Work Productivity and Activity Impairment (WPAI)

WPAI is a validated self-report questionnaire (Reilly, Zbrozek, Dasbach, 1993) measuring how much a health problem affects paid work and daily activities, usually over the past 7 days. It splits the loss into absenteeism — work time literally missed — and presenteeism — reduced productivity while physically at work — and the second is usually the larger, more hidden cost component.

Why it matters

Simple sick-day counts only see absenteeism. A clinician or knowledge worker who never takes a day off but works at 60% capacity through a chronic condition contributes zero to an absence register while still generating a large, real productivity loss — WPAI is designed specifically to surface that invisible cost. Because it is a validated instrument rather than a bespoke survey, its scores are usable in patient-reported outcome evidence packages and cost-of-illness studies without the reviewer needing to re-validate the measure. As a self-report instrument, it is itself a form of PROM, distinguished mainly by its focus on work and activity rather than symptoms or quality of life.

The math

Absenteeism % = hours_missed_due_to_health / (hours_missed_due_to_health + hours_worked) × 100

Presenteeism %  = self-rated 0–10 impairment while working, × 10
                  (elicited directly via questionnaire, not derived here)

Overall Work Impairment % =
    Absenteeism% + (1 − Absenteeism%/100) × Presenteeism%
    (combines the two so the total can never exceed 100%)

Productivity cost = Overall Work Impairment% / 100 × period_earnings

The overall-impairment formula is deliberately not a simple sum: adding the two percentages directly could exceed 100%, so presenteeism is applied only to the remaining (non-absent) share of work time.

Worked example

An employee with migraine reports being scheduled for a 40-hour week but misses 4 hours of it:

hours_missed = 4, hours_worked = 36
Absenteeism% = 4 / (4 + 36) × 100 = 10%

They separately self-rate their productivity impact while working as 3 out of 10 on the WPAI questionnaire, i.e. Presenteeism% = 30% (this step is a raw questionnaire answer, not something derived from other numbers):

Overall Work Impairment% = 10 + (1 − 10/100) × 30
                          = 10 + 0.9 × 30
                          = 10 + 27
                          = 37%

Over a 5-day week earning £800 (£160/day):

Productivity cost = 37/100 × 800 = £296

Notice that a naive sick-day count would have recorded only the 4 hours (10%) missed — the presenteeism component nearly triples the true impairment once counted.

Software engineering connection

This maps directly onto engineering-team health metrics:

Absenteeism is sick leave and PTO — visible, already tracked, and the easy part.
Presenteeism is the burned-out or context-switching-overloaded engineer who is present in every stand-up while operating at reduced capacity — usually the larger and more hidden cost, and invisible to headcount or attendance data. It shows up instead as reduced throughput in DORA and flow metrics, or as slower resolution of the very technical debt whose "interest" compounds the impairment further.
The engineering lesson is the same as the clinical one: measuring only absence and calling it "productivity loss" systematically understates the real cost, because it misses everyone who is present but impaired.

Pitfalls

Self-report recall bias. A 7-day recall window is subject to the same reporting distortions as any retrospective self-report.
Treating the 0–10 presenteeism scale as a true physical measurement. It is ordinal, elicited by self-rating, not a validated physical quantity — treating differences on it as strictly linear or interval is a modeling convenience, not a validated physical fact.
Pooling scores across WPAI variants. WPAI has several condition-specific versions — WPAI:GH (general health), WPAI:SHP (specific health problem), and disease-specific variants — and scores from different variants should not be pooled or compared without first checking they are the same instrument version.

Sources

Reilly MC, Zbrozek AS, Dasbach EJ. "The validity and reproducibility of a work productivity and activity impairment instrument." PharmacoEconomics 1993;4(5):353-65.
WPAI instrument documentation, Reilly Associates — the official scoring reference. https://www.reillyassociates.net/