Buying the Record: Africa's EMR Budget and the Gap It Hides

There is a pattern in how African health ministries buy electronic medical records, and it is visible in the announcements. A country signs a national strategy, a donor or a development bank puts an envelope behind it, a vendor is selected, a photo is taken at the first facility, and a number is quoted for the country. The number is almost always the envelope. It is almost never the cost of a facility.

Those are not the same number, and the distance between them is where most programmes fail four years later. Rather than a purchase, a national EMR is the equivalent of a building handed over without a maintenance contract: the ribbon-cutting is funded, the wiring, the electricity bill, the staff who keep it running and the migration of ten years of paper are not.

The commitment is continental and it is happening now

South Africa's e-Impilo, the national electronic medical record for public primary care, went live in March 2025 and was in use at 55 facilities by late 2025, with expansion to 3,300 more planned over two years (Jonda Health, October 2026). Ghana's Lightwave system is reported to be entering a second phase described as a US$100 million project covering 950 facilities (B&FT, November 2025). Rwanda is accelerating the national rollout of e-Buzima, with the United States planning to allocate US$6,806,292 over the first two years of an agreement (published agreement text). Senegal's strategy, validated in April 2025, carries a US$50 million World Bank envelope, and Mali's 2024 to 2028 plan carries EUR 32 million (ICTworks inventory, July 2026). Zambia is extending its national SmartCare record to all facilities in a country whose electrification rate was about 44.5 per cent in 2023 and whose hospitals now run 17-hour load-shedding schedules (Global Fund Zambia case study, April 2025).

Read those figures together and a shape appears: envelopes of tens to low hundreds of millions, for countries of ten to sixty million people, delivering records at thousands of facilities. Despite all of this, the arithmetic that matters, i.e. cost per facility per year including running costs, is almost never published.

What the money actually is

Africa spent roughly US$145 billion on health in 2022, about 5 per cent of continental GDP, and only Cabo Verde and South Africa met the Abuja Declaration target of 15 per cent of government spending on health (ONE health financing data). Against that, "digital health" is not a budget line in any finance ministry on the continent. It is an aggregation that analysts make afterwards, from procurement records, donor agreements and project documents.

Notably, three different clocks are being conflated in this aggregation: donor projects of three to five years with an evaluation at the close, such as the Global Fund's US$50 million Digital Health Impact Accelerator running 2024 to 2026 across five countries (Global Fund); concessional loans that build capital infrastructure but must be repaid, such as Rwanda's US$200 million World Bank project plus a US$100 million AIIB facility, in which health is one service line among several (RISA); and domestic recurrent budget, the only pot that pays salaries, connectivity and support after the project closes. In low income countries, external funding accounted for 31 per cent of health spending in 2022 (ONE).

The scale mismatch is stark. Africa carries roughly 25 per cent of the world's disease burden and receives less than 1 per cent of global digital health investment (Omolo, Lancet Global Health, 2026). The World Bank's own review put its entire digital-health-related lending worldwide at slightly under US$4 billion across 2012 to 2022 (World Bank, Digital-in-Health, 2023).

One number that can be derived is worth stating. Ghana's reported second phase implies roughly US$105,000 per facility in capital terms. That is not a benchmark, because the scope is undefined: it may or may not include devices, connectivity, training, data migration and five years of support. It is, however, the only kind of number that allows a finance ministry to compare a bid. South Africa is building at more than three times Ghana's facility count and has published no contract value, no per-facility cost and no annual running cost for e-Impilo. The country with the largest programme on the continent has the least legible price.

Three deficits money alone cannot close

The WHO African Region's maturity assessment covering 32 countries found most sitting at Phase 3 of 5, none at Phase 1 or Phase 5, five at Phase 4 and six at Phase 2 (State of Digital Health in 32 Countries of the WHO African Region, 2025). Three findings matter more than the average.

Governance is the strongest domain and the workforce the weakest. Mean scores for leadership and governance rose from 3.00 in 2023 to 3.75 in 2025. Ministries can write strategy and chair committees. They cannot yet staff the systems they commission. A record that nobody is trained and rostered to maintain is a database with a login screen.

Infrastructure remains the binding constraint at the point of care. A review of sub-Saharan Africa's EHR lag identified high procurement and maintenance costs, poor electricity supply, inconsistent connectivity and limited computer skills among primary users as the recurring barriers, not clinical resistance (PMC). Zambia makes the point concrete: hydroelectric dependence has produced 17-hour outage windows that hit power, internet and telecommunications at facilities simultaneously.

