Laboratory Workload Measurement
Converting laboratory activity into comparable units of staff time, so that staffing and capacity can be argued from measurement rather than impression.
What it is
Specimen counts are the usual proxy for laboratory workload, and they are a poor one: a Whipple resection and a skin shave both count as one specimen while differing by an order of magnitude in technician time. Workload measurement replaces the count with a time-weighted unit, so that two departments — or the same department in two years — can be compared on something closer to the actual work done.
The general shape is always the same. Time each process, pick a reference process, express everything as a fraction of it, then multiply by volume to get a department’s total. The choice of reference is a design decision rather than a detail, because the whole scale inherits the measurement error of whatever is chosen as the anchor.
The worked example here is the Spanish Society of Anatomical Pathology’s pilot (Tresserra Casas et al., Rev Esp Patol 2026;59(3):100882), which timed technical processes over five days in ten hospitals and defined a Technical Workload Unit of 11 minutes, anchored to large-specimen embedding as the most time-consuming task. On that scale microtomy is 0.09 TWU, immunohistochemistry about 0.19 TWU, registration about 0.36 TWU, and archiving about 0.04 TWU.
Three things determine whether such a number transfers to another laboratory, and they are worth checking before adopting anyone’s published unit:
- Case mix. A per-specimen median is an average over the specimen mix it was measured in.
- Batch size. Microtomy and IHC are batched, so per-specimen time falls as runs get larger. A per-specimen figure is only portable alongside the batch size behind it.
- Scope of practice. What a technician does versus what a pathologist or an assistant does varies by country and by department, so process boundaries are not comparable by default.
Why it matters for my work
The group already measures operational time, but from the instrument side rather than the staffing side: Scanning Time in Real Life quantifies scanner utilisation, transfer, and queueing. Workload measurement is the human counterpart, and the two answer different questions — one asks whether the scanners can keep up, the other asks whether the people can. A capacity argument usually needs both.
It also gives the group an external benchmark. A national society’s multi-centre figures are a more defensible comparator than internal numbers alone, so local measurements can be reported as a difference from a published reference rather than in isolation. That said, the SEAP figures are a self-described pilot from one country, and the source note lists why they should be re-measured locally rather than adopted — the strongest reason being that ten hospitals were pooled into a single median with no between-site variation reported, which is exactly the number that would tell you whether the benchmark travels.
How it connects
Turnaround Time — the other half of the operational picture, and often confused with this one: turnaround measures elapsed time from the specimen’s point of view, workload measures staff time from the department’s. A department can have excellent turnaround and unsustainable workload at the same time.
Scanning Time in Real Life — measures the instrument side of the same laboratory; combining its queueing findings with per-process staff time is what turns either into a capacity model.
Konsultasyon — consultation cases carry workload that specimen counts miss entirely, since a referred-in case may consume far more time than its single accession suggests.
Patoloji Bilgi Yönetim Sistemi (LIS) — the LIS is where specimen counts and timestamps come from, and its documented caveat that field reliability varies by subspecialty applies directly to any workload denominator built from it.
Tresserra 2026 — Technical Workload Unit (SEAP pilot) — the only measured unit this page rests on: a Spanish multi-centre time-and-motion pilot defining a Technical Workload Unit of 11 minutes, anchored to large-specimen embedding.
Open questions
- Does the group have per-process technician timings of its own, or only specimen counts and scanner logs? Nothing in the repo records staff-time measurement.
[unverified] - If Memorial were to measure this, the design should fix the flaws in the source: report between-observer and between-day variation, record batch sizes, and stratify “large specimen” by complexity. That would make a local study publishable rather than merely internal.
- Would a workload unit anchored to a high-volume, low-variance process (microtomy) be more stable than one anchored to the most time-consuming one? The SEAP choice is defensible but ties the scale to the noisiest estimate.
- How does this relate to the SUT billing data already used in
MemorialPathStats? Billing codes are a third proxy for workload, with their own distortions. Not yet documented here.[unverified]