
How the Algorithm Overcharges the Poor
The investigation found a systematic pattern: the algorithm overestimates incomes for poorer households and underestimates incomes for wealthier ones. As a result, poor families receive inflated premium bills, while better-off households pay far less than they should.
This is not a minor calibration error. Some beneficiaries face monthly contributions equal to 10% to 20% of their total income. For a family already struggling to afford food, that kind of burden is simply unmanageable. Moreover, those who cannot pay risk being turned away from health facilities altogether or receiving steep hospital bills they cannot settle.
The system, therefore, achieves the opposite of its goal. Instead of reducing inequality in healthcare access, it deepens it.
Real Stories of Unaffordable Premiums
The investigation puts human faces on the data. One single mother in Kenya was assigned a monthly premium of 3,500 Kenyan shillings — a sum that far exceeded what her income could support.
Patrick Maina is another example. A former motorbike taxi driver, Maina broke his spine in a 2021 accident. He contributes Ksh. 500 (roughly $4) every month. Yet when he falls ill, SHA provides no assistance. He pays into a system that gives him nothing in return.
These are not isolated cases. SHA volunteers reported visiting households in Nairobi and watching families who were already struggling to eat receive premiums entirely beyond their reach.
Warnings Were Ignored Before Launch
Perhaps the most damaging detail from the investigation is this: the government knew about the problem before the system went live. A pre-deployment report by data consultancy IDinsight — obtained by journalists — warned that the SHA system was “inequitable, particularly for low-income households.” The government deployed it anyway.
This decision raises urgent accountability questions. When experts flag an algorithm as unfair before it is used on 20 million people, choosing to proceed is a policy choice — not a technical failure. Critics, including Dr. Brian Lishenga of Kenya’s Rural and Urban Private Hospitals Association, have called SHA a “great tool for helping the government run away from responsibility.” The algorithm provides convenient cover: blame the code, not the policymakers.
The Collapse of the SHA System
The SHA system is now struggling to survive. Of more than 20 million registered users, only about 5 million pay their premiums regularly. That low compliance rate has created a severe funding gap. Hospitals across Kenya are reporting large deficits as SHA reimbursements remain unpaid.
Hospitals Bear the Financial Burden
Healthcare facilities have been left holding the costs. As reimbursements stall, many hospitals cannot pay suppliers or staff. This operational strain directly affects patient care — and it is the poorest, least mobile patients who suffer most when local clinics lack supplies or capacity.
On March 20, 2026, Kenya’s National Assembly Departmental Committee on Health declared SHA unsustainable. Former Deputy President Rigathi Gachagua had already predicted in March that SHA would “collapse in another six months.” If the programme is dismantled, millions of Kenyans will face heavy out-of-pocket costs with no safety net.
Meanwhile, Kenya’s elected representatives face no such vulnerability. MPs receive annual medical cover worth Ksh. 10.65 million, fully funded by taxpayers, including maternity, dental, and optical benefits for their families.
Why This Problem Matters Globally
Kenya’s SHA crisis is a warning for any country that considers replacing human-centred welfare with algorithmic decision-making. The wider lesson — that machine-learning means-testing systems shift accountability from policy decisions to opaque code — applies far beyond Kenya.
Furthermore, Kenya’s national budget tells a troubling story. The country allocates roughly Ksh. 132 billion to health in the 2025/2026 budget. That represents less than 4% of total government spending. At the same time, nearly 67% of government revenue goes toward debt repayment. Universal health coverage cannot succeed under those fiscal constraints, no matter how sophisticated the algorithm.
The funding gap has pushed ordinary Kenyans to build their own safety nets. Mchanga, Kenya’s largest crowdfunding platform, has recorded over 134,000 fundraisers since its founding — 40% of them for medical expenses. In 2024 alone, up to 16,000 families used the platform to raise money for hospital bills. Thousands more depend on WhatsApp groups and informal community channels. This is not digital innovation. It is desperation filling a policy vacuum.
What Kenya Must Do Next
Kenya’s government must acknowledge that the SHA algorithm is broken — and act decisively. Three steps are essential. First, exemptions must be introduced for households earning below the poverty line. Mandatory contributions from people who cannot afford food defeat the purpose of universal health coverage. Second, the algorithm needs independent auditing and recalibration, with a transparent process that the public can scrutinize. Third, healthcare funding must increase meaningfully as a share of the national budget. Technology cannot substitute for adequate investment.
AI holds real promise in healthcare — for diagnostics, resource allocation, and fraud detection. However, that promise collapses when the technology is deployed to shift costs onto those least able to bear them. Kenya’s SHA experiment is not just a cautionary tale about artificial intelligence. It is a reminder that every algorithm reflects a policy choice — and policy choices have consequences.