The continent's strongest grid does not settle the question either, and it is the one e-Impilo intends to run at 3,300 facilities. South Africa's load-shedding peaked in 2023 at 6,948 hours, the equivalent of 290 days and the highest on record. It fell to 1,656 hours in 2024 and to 236 hours in the first half of 2025 (CSIR, September 2025). By October 2026 Eskom reported more than 500 consecutive days without load-shedding, with an energy availability factor of 68.11 per cent, still below 80 (Eskom, October 2026). That recovery is real and it is recent. The uninterrupted run dates from 15 May 2025, after power cuts returned at the end of January and persisted into May. A national record is a decade-long commitment to a system that needs electricity every day, and the grid behind it has produced one clean year after four difficult ones.

Connection is not reliability either. More than nine in ten South African households have mains electricity, but three-quarters of those without it live in urban areas and half of those in Gauteng (Africa Check, October 2026). The facilities where a shared record does the most work, the periurban clinics and the informal-settlement sites, are therefore the least likely to have the power and connectivity a shared record assumes.

Fragmentation is a design choice, not a fate. South Africa's provinces start from very different places: five use a mix of hospital systems, while four reported none at all in 2025 (Jonda Health). Rwanda, with a fraction of the money, has one ministry-owned record, one exchange layer and national identity linkage. The comparison is uncomfortable because it suggests the variable that matters is architectural discipline, not budget size.

Policy: strong on paper, expiring in practice

The policy layer looks healthy in aggregate. WHO counted countries in the African Region with digital health policies and strategies rising from 20 to 36 of 47 between 2016 and 2022 (WHO AHOP policy brief, 2024).

Then look at the instruments individually. As of July 2026, South Africa's National Digital Health Strategy 2019 to 2024 is expired and its successor is referenced in donor documents but not released; Rwanda's standalone plan expired in 2023, Tanzania's in June 2024, Uganda is in its final year with no successor published, and Malawi's expires without an identified replacement (ICTworks inventory, July 2026).

An expired strategy is not a neutral administrative fact. It removes the legal basis for budget lines, for data-sharing agreements, for procurement specifications and for the interoperability standards a vendor must meet. A programme running on an expired strategy is running on precedent. Where policy has been made durable it is by statute rather than strategy: Kenya's Digital Health Act is the clearest case, though it has been challenged in the High Court for inadequate public participation and has operated under a Court of Appeal stay.

Data protection is the second half of the question. At least 44 African countries have enacted dedicated laws, by one 2025 count roughly 80 per cent of African Union members (Digital Policy Alert), yet the African Union's Malabo Convention, in force since 2023, has been ratified by fifteen of 55 (Digital Watch Observatory). Domestic law exists; the continental instrument that would make cross-border health data flows lawful is thin.

Barriers, ranked by damage

The barriers are usually listed as a set. They are better ranked, because the order determines what a ministry should fix first.

Procurement and lock-in. A five-year EMR contract with proprietary data structures converts a software choice into a permanent one. Exit costs are never in the bid. This is the barrier that quietly sets the terms of the other four, which is why the single most useful demand of any ministry costs nothing to meet: publish cost per facility per year, over ten years, split capital and recurrent, so that a parliament, a scheme or a successor minister can test the bid against something.

Donor dependency in the recurrent layer. Ethiopia's eCHIS has been financed through donor funds with no multiyear government budget line, documented as a structural weakness in the implementation literature (Daka et al., Health Research Policy and Systems, 2025). Capability built on grants can be excellent and still be unplannable.

Connectivity and power. Already measured, and already named by clinicians as the reason a system is not used.

Workforce, standards and identity. The weakest domain in the region's own assessment; interoperability named in strategy but not enforced in procurement, so every vendor becomes an island; and no longitudinal record without identity linkage at facility level.

The political clock. A record system is a ten-year build. A minister's horizon is three years and an election. Programmes that survive are the ones whose sponsors deliberately made them boring: statutory, budgeted and unwatched.

