The price of a health system is no longer money

Botswana is a country of 2.4 million people with a health system that works, by the measure of a country that has spent its own money on it for sixty years. In the last twelve months that system nearly broke. A public health emergency was declared, a medical airlift arrived from Abu Dhabi, and a pharmaceutical memorandum was signed with India. Then two foreign parties arrived with money, and both stayed.

What they brought, and what they asked for in return, is the subject of this article. The money is the part that was announced. The data is the part that was not. And the data, not the money, is what outlives the contract.

The figure above is the argument of this article in one drawing. Medicines are renewed every year, from the first shipment. A records platform runs for the ten year term. A genomic dataset is different: its value is built once, it is unique to a population, and it outlives the contract that created it. Whatever the contract says about the first two, it is the third that decides whether the country still owns its health system when the contract ends.

On 26 August 2026, M42 announced a ten year partnership with the Government of Botswana to strengthen the country's health sector. The announcement named digital health transformation and a national genomics programme as the core scope, with supply of pharmaceuticals and essential medical products alongside. Three days earlier, the Ministry of Health had issued its own statement. It named a different counterparty, Abu Dhabi Health Data Services, and put pharmaceuticals first, with genomics appearing as an avenue for cooperation rather than a headline.

The two statements described the same arrangement differently. That is not unusual in itself. What matters is what neither party published. The financial terms were not disclosed. The data terms were not published. Neither statement said where the primary copy of the records would live, who would administer it, what the data could be used for, or what would happen at the end of ten years.

This was not the only foreign party in Botswana's health system. Since December 2025, a United States health memorandum had been running alongside it, funding the exchange of records, laboratory systems and a medical warehouse. That memorandum was not an international agreement, and it carried no legal obligations. But it had a data sharing arrangement attached, and that arrangement ran until reporting was complete, which is to say, for as long as the data existed.

To understand why these two agreements landed in August 2026, it helps to see the year that preceded them. What follows is a timeline of public events, drawn from news releases, parliamentary records and official statements. The two navy markers are the ones that reach health data.

The first navy marker is the United States memorandum of December 2025. The second is the Abu Dhabi agreement of August 2026. Between them, a public health emergency, a medical airlift, a World Bank loan, a pharmaceutical memorandum, a national budget that cut the health allocation, and a minister who reported medicine shortages in June, eight months after the emergency was declared. The country was still short of medicine when the ten year partnership was signed.

What the timeline does not show is the structural condition that made these agreements possible. Botswana funds its health system from its own revenue, and it has done so for decades. It is not a country that depends on external donors to keep its clinics open, in the way that Malawi or Zambia do. But in the twelve months shown above, it could not. That is the gap the two foreign parties entered, and it is the reason the terms of entry matter.

The United States memorandum was published, eventually. The text was released under freedom of information in August 2026, eight months after it was signed. It is a short document. It funds the exchange of health records, the laboratory system and a medical warehouse. It provides for audit access under Botswana's Data Protection Act. It is explicitly not an international agreement, and it creates no legal rights or obligations for either party.

What it does not contain is any description of the data sharing arrangement attached to it. That arrangement is referenced, but its text was not released. What is known about it comes from a leaked model specimen sharing agreement and from reporting on similar arrangements with other countries. In that template, data sharing runs for five years, but the text says it continues until reporting is complete. In one reported version, the term extends to twenty five years.

The matrix above places Botswana's own law, the published memorandum, and the two undisclosed arrangements side by side. Botswana's Data Protection Act, in force since January 2025, is a serious statute. It classifies clinical records as sensitive data, restricts onward transfer, and requires that a copy remains in Botswana. It applies to genomic data. It requires an impact assessment. It says that when a processing relationship ends, the processor must return or delete the data.

The law exists. The question is whether the contracts are written to use it.

The chart above is the central comparison of this article. The money lasts between three and ten years. The data claim, in the one template that has been seen, lasts until reporting is complete, which the text defines as up to twenty five years. The Abu Dhabi data term is not published, and is shown without a bar. A contract that runs for ten years but governs a dataset that outlives it is not a ten year contract. It is a permanent arrangement with a ten year service layer on top.

The reason this matters more for Botswana than it would for a country that has always relied on external funding is that Botswana is not one. The chart below shows external funding as a share of current health expenditure across nine African countries in 2023. Botswana is the navy bar, at 6.5 per cent. Malawi, at the top, is 65.5 per cent.

