Most founders build features first, then scramble for proof when an NHS buyer asks the awkward question: where’s your evidence this works? A better way is to plan your evidence like you plan your product.
The NICE Evidence Standards Framework (ESF) gives you a clear picture of what “good” looks like for digital health. In this week’s Monday Masterclass we’ll turn ESF into a practical Now → Next → Later evidence roadmap you can execute. This is the same approach we coach our mentees on all the time.
What the NICE ESF is, in plain English
The ESF is a set of standards NHS evaluators use to judge whether a digital health technology has the right kind of evidence. It tells you what to prove, and to what depth, based on how risky your product’s intended use is. That risk determines the bar you need to clear.
NICE classifies digital health tools by intended purpose and risk into three tiers. The higher the risk, the stronger and more clinical your evidence needs to be. Most regulated medical devices and IVDs sit in the highest tier.
NICE updated the ESF in 2022 to cover AI and data-driven tech with adaptive algorithms, and to align more tightly with regulation. If you’re building with machine learning, expect to show how you manage data, performance, and change over time.
The three tiers at a glance
- Tier A: low-risk system or efficiency tools. Prove they work as described and are usable in context.
- Tier B: medium-risk health management and communication tools. Show acceptability, engagement, and early outcome signals in the real world.
- Tier C: high-risk clinical interventions and most regulated devices/SaMD. Demonstrate effectiveness and safety with trial-level or comparative evidence, and start building economic value cases.
Step 1: Classify your product before you design your study
Your tier flows from your intended purpose this year, not your five-year vision. Write down the claim you want to make in the next 12 months in a single sentence. Then sanity-check it:
- Is it purely workflow or admin? You’re likely Tier A.
- Is it health management without diagnosing or treating? You’re often Tier B.
- Does it inform diagnosis or treatment, or deliver therapy? Assume Tier C and plan accordingly.
Getting this right avoids two common traps: under-evidence for high-risk claims, or over-engineering studies for low-risk use cases.
Step 2: Turn ESF into a Now → Next → Later plan
Use Now → Next → Later to pace your proof, line up the right partners, and tie evidence to funding. If you missed our primer on this roadmap model, read it here and come back: Can you make your business model work on paper?
Now (0–3 months): feasibility, safety basics, usability
Your goal is credible early proof and no nasty surprises in governance.
- Evidence: user acceptability and usability in real settings; basic performance checks; early signals your outcomes are moving in the right direction.
- Governance: start your clinical safety approach and risk logs; get your data protection and security hygiene in order; prepare for DTAC by knowing what evidence buyers will ask for at procurement. NHS Transformation Directorate
- Output: a short “Feasibility & Acceptability” pack you can share with pilot sites.
Next (3–12 months): pilot outcomes and real-world evidence
This phase turns promise into practical results.
- Evidence: a single-site or multi-site pilot; pre-post outcomes; clinician and patient-reported measures; operational metrics; early economic signals.
- Regulatory prep: if you’re a medical device or SaMD, keep a tidy technical file and a live clinical evaluation plan. Maintain your clinical safety case and DTAC artefacts so a trust can actually buy you.
- Output: a clear pilot report with methods, results, and limitations an NHS evaluator can trust.
Later (12–24+ months): comparative effectiveness and economics
Now you go beyond “works here” to “works better than what we do today and is value for money.”
- Evidence: a controlled comparative study or pragmatic real-world evaluation; cost-effectiveness or budget impact modelling aligned to the pathway you want to live in.
- Adoption: you are ready for formal evaluations or scaled procurement conversations, and can consider NICE engagement if the product is a fit for their programmes. NICE
Three practical archetypes (no tables, just what to do)
1) Digital health app (likely Tier B)
What it does: supports self-management, behaviour change, or symptom tracking without diagnosing or treating.
Now
- Define your primary outcome in language clinicians care about.
- Run quick usability sessions with target users and one clinical champion.
- Document privacy and security basics; list the minimal DTAC evidence you can assemble today.
Next
- Run a 6–12 week pilot in one setting. Pre-post scores on the primary outcome, retention, completion, and any workflow wins.
- Collect simple cost signals. For instance, fewer unnecessary appointments or shorter reviews.
- Keep a tidy pack: protocol, consent, measures, results, and your limitations.
Later
- Do a comparative study in two or more sites or a pragmatic RWE design with a clear comparator.
- Build a simple budget impact model that matches the local pathway.
- Publish. Even a short service evaluation write-up increases credibility.
2) Clinical software / SaMD (usually Tier C)
What it does: informs clinical decisions or delivers an intervention.
Now
- Freeze this year’s intended claims. If they’re diagnostic or therapeutic, plan for medical device obligations.
- Start your clinical safety case (DCB0129 for developers) and nominate a Clinical Safety Officer.
- Design a feasibility study focused on accuracy, reliability, and safe use in context.
Next
- Run a prospective pilot with ground truth where relevant. Pre-specify metrics that matter to clinicians.
- Keep your technical file current and your clinical evaluation plan alive.
- Maintain DTAC artefacts so a trust can adopt post-pilot without months of rework.
Later
- Step up to a controlled comparative design or equivalent real-world evaluation.
- Start health economic modelling alongside.
- If appropriate, prepare for a NICE interaction and ensure your dossier is publication-ready.
3) Hardware device (usually Tier C)
What it does: measures or treats with a physical component, often paired with software.
Now
- Map your risk class and essential performance; plan bench tests and human factors work.
- Start the clinical safety case and basic cybersecurity hygiene if it connects.
Next
- Early clinical use under a defined protocol. Capture safety events, performance against reference standards, and usability in the wild.
- Build evidence an NHS buyer can understand: what changes for staff and patients, and what it costs.
Later
- Comparative clinical study with outcomes that matter to the pathway and a credible budget impact model.
- Publish and prime your evidence for external evaluation and scaled procurement conversations.
Common traps we see — and how to avoid them
- Over-claiming too early. Wellness copy that implies diagnosis or treatment flips you into device rules, without the evidence or documentation to back it up. Tighten the claim, or step up your evidence and compliance.
- Collecting the wrong outcomes. If a respiratory clinic cares about exacerbations and admissions, your pilot must measure those, not just app engagement.
- Leaving governance to the end. DTAC, clinical safety (DCB0129/0160), DPIA and security are not optional. NHS buyers expect to see them at procurement. Start early and keep a neat pack.
How this fits your wider roadmap
Evidence isn’t a side-quest. It sits alongside product and commercial milestones in your Now → Next → Later plan, so fundraising and NHS conversations happen with credibility. We’ll be coming back to this topic in upcoming posts showing how you can use your evidence generation strategy to structure your fundraising activities and get clarity on the key milestones you need to hit at each phase.
Want more like this?
If you found this helpful and want more posts like this, subscribe to Monday Masterclass below. If you’d like to book a paid consultation with our mentors to discuss in more detail you can drop us an email here.
Educational only. This is not legal or regulatory advice. Always check current NHS, NICE, and MHRA guidance for your specific product and pathway.