Health Economics Metrics

Topic 29 of 76 · Health system operational economics

Length of Stay (LOS)

Length of stay is the number of days from hospital admission to discharge — the core flow-efficiency metric of inpatient care. UK acute means run around 4–5 days; every excess day consumes a scarce bed and exposes the patient to hospital-acquired risks.

Why it matters

LOS drives almost everything in acute-hospital economics: bed capacity, elective throughput, emergency flow, staffing. Reducing average LOS by even fractions of a day at scale releases enormous capacity (see bed days saved). LOS is also a quality signal in both directions — too long suggests process failure (delayed diagnostics, discharge paperwork, social-care waits); too short can mean premature discharge, which shows up later as readmissions.

The math

LOS (per spell)  = discharge date − admission date
Average LOS      = occupied bed days / discharges (report mean AND median;
                   LOS is heavily right-skewed by long-stay outliers)

Comparisons require case-mix adjustment (age, diagnosis, acuity),
or you are measuring who the hospital admits, not how it performs.

Little's Law connects the flow variables: beds occupied = admission rate × average LOS — the same law that governs software queues (see flow metrics).

Worked example

A trust admits 40 emergency medical patients/day at mean LOS 6.0 days: 240 beds permanently occupied (40 × 6). Discharge-coordination software (task tracking, pharmacy-to-take-out automation, transport booking) cuts the non-clinical tail of stays by 0.4 days on average.

Beds needed = 40 × 5.6 = 224 → 16 beds freed continuously
            = 16 × 365 = 5,840 bed days/year

Value the 5,840 bed days by mechanism (refill/close/slack) per bed days saved. Note what moved: not medicine, but waiting — the patient was medically fit; the system was still doing paperwork. That's a queueing problem, and software is good at queueing problems.

Software engineering connection

LOS is the hospital's cycle time, and the improvement playbook is identical to delivery-flow work: instrument the stages (admission → treatment → medically-fit → actually-discharged), find where time pools (it's the handoffs), remove wait states rather than adding capacity. The "medically fit for discharge but still occupying a bed" cohort is the hospital's version of a PR approved but not merged. Direct software opportunities: discharge task orchestration, diagnostic turnaround, e-prescribing of discharge meds, social-care referral integration.

Pitfalls

  • Mean-only reporting — outliers dominate; a falling mean can hide a growing long-stay tail.
  • No case-mix adjustment in before/after claims: admission thresholds change seasonally and secularly.
  • LOS reduction that reappears as readmission — always pair LOS claims with 30-day readmission data.

Sources