Botswana's health system is not donor funded. It is tax funded, and it has been for sixty years. That is what makes the entry of foreign parties into its data layer a different proposition from the humanitarian programmes that operate in countries at the top of this chart. A country that pays for its own system has something to lose that a donor funded system does not: ownership of the records that system generates.

When a Gulf health platform signs a ten year partnership with an African health ministry, what is being exchanged? The announcement says medicines, digital health and genomics. But the thing that has lasting value, the thing that cannot be reissued or reordered, is the data. Medicines are consumed. Platforms are replaced. A genomic dataset, once sequenced, is unique to that population for as long as that population exists.

This is not a speculative concern. Gulf health platforms have been acquiring and operating health assets abroad for several years, and the map below shows where.

The map is not a description of either Botswana agreement. It is the context in which those agreements sit. Gulf platforms now operate hospitals in Iraq, Libya and Algeria, hold acquisitions in the United Kingdom and Greece, and have signed memorandums as far as Uzbekistan. The Botswana partnership is the first that puts a national genomics programme in the core scope.

The question is not whether Gulf platforms should operate abroad. They bring capital, operational capability and, in several cases, genuine clinical quality. The question is what the sovereign buyer writes into the contract that governs the relationship, and whether those terms are published.

None of this is new. Three previous waves of foreign involvement in national health systems tell the same story, and the lesson in each case is the one Botswana now faces.

Lesotho shows what happens when a cost ceiling moves with the health budget but the contract is fixed: the cost outgrew the budget, and the contract was terminated early. Türkiye shows what happens when the obligations are set in a foreign currency against revenue in a local one: the payments are fixed, the currency moves, and the bill grows. Iceland shows what happens when the consent and legal basis for a national health database are too thin: the Supreme Court ruled that a daughter could keep her father's records out, the database was never built, and the exclusive licence lapsed.

Botswana does not need to write a new law. It has one. The Data Protection Act 18 of 2024 came into force in January 2025, eleven months before the United States memorandum was signed. It classifies health data as sensitive. It restricts transfer. It requires that a copy remains in the country. It demands an impact assessment for genomic data. It sets out what must happen when a processing relationship ends.

What Botswana needs is not a new statute. It needs contracts that use the statute it has. The law is the floor. The contract is where the terms that the law does not reach, the commercial terms, the intellectual property terms, the benefit sharing terms, are written or left blank.

The Nordic countries have spent forty years building something Botswana is now building in ten: a national health data infrastructure, governed by law, operated under permit, and audited in public. They have made mistakes, and they have corrected them. Their experience is not a template that can be copied, but it is a set of practices that can be learned from.

The flow above is how a Nordic registry decides who gets the data. An identity number links one person across every registry. The registries hold prescriptions, diagnoses and outcomes. A data permit is required, from the registry holder, every time. Ethics review applies where required. Analysis happens in an approved environment, on approved servers, with extracts leaving only under a permit. And the audit trail is published, yearly, so that the public can see who used what and for what purpose.

This is not a description of a single Nordic system. It is a synthesis of practices from Findata in Finland, the Danish Health Data Authority, Swedish register holders, and the Norwegian REK system. They differ in detail. They share a principle: the country that generates the data administers the data, and every use is approved, recorded and reported.

None of this is beyond Botswana's capacity. It has the law. It has the institutions. What it needs, and what a Nordic supplier can provide inside a Gulf platform's delivery contract, is the design of the controls: the permit flow, the approval environment, the audit trail, and the benefit sharing terms that make sure what the data produces comes back to the health system that generated it.

What follows is not a policy recommendation. It is a set of questions a sovereign buyer can ask, and that an operator can answer, before the first genome is sequenced. None of them requires the commercial terms to be published. All of them can be.

This article reaches a conclusion: that the price of a health system is no longer money, and that a sovereign buyer that does not write its data terms into the contract will find, at the end of the contract, that it has sold something it did not intend to sell. That conclusion could be wrong. Three tests, set for September 2027, would weaken it.

If a second country signs a contract that combines service delivery with a national data responsibility, and publishes its terms, the near term concern weakens. If Botswana publishes a delivery milestone, the execution concern weakens. If either the American data sharing arrangement or the Abu Dhabi data terms become obtainable by any route, the governance concern weakens. If none of these happens, the concern of this article stands, and it strengthens with every year that passes without publication.

The title of this article is a statement, not a warning. The price of a health system is no longer money. It is the terms under which the data generated by that system is held, used, and returned. A sovereign buyer that writes those terms, and publishes them, keeps its health system. One that does not will discover, at the end of the contract, that the most valuable thing it ever produced was the one it never priced.