BMJ: Engineering for a Medical Publisher That Went Digital First

Most publishers are still working out what digital means for them. The BMJ settled that question in 1995.

It was one of the first major medical journals to launch a website, and has made most of its content freely available online ever since. That head start changes what an engineering engagement looks like. There is no transformation to argue for and no case to make internally. What there is instead is three decades of accumulated digital scholarly record, and a very low tolerance for anything that disturbs it.

The Client

BMJ is a global healthcare knowledge provider committed to improving patient outcomes, through clinical decision support tools, educational resources, events and career services for healthcare professionals worldwide.

That description is worth reading carefully, because it is not a description of a journal. It is four different products with four different technical shapes — a decision support tool, a learning platform, an events operation and a careers service — sharing one masthead and one standard of credibility.

The Challenge

Working with a publisher of this kind concentrates the difficulty in places that are unusual for a content project.

Citations are permanent obligations. In clinical and academic publishing, a URL in a reference list is part of the record. Links accumulated over thirty years cannot be treated as an SEO consideration to tidy after a migration; they are commitments, and breaking them damages something that cannot be repurchased.

Clinical content raises the cost of being wrong. A retail site with a rendering defect has a bug. A clinical decision support tool that displays something incorrectly has a different category of problem, and the engineering standard follows from that.

Entitlement is genuinely complicated. Institutional subscriptions, individual memberships, open access articles and freely available content coexist. A reader may be entitled to one thing through their hospital, another personally, and a third because it was always free. Access control has to resolve all of that without blocking a legitimate reader.

Editorial authority is the asset. Everything technical is subordinate to the credibility of the masthead, and any feature that risks it is not worth its efficiency.

What We Do

iLeaf provides a comprehensive suite of engineering services to BMJ across their digital estate — the platform work behind clinical decision support tools, educational and medical education resources, and the surrounding digital products.

The work is characterised less by any single build than by the standard it is held to: changes to a live scholarly platform, made without interrupting a service that clinicians use in practice and researchers cite permanently.

The Approach

Continuity is a requirement, not a cleanup task. URL and citation continuity is planned before a change and verified after it, against real traffic and real citation data.

Entitlement is tested against reality. Access rules are validated with actual institutional subscription data rather than a model of it, because the failure modes in both directions are expensive — a licensing breach one way, a blocked paying institution the other.

Change is incremental against a live service. A platform serving clinicians does not get a switchover day. Work proceeds capability by capability, each releasable and reversible on its own.

Editorial judgement stays with editors. Automation supports production and discovery; it does not make claims on the masthead's behalf.

Why It Matters

BMJ is the engagement we point to when a publisher asks whether we understand their constraints, because it demonstrates the specific thing that is hard to demonstrate otherwise: sustained engineering on a platform where the archive is the asset, the citations are obligations, and the credibility is not ours to spend.

---

Read more about how we work with media and publishing and healthcare, or start a conversation.