From Babel to a Common Tongue: Healthcare's Long Search for a Shared Language Part 3 of 5

Part two ended on an uncomfortable ledger: doctors feeding screens while patients waited. But a second failure hides inside that story, quieter and in some ways stranger. All those specialised systems we met, the laboratory platforms and pharmacy modules and imaging archives, generated mountains of information about every patient, and then largely refused to share it. Anyone who has worked in health systems on different continents will recognise the scene: a referral letter printed, signed, scanned and faxed between two buildings that each hold the same patient’s record in a different database. This third instalment asks why the machines never learned to talk, and follows the unlikely road towards a common tongue.

“Help us make every doctor’s office in America a fax free zone by 2020!”

Seema Verma, Administrator, US Centers for Medicare and Medicaid Services, 2018 [1]

Many Systems, No Conversation

Health Level Seven, the standards body founded in 1987 that we met in part two, gave the industry its workhorse: version 2, a compact messaging format that became the backbone of hospital data exchange and remains in wide use today [2, 3]. It worked, after a fashion, but it was written for the computing of its era and left so much room for local interpretation that integration engineers joke that once you have seen one HL7 interface, you have seen one HL7 interface. Every hospital spoke its own dialect, and every new connection became a bespoke translation project. The planned successor, version 3, attempted a far more rigorous information model, but its ambition outran its adoption; as one educational review puts it, version 3 “was not as widely adopted as V2, in part due to its complexity” [3]. A decade of committee work had produced a standard the market quietly declined to use.

The Fax That Would Not Die

The consequences were visible in every clinic. By the late 2010s, as part two showed, virtually every American hospital ran an electronic record. Yet in 2021, 69% of those hospitals were still posting or faxing patient summaries when sending them to another healthcare professional, and 78% still received records that way [4]. The absurdity reached the point where the head of the US Medicare programme stood before an audience of developers in 2018 and pleaded for the fax machine’s extinction [1]. Digitising the filing cabinet, it turned out, was not the same as connecting it. A record that cannot travel is simply a more expensive cabinet.

A Volunteer Writes a Standard

The turn came from outside the committee rooms. In 2011, a health informatician named Grahame Grieve argued in a blog post that health data exchange should work the way the modern web works [3]. His proposal, begun as volunteer work, grew through HL7’s open community process into FHIR, the Fast Healthcare Interoperability Resources standard: patients, medications and observations described as tidy web resources that any developer can request the way a phone requests a weather forecast. The first draft appeared in 2013 with just 49 resources; by 2019, Release 4 carried the standard’s first normative, stable content, and it is openly published for anyone to use [3].

It is worth pausing on how unusual that is. The standards that failed had budgets, committees and mandates. The one that succeeded began as a spare-time argument that the web had already solved this problem, was built out by a community of contributors, and won by being useful before it was official. Healthcare, an industry that measures its spending in trillions, acquired its common tongue from a volunteer who refused to accept that a lab result should be harder to fetch than a weather forecast.

When Sharing Became the Law

Standards alone were not enough; sharing had to become compulsory. In December 2016 the US Congress passed the 21st Century Cures Act, which outlawed “information blocking”, the practice of hoarding patient data for commercial advantage [5]. A separate set of requirements in the 2020 final rule obliged vendors to offer standardised, FHIR-based interfaces, so that patients can pull their own records into applications of their choosing at no cost [6]. And in December 2023, a national trust framework called TEFCA became operational, with five initial networks exchanging data under a common agreement [7]. Add up the dates and the lesson is sobering: it took a workable standard, an act of Congress and the better part of two decades before information could even begin to follow the patient.

A View from the South

Africa’s version of this story contains an inversion that deserves to be better known: on public health reporting, the continent arguably leads the world. DHIS2, an open source platform stewarded by the University of Oslo, is the largest health management information system on earth, running as the official government platform in more than 80 low- and middle-income countries, South Africa among them [8]. Where the continent still struggles is the layer beneath: connecting the care of the individual patient. A Global Fund review published in November 2025 found South Africa’s core systems, including TIER.Net for HIV, the district health information system and the tuberculosis registers, “fragmented and not interoperable” [9]. The response is no longer hypothetical. A national patient registry already verifies identity against the Home Affairs fingerprint database, and after beta testing in 36 clinics across five provinces, the first phase of the national electronic health record went formally live in March 2025, with full-scale implementation targeted from 2026, although several provinces still run no hospital information system at all [9, 10]. The lesson worth importing from the American detour is simple: standards first, mandates early, and never mistake digitisation for connection.

Pipes Before Intelligence

Interoperability is the least glamorous subject in healthcare technology, and possibly the most important. Without it, every promise made about data-driven care collapses into retyping and guesswork. With it, something remarkable becomes possible: the fragments of a patient’s story, scattered across the patchwork we described in part two, can finally be assembled in one place at the moment of care.

And the timing matters more than anyone planned. Just as the pipes began to connect, a technology arrived that could actually read what flows through them. For half a century it had been a distant glow on healthcare’s horizon, flickering, stalling, always promising. In part four, we follow that light as it approaches, and ask what happened when it finally switched on.

References

1. Centers for Medicare and Medicaid Services. Remarks by Administrator Seema Verma at the ONC Interoperability Forum, Washington DC, 6 August 2018. https://www.cms.gov/newsroom/press-releases/speech-remarks-administrator-seema-verma-onc-interoperability-forum-washington-dc

2. Health Level Seven International. About HL7. https://www.hl7.org/about/

3. National Institutes of Health, Office of Data Science Strategy. History of FHIR (FHIR for Research training module). https://nih-odss.github.io/fhir-for-research/modules/fhir-history

4. Office of the National Coordinator for Health IT. Interoperability and Methods of Exchange Among Hospitals in 2021. HealthIT.gov Data Brief. https://healthit.gov/data/data-briefs/interoperability-and-methods-exchange-among-hospitals-2021

5. The White House (archived). Remarks by the President at the 21st Century Cures Act Bill Signing, 13 December 2016. https://obamawhitehouse.archives.gov/the-press-office/2016/12/13/remarks-president-and-vice-president-21st-century-cures-act-bill-signing

6. Federal Register. 21st Century Cures Act: Interoperability, Information Blocking, and the ONC Health IT Certification Program. Final rule, 1 May 2020. https://www.federalregister.gov/documents/2020/05/01/2020-07419/21st-century-cures-act-interoperability-information-blocking-and-the-onc-health-it-certification

7. Healthcare Dive. TEFCA goes live in “Big Bang” for health data exchange. 12 December 2023. https://www.healthcaredive.com/news/tefca-goes-live-health-data-sharing-interoperability/702280/

8. DHIS2. About DHIS2. HISP Centre, University of Oslo. https://dhis2.org/about-2/

9. The Global Fund. South Africa Digital Health Case Study. November 2025. https://www.theglobalfund.org/media/x1dh0ls4/publication_south-africa-digital-health_casestudy_en.pdf

10. Parliamentary Monitoring Group. Question NW4495 to the Minister of Health: reply on electronic medical record implementation. https://pmg.org.za/committee-question/31734/