Skip to content
fatiguemanagement.co.uk

Fatigue risk model inputs and outputs

Fatigue risk models turn structured information about a work pattern into modelled estimates of fatigue exposure. This page explains, in plain English, the typical inputs these models use and the outputs they produce — and why the outputs are only as reliable as the inputs.

It is not legal advice. Different tools use different inputs; there is no single universal set.

Typical inputs

Section titled “Typical inputs”

Depending on the model and how a tool is configured, common inputs include:

Input Why it matters
Shift start and finish times When work happens affects fatigue through the body clock — see timing and circadian effects
Shift duration Longer duties can increase time at lower alertness
Night work and early starts These coincide with the circadian low and can compress sleep — see night shift fatigue and early starts and fatigue
Rest between shifts Determines realistic sleep opportunity before the next duty
Consecutive duties Fatigue can build across a run of shifts — see cumulative fatigue in models
Breaks Within-shift recovery affects the job/break component — see workload, breaks and fatigue models
Workload / job demand Vigilance-heavy or physically demanding work can add fatigue where the model supports it
Commute / travel Some organisational policies include travel burden — see duty, travel, and commute

Not every model uses every input, and not every model uses identical inputs. Travel handling in particular varies — there is no single commute assumption that applies universally.

Typical outputs

Section titled “Typical outputs”

Models based on HSE RR446 commonly produce two related outputs:

  • Fatigue Index (FI) — a modelled estimate of the likelihood of very high sleepiness during the duty.
  • Risk Index (RI) — a modelled estimate of relative risk compared with a reference pattern.

See Fatigue Index vs Risk Index for how these differ. Both are modelled estimates of relative exposure — not measures of real-time alertness, and not legal compliance thresholds.

Some tools also break outputs down by the underlying components (timing, cumulative, and job/breaks) so reviewers can see why a pattern scores as it does.

Why output quality depends on input quality

Section titled “Why output quality depends on input quality”

A model can only work with the information it is given. Common input problems and their effects:

Input problem Effect on output
Missing earlier shifts in the schedule Cumulative estimate wrong — often the largest error
Planned times entered when actual hours differed Underestimates exposure after overtime or overruns
Unrealistic break assumptions Understates within-shift demand
Incorrect day/night classification Misaligns the timing estimate
Wrong or absent travel data Misstates total burden where travel is in scope

This is why comparing planned vs actual fatigue matters — outputs based on an out-of-date plan can look reassuring while real exposure has changed.

Interpreting outputs responsibly

Section titled “Interpreting outputs responsibly”
  • Outputs may indicate which patterns are likely to be more tiring — they can support review, not replace it.
  • A favourable output does not prove a shift is safe; an unfavourable one does not automatically mean work cannot proceed.
  • Outputs should be interpreted with competent judgement, alongside worker feedback and wider fatigue risk assessment.
  • Any thresholds applied to outputs are organisational policy thresholds that may vary.

For the boundaries of what models can and cannot tell you, see limitations of fatigue models.

Practical steps

Section titled “Practical steps”
  1. Enter accurate schedule history — include preceding shifts and rest, not just the duty being assessed.
  2. Use actual hours where reviewing what really happened.
  3. Be honest about breaks and workload rather than assuming best case.
  4. Record assumptions — what was entered, and who reviewed it — see fatigue records and audit.
  5. Reassess when timings, travel, or assignments change materially.
Section titled “Related pages”

References

Section titled “References”