Which returns the argument to where it started. The continent's EMR bill is not the problem. The bill is, in real terms, small: less than one per cent of global digital health investment against a quarter of the world's disease burden. The problem is that the bill is quoted as an envelope rather than as a cost, so nobody can see what will be owed in year six. A record system is a maintenance obligation disguised as a procurement. Countries that budget the maintenance will still be running their systems in 2036. Countries that budget the launch will be announcing the successor programme.

---

References

1. Jonda Health, "Beyond NHI: Why Health Data Readiness is Urgent in South Africa", 16 October 2026. https://www.jonda.health/articles/south-africa-e-impilo-nhi-health-data-readiness
2. B&FT, "Lightwave sets the record straight: the real story behind e-health system controversy", 7 November 2025. https://thebftonline.com/article/lightwave-sets-the-record-straight-the-real-story-behind-e-health-system-controversy
3. Global Fund, "Zambia Digital Health Case Study", April 2025. https://www.theglobalfund.org/media/w4gdfjob/publication_zambia-digital-health_casestudy_en.pdf
4. United States Government, Rwanda programme document (e-Buzima allocation). https://www.citizen.org/wp-content/uploads/RWANDA_FL-2026-00021_August-2026-Production_RELEASE.pdf
5. ICTworks, "Every African Country's National Digital Health Strategy in 2026", July 2026. https://www.ictworks.org/updated-every-african-countrys-national-digital-health-strategy-in-2026
6. ONE Data, Health Financing. https://data.one.org/analysis/health-financing
7. The Global Fund, Digital Health and the Digital Health Impact Accelerator. https://www.theglobalfund.org/en/resilient-sustainable-systems-for-health/digital-health
8. Rwanda Information Society Authority, Rwanda Digital Acceleration Project. https://www.risa.gov.rw/projects/rdap
9. Ministry of Finance and Economic Planning (Rwanda), RDAP AIIB facility, December 2021. https://www.minecofin.gov.rw/news-detail/rwanda-digital-acceleration-project-receives-us-100-boost-1
10. Omolo CA, "Bridging the gap: why Africa must invest in AI infrastructure for health care", Lancet Global Health, 2026. https://www.thelancet.com/journals/langlo/article/PIIS2214-109X(26)00100-2/fulltext
11. World Bank, Digital-in-Health: Unlocking the Value for Everyone, 2023. https://documents1.worldbank.org/curated/en/099081723223033632/pdf/P17507505033520b109f180abcad94cfd14.pdf
12. "State of Digital Health in 32 Countries of the WHO African Region", 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC13361429
13. "Why sub-Saharan Africa lags in electronic health record adoption", PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC5654179
14. WHO AHOP, Health Data Digitalization in Africa: Unlocking the potential, 2024. https://iris.who.int/server/api/core/bitstreams/a24b9cd3-cc02-4062-9ad1-6d8150bdb93c/content
15. Digital Policy Alert, "Data Protection in Africa Roundup, 2025". https://digitalpolicyalert.org/blog/data-protection-in-africa-roundup
16. Digital Watch Observatory, "African Union Convention on Cyber Security and Personal Data Protection". https://dig.watch/resource/african-union-convention-on-cyber-security-and-personal-data-protection-african-union
17. Daka DW et al., "Electronic community health information system in Ethiopia", Health Research Policy and Systems, 2025. https://link.springer.com/article/10.1186/s12961-025-01355-3
18. Simeka Health, "2026 Budget Update" (national health allocation, R310.4 billion 2026/27). https://www.simekahealth.co.za/2026-budget-update
19. Global Fund, "South Africa Digital Health Case Study", November 2025. https://www.theglobalfund.org/media/x1dh0ls4/publication_south-africa-digital-health_casestudy_en.pdf
20. Eskom, "More than 500 days without loadshedding, EAF improves to 68.11%, diesel costs fall by R4.79 billion", 2 October 2026. https://www.eskom.co.za/more-than-500-days-without-loadshedding-eaf-improves-to-68-11-diesel-costs-fall-by-r4-79-billion/
21. Africa Check, "#LGE2026 FACTSHEET: what the data tells us about electricity access in South Africa", 14 October 2026. https://africacheck.org/fact-checks/factsheets/lge2026-factsheet-what-data-tells-us-about-electricity-access-south-africa
22. CSIR Energy Research Centre, "Utility-scale power generation statistics in South Africa 2025 (1 January 2025 - 30 June 2025)", September 2025. https://www.csir.co.za/sites/default/files/2025-10/Utility%20Statistics%20Report_29092025.pdf