Explainer: new & follow-up attendances

Where outpatient demand really comes from: first attendances are generated by referrals at the start of the pathway; follow-ups are generated by treatment at the end. Attribute both to the same driver and the model misstates slot requirements — in size and in timing. This page shows the pathway, quantifies the error on the real NUH model, walks the slot arithmetic, and demonstrates how capacity choices move the waiting population along the pathway.

Loading model…

1  One pathway, two demand generators

referrals queue 1 awaiting FIRST outpatient attendance first attendance consumes a NEW slot discharged / leave the list (most “other removals” attended first) onward queue 2 awaiting decision, diagnostics & treatment treatment the clock stop follow-ups (N:FU ratio) — consume FU slots both queues together are “the RTT waiting list” — one number, two populations

The model's activity conversion follows this structure: first attendances = referrals (raw referrals — even the ~10.7% who later leave the list without treatment mostly attend a first appointment), and follow-ups = clock stops × N:FU ratio (treatment generates the review burden). An earlier draft attributed firsts to clock stops; a reviewer caught it, and section 2 shows why it matters.

2  Why attribution matters — the clearance surge

During the recovery, required treatments surge (demand + backlog clearance) while referrals only grow at trend. Tie first attendances to treatments and new-appointment demand appears to surge with the theatre programme — capacity that would be commissioned in the wrong clinics at the wrong time. Computed live from the NUH baseline:

The two drivers (per month, trust total)

OP slots required under each attribution

3  The slot arithmetic

Slots ≠ attendances: every attendance needs a bookable slot, uplifted for patients who do not attend (DNA) and for clinic time that cannot be filled (utilisation): slots = (firsts + follow-ups) ÷ (1 − DNA) ÷ utilisation. Follow-up burden is the biggest lever — this is where PIFU (patient-initiated follow-up) acts.

4  How capacity choices move the waiting population

The two queues are served by different capacity: queue 1 by new slots, queue 2 by treatment capacity. Boost one without the other and the waiting population doesn't shrink — it moves. The simulator runs the tandem queue for two years; drag the capacities and watch where the queue goes.

The two waiting populations (tandem queue, weekly steps)

Classic patterns to try: treatment surge only (treatment slots up, new slots at referral rate) — queue 2 drains but queue 1 is untouched and the "list" barely moves at first; front-door surge only (new slots up) — queue 1 drains straight into queue 2, which balloons: the list changes shape, not size, and the treatment requirement arrives a few months later. A recovery plan needs both, phased: front-door capacity leads, treatment capacity follows the conversion wave. This is why the main model drives first attendances from referrals and treatments from the clearance path — the two surges are different sizes at different times.