Regulated Products
If an NHS organisation is going to use your product, you have to clear DTAC before the pilot, not just before procurement. Here's what that involves in 2026, the roles you'll need, and how to sequence it so it doesn't derail the build.
| Name | What it is | Who it applies to |
|---|---|---|
| DTAC | The assessment NHS bodies use to check a product before adopting it | Anyone selling or piloting digital tech into the NHS |
| DCB0129 | Clinical risk management standard for the makers of health IT | You, the manufacturer |
| DCB0160 | The matching standard for the NHS organisation deploying it | Your NHS customer, not you |
| DSPT | Annual data security and protection self-assessment | Any organisation handling NHS patient data |
| DPIA | Data protection impact assessment under UK GDPR | Any product processing personal health data |
DTAC assesses five areas: clinical safety, data protection, technical security, interoperability, and usability and accessibility. It is not a certification you pass once. It's a structured set of evidence an NHS organisation reviews before they let your product near patients or staff, and crucially it's required before pilots as well as full procurement. Teams are routinely caught out by planning a “quick pilot” and discovering it needs the same evidence pack as a contract.
The February 2026 update, widely called DTAC 2.0, trimmed roughly a quarter of the questions by removing overlap with the DSPT and the pre-acquisition questionnaire. It's less repetitive than it was, but the underlying bar has not dropped.
This is the one that shapes the product rather than just documenting it. DCB0129 requires you to systematically identify, assess, and mitigate the clinical risks that could arise from someone using your software: a wrong dose shown, a missed alert, an ambiguous label, a screen that reads differently under pressure.
In practice it means:
The Data Security and Protection Toolkit is an annual self-assessment against a set of data security standards, with evidence attached. If you process NHS patient data you complete it yearly, and your DTAC submission references it. Start it early: gathering the evidence (policies, training records, access controls, penetration test results) takes longer than filling in the form.
| Item | What it costs you |
|---|---|
| Clinical Safety Officer | A retained clinician, from discovery through every release |
| Hazard log and safety case | Ongoing analyst and clinical time; a few weeks of concentrated work up front |
| DSPT completion | Weeks of evidence-gathering, then annual upkeep |
| DPIA | 1 to 2 weeks with a data protection specialist |
| Extra QA against hazards | 10 to 20% on top of a normal test cycle |
| Timeline | Typically 4 to 8 weeks of additional elapsed time, mostly overlappable with the build |
The teams that struggle treat compliance as a phase at the end. The teams that don't fold it in from the start:
Done this way, the compliance work adds weeks, not months, and it genuinely improves the product: the hazard analysis forces you to design the unhappy paths properly, which is where health software usually fails.
A health product that clears assurance smoothly usually shares a few traits: a Clinical Safety Officer who was in the room from the first design review, a hazard log that started as a spreadsheet in discovery week one, a deliberately narrow first release so the safety case is small and defensible, and a DPIA and DSPT that were opened months before they were needed. The teams that treat it as a design input rather than a launch gate are the ones that don't slip.
If you're building something the NHS will use, talk to us early, ideally before design. Folding clinical safety in from the start is far cheaper than bolting it on. You can see the kind of work we mean on our healthcare projects page.
Minoo
Minoo works on products in regulated and clinical settings, where shipping fast and shipping safely are the same conversation.
Turn an idea into a product, brand, or campaign, with one team from strategy to launch.
start a projectEngineering
The best tech stack for a first build is a boring one: proven, well-documented, and easy to hire for. Here's what “boring” actually means in 2026, the traps that sink second years, and a sensible default to start from.
11 Sep 2026 · 10 min read
Regulated Products
You rarely need your own licence to launch a payments or e-money product. Here are the routes that let you go live in weeks instead of a year, what you still own, and the safeguarding change landing in May 2026.
8 Sep 2026 · 11 min read
Product Strategy
Every founder asks it, and “it depends” is a useless answer. Here's the real 2026 range for a UK build, the five things that actually set your number, how offshore and native change it, and the three places budgets quietly leak.
2 Sep 2026 · 11 min read