Health Economics Metrics

Topic 39 of 76 · Health system operational economics

Downstream Resource Optimization

Saving an hour for a senior practitioner — a GP, a senior registrar, a consultant — often prevents bottleneck delays for an entire multi-disciplinary team (MDT) of nurses, administrative clerks, and therapists who are waiting on clinical sign-offs. The value of unblocking the bottleneck is the throughput of everyone downstream of it.

Why it matters

Health care runs on authorization chains: discharges wait for consultant sign-off, treatment plans wait for MDT review, referrals wait for triage. When the gating role is delayed, the cost is not one person's hour — it is idle or blocked time across every dependent role, plus patient time in limbo (extra bed days, longer RTT waits). This is the theory of constraints applied to clinical pathways: an hour saved at the constraint is worth the whole system's marginal throughput; an hour saved elsewhere is worth much less.

The math

Value of unblocking = Σ over downstream roles (blocked hours released × unit cost)
                    + pathway throughput gain × value per pathway completion

Contrast: value of the same hour saved at a non-gating role ≈ that role's
capacity value alone (see practitioner-time.md).

Identify the constraint empirically: where does work queue longest? Whose inbox do delays trace back to?

Worked example

A ward's discharges require consultant review each morning. The consultant spends 90 min/day assembling information scattered across systems; reviews finish by 14:00, and 6 discharges/day complete too late for that day — each costing an avoidable bed day.

A discharge-summary dashboard (labs, meds, flags in one view) cuts assembly to 20 minutes; reviews finish by 11:30:

Bed days avoided     = 4 of the 6 late discharges × 365 ≈ 1,460 bed days/year
Downstream unblocking: 2 discharge coordinators + pharmacy + transport
                       previously idle-then-crunched each afternoon —
                       ~3 staff-hours/day of blocked time released ≈ 1,100 hrs/yr

The consultant's own 70 minutes is the smallest part of the value — the point of this metric. Value the bed days by mechanism (see bed days saved) and the staff hours as capacity.

Software engineering connection

This is code review, architecture sign-off, and the staff engineer's inbox. When five engineers wait a day for the one person who can approve a design, the cost is five engineer-days plus a day of cost of delay on the work itself — not one reviewer-hour. Tooling that compresses the gating role's task (better review context, automated pre-checks, dashboards that assemble what the approver needs) buys system throughput, not individual convenience. Measure pickup/wait time at the constraint (see flow metrics) — it is the software equivalent of the 14:00 discharge cliff.

Pitfalls

  • Optimizing a non-constraint: beautiful tooling for a role nothing queues behind produces near-zero system value.
  • Constraint migration: unblock the consultant and the constraint moves (to pharmacy, to transport) — model the next constraint before claiming full throughput gains.
  • Counting downstream hours as cash: blocked-time release is capacity, subject to the usual redeployment test.

Sources