Health Economics Metrics

Topic 34 of 76 · Health system operational economics

Referral to Treatment (RTT)

Referral to treatment is the elapsed time from a GP's referral to the start of consultant-led treatment. The NHS Constitution sets the standard: 92% of patients should start treatment within 18 weeks. RTT is the single most politically visible operational metric in the English NHS.

Why it matters

Trusts that miss RTT targets face regulatory scrutiny, intervention, and reputational damage; the national waiting list is a front-page number. Every week a patient waits is health lost (waiting in a worse health state — see the QALY math below) and often cost gained (conditions deteriorate; see earlier intervention). Software that saves time anywhere in the referral-to-treatment pathway — triage, diagnostics turnaround, clinic capacity, scheduling — directly mitigates the operational and financial consequences of failing the standard, which is why RTT impact is a first-class benefit line in NHS digital business cases.

The math

RTT performance = patients treated within 18 weeks / total treated × 100
Waiting-time health cost per patient = wait duration × (utility_treated − utility_waiting)

Pathway view: RTT = Σ stage durations (referral triage → first appointment →
diagnostics → decision → treatment) — improve the longest queue, not the
busiest stage (see flow-metrics.md).

Worked example

A specialty treats 5,000 pathway patients/year; mean wait 24 weeks; waiting utility 0.68 vs treated 0.80.

Digital triage plus straight-to-test protocols remove 5 weeks of pure queueing:

QALY gain = 5,000 × (5/52) × (0.80 − 0.68) = 57.7 QALYs/year
Monetized at £20,000–£30,000/QALY (see willingness-to-pay-thresholds.md):
  ≈ £1.15M–£1.73M/year of health value

— plus the trust moves from breaching to meeting the 18-week standard, which has governance value no spreadsheet fully captures.

Software engineering connection

RTT is a lead-time metric over a multi-stage queue — the hospital's version of commit-to-production lead time (see DORA metrics). The improvement method is identical: instrument every stage, find where calendar time pools (it is nearly always handoffs and queues, not clinical work), and remove wait states. Typical software wins: e-triage that routes referrals in hours instead of weekly batches, diagnostic-results push instead of follow-up appointments, and automated straight-to-test criteria. Value the improvement with cost of delay denominated in QALYs/week.

Pitfalls

  • Improving a stage that isn't the constraint — cutting first-appointment waits while diagnostics queues grow just moves the pool.
  • Gaming: pathway resets and clock pauses can improve reported RTT without treating anyone sooner; audit the underlying distribution.
  • Claiming the whole pathway improvement for one tool when several changes landed together — attribution needs a comparator.

Sources