Evidence Plans as Test-Driven Development: Building NHS Confidence Early
Too many NHS programmes write evaluation plans at the end, just before the "scale-up" stage. By then it's too late — data hasn't been collected, metrics are unclear, and pilots can't show impact.
The result? Programmes stall at the "valley of death" between pilot and adoption.
What if evidence was built in from the start, the way engineers use test-driven development (TDD)?
In software, TDD means writing tests before you write code. In NHS transformation, it means writing evidence plans before you deploy a single tool.
Why Evidence Matters
- Buyers demand it: NHSE, NICE, and ICS boards need proof before they commit.
- Clinicians trust it: Without credible outcomes, frontline staff won't adopt.
- Patients deserve it: New models should demonstrate safety and benefit, not just promise them.
Tools for NHS Teams
1. Evidence as a Sprint Deliverable
Every sprint ends not just with a working feature, but with a tested outcome metric.
- Example: Remote monitoring rollout → metric is reduced DNAs within 4 weeks.
- In practice, this may mean starting with proxy measures (e.g. number of referrals logged, % completed remotely) until systems catch up.
2. Logic Models and Driver Diagrams
Map the chain between interventions and outcomes:
- Primary outcomes: reduced admissions, faster discharge.
- Secondary drivers: staff engagement, referral quality.
- Interventions: training, dashboards, workflow redesign.
This creates a transparent evidence map for boards.
3. Evidence Bundles
Pre-build "evidence bundles" that cover:
- Safety: Hazard logs, incident reports.
- Effectiveness: Before/after outcome data.
- Experience: PROMs, PREMs.
- Economics: Cost-consequence analysis.
Bundles should be concise (2 pages) for board use, with links to detail where needed.
Evidence in Practice
- NICE Early Value Assessment (EVA): Teams that collected outcome metrics from day one advanced faster through adoption. Those that waited had to repeat pilots.
- Denmark & Netherlands: Community pilots embedded evaluation at the start. Evidence was linked to reimbursement decisions, ensuring scale-up credibility.
Making Evidence Plans NHS-Ready
Data Infrastructure: NHS systems are fragmented. Start with proxy metrics/manual audits, then automate as maturity grows.
Expertise: Most PMOs lack evaluation skills. Budget for 0.5–1.0 WTE evaluation specialist per pathway (~£50–70k). Use shared templates or pooled regional resources to scale expertise.
Baselines: Run baseline sprints in the first 4–6 weeks. Document baseline reliability (low/medium/high) to avoid false precision.
Frontline Burden: Automate PROMs/PREMs (SMS, app, kiosks). Fund 0.1–0.2 WTE clinical time per pathway for evidence collection.
Attribution: Use contribution analysis, not strict attribution. Document context factors and parallel programmes. Graph-based mapping can help visualise overlaps.
Governance: Evidence bundles = 2-page summaries with headlines, baselines, confidence rating, risks. Negative results trigger Pathway Steering Group arbitration (pivot/stop/scale).
Resources: Allocate 10–15% of budget. Typical breakdown (per £1m programme):
- Evaluation lead: £50–70k
- Analyst: £40–50k
- Clinical time: £10–20k
- PROM/PREM system: £20–40k
- Total: £120–180k
The Economics of Evidence
Delayed evidence collection means wasted investment:
- A £1m pilot without data → cannot scale → 100% value lost.
- A £1m pilot with structured evidence → supports adoption → avoids repeated pilots and parallel proofs.
Investing 10–15% of budget up front in evidence saves millions downstream.
Takeaway
Evidence is the NHS adoption currency.
By treating evaluation like test-driven development — planned up front, tested incrementally, and bundled for governance — NHS teams can move faster and with more credibility.
Evidence plans aren't paperwork at the end. They are the test cases that prove change is safe, effective, and worth scaling.
👉 Next in this series: The Economics of Delay: Why NHS Waiting Times Cost More Than They Save.