Designing for accountability:
a QA framework that outlasts the project

Where it started

Something I have noticed across a lot of organisations is that quality assurance in digital learning tends to be treated as an end-stage activity. You build the content, you check it for errors, you publish it. The assumption is that quality is primarily a production problem. In a regulated healthcare environment, that assumption causes real harm. An error in patient-facing health content is not a design issue that can be quietly fixed in the next version. It has consequences for real people.

When I audited the existing QA process at MyWay Digital Health, I found something that is actually quite common in fast-growing organisations. The process had been designed for a much smaller, simpler content operation than the one the organisation had become. There was no consistent sign-off process for clinical accuracy. Accessibility checks against WCAG 2.1 were not structured. Multilingual content going out to Arabic, Malaysian, and Singaporean markets had no formal cultural quality controls. And there was no audit trail that could demonstrate compliance if the platform was ever reviewed externally. The organisation was serving over 26,000 users on an NHS-integrated platform and its governance infrastructure had not kept pace.

What made it complex

The redesign needed to add rigour without adding so much process that it slowed everything down. That balance is harder to get right than it sounds. Too light and the framework does not actually catch anything. Too heavy and teams start working around it. I also had to make sure it worked across four different types of review, clinical accuracy, accessibility, multilingual and cultural quality, and instructional design, without those reviews blocking each other or creating rework late in the process.

How I approached it

The key decision was to run the four review streams in parallel rather than in sequence. Sequential review is how most organisations do it, and it creates a particular kind of problem: a clinical reviewer flags something in week three that requires changes to content that has already been through accessibility and translation review in weeks one and two. Everything has to go back around. Parallel review means all four streams are active from the start, issues surface earlier when they are easier to fix, and the overall cycle time actually comes down even though you are doing more.

Each stream had defined sign-off criteria and a documented audit trail. Version control was embedded throughout so that any piece of content reaching publication could be traced back through every stage of its development. That traceability is what makes a framework genuinely audit-ready rather than just well-intentioned.

What came out of it

The framework was adopted as the organisational standard across all digital learning content. Course completion rates across redesigned content reached 76% against a 60% industry benchmark, which reflects the cumulative effect of designing with quality built in from the start. The governance model we built for the multilingual Ramadan course came directly out of this framework and was one of the first real tests of it in practice.