How to Screen Overtime Fatigue Risk in 21 Days

Overtime rarely arrives as a single safety decision. A supervisor adds two hours, a planner extends a turnaround, and a manager approves the change because the work is almost finished. The fatigue exposure grows through those small approvals, while the schedule still looks manageable on paper.

This guide shows EHS and operations leaders how to screen overtime fatigue risk in 21 days. The aim is not to diagnose a worker or replace medical advice. The aim is to identify when work design, recovery time, and decision quality have moved outside the operating envelope.

What you need before starting

Assign one operational owner, one EHS reviewer, and one workforce representative who understands the real schedule. Collect three weeks of planned hours, actual hours, shift changes, call-outs, travel time, and high-consequence tasks. If your site already uses a fatigue risk escalation trigger, keep its escalation logic and use this review to improve the evidence behind each decision.

Do not begin with a survey alone. Fatigue risk becomes visible when schedule data is compared with task demand, recovery opportunity, and the controls that supervisors can actually use.

Step 1: Define the overtime change

Write down exactly what will change. Record the number of extra hours, the affected shifts, the duration of the arrangement, the teams involved, and the reason for the request. A vague statement such as “the crew may work longer” cannot support a reliable review because it hides the exposure window.

Separate planned overtime from emergency overtime. Planned changes can be screened before approval, whereas emergency work needs a rapid decision rule that identifies who may authorize it and when the work must pause. This distinction prevents an exceptional response from becoming a normal roster pattern.

At the end of this step, the owner should be able to answer which people, tasks, and dates are inside the review. If that boundary is unclear, do not move to approval.

Step 2: Map recovery time, not only hours worked

Total hours matter, but recovery time explains why the same overtime pattern affects people differently. Record the interval between shifts, commuting time when it is material, overnight call-outs, quick returns, and the number of consecutive workdays. A schedule with acceptable weekly hours can still create a short recovery window before a critical task.

Review the difference between planned and actual recovery. If a shift ends at 10 p.m. but travel, handover, and transport keep the worker away from home until midnight, the roster is not providing a two-hour recovery advantage merely because the timecard closes at 10 p.m.

Use the existing night-shift fatigue check when night work is included, but do not treat that checklist as a substitute for the overtime map. The new question is how the additional hours change recovery across the full cycle.

Step 3: Mark high-consequence work

Classify the tasks that could create severe harm if attention, reaction time, or judgement deteriorates. Examples include energy isolation, lifting near people, work at height, confined-space entry, driving, process changes, and emergency response. The screening threshold should be tighter for these tasks because a small decline in decision quality can have a large consequence.

Ask the task owner which decisions must be made without delay and which controls depend on a second person. This reveals whether the overtime change affects a critical barrier, a lone-worker arrangement, or a handover where important information can be lost.

James Reason’s work on latent failures is useful here because fatigue is rarely an isolated personal issue. The schedule can interact with weak planning, thin supervision, poor equipment design, and production pressure, creating a pathway in which a predictable human limitation becomes an incident precursor.

Step 4: Compare the plan with field reality

Sample the last seven days of actual work rather than relying only on the proposed roster. Compare start and finish times, breaks, task duration, reassignment, and unplanned call-outs with the schedule that managers approved. The gap between the two is evidence about how the system behaves under pressure.

Interview a supervisor and two workers from different shifts. Ask when the overtime was requested, who could decline it, whether the break occurred, and what work was delayed or compressed. Use specific examples instead of asking whether people felt tired, because a general question can produce polite agreement while missing the decision that created exposure.

When the evidence shows repeated divergence, document the pattern as an operational control issue. It should not be described only as a worker’s failure to manage fatigue.

Step 5: Test the decision rights

Identify who can approve overtime, who can reject a fatigued assignment, and who can move a high-consequence task to another crew. Then test those rights with a realistic scenario. For example, ask what happens when a technician reaches the end of a long shift while an energized system still needs isolation.

A decision right is real only when the person can use it without needing informal permission from a manager who is measured on production completion. If the answer depends on personal courage, the control is fragile. Establish an escalation route that names the next decision-maker and the response time.

Andreza Araujo’s work on safety culture emphasizes the difference between a declared value and a repeated operating choice. In this review, the operating choice is whether the site protects recovery and critical decisions when the schedule becomes inconvenient.

Step 6: Set temporary controls for the 21-day test

Choose controls that can operate during the review rather than waiting for a full roster redesign. Depending on the evidence, they may include a second-person check for critical isolations, a ban on assigning certain tasks after a defined recovery breach, a protected break, transport support, a supervisor handover, or a named fatigue escalation trigger.

Each control needs an owner, a start date, a verification method, and a condition for removal. “Remind the team to rest” is not a control because it does not specify what changes in the work. “No lone isolation after a short recovery interval, verified by the shift supervisor before the permit is released” can be tested.

Keep the set small. Three controls that are used and checked will produce better evidence than ten actions that exist only in a meeting record.

Step 7: Review evidence at days 7 and 14

Do not wait until day 21 to discover that the review was producing no usable evidence. At day 7, check whether actual hours, recovery intervals, high-consequence assignments, and exceptions are being recorded consistently. At day 14, compare the first and second weeks and look for repeated breaches rather than isolated variation.

Review near misses, stop-work decisions, delayed maintenance, rework, handover errors, and complaints about schedule pressure alongside the roster data. These signals do not prove that fatigue caused an event, but they can show where the work system is losing margin.

If a measure is not changing a decision, remove it or replace it. A metric that only fills a dashboard can create the appearance of control while the operational question remains unanswered.

Step 8: Decide what happens after day 21

At the end of the test, make one of three decisions. Keep the arrangement with defined controls, modify the roster and repeat the review, or stop the overtime pattern because the risk cannot be controlled under the current design. Record the evidence for the decision, including what was observed and what remained uncertain.

The final review should also identify the next owner. A temporary control becomes a durable control only when someone is responsible for checking it after the project or emergency has ended. Link the result to the site’s broader psychosocial risk screening process when the change affects workload, recovery, role conflict, or worker participation.

If the review finds a health concern, refer the person to the appropriate occupational health or medical pathway. Schedule screening is a management control, not a clinical assessment.

Final checklist for the review owner

1. Confirm the overtime scope, dates, shifts, and affected tasks.
2. Compare planned hours with actual hours and recovery intervals.
3. Mark high-consequence work and decisions that require a second person.
4. Test whether supervisors and workers can use the stated decision rights.
5. Assign no more than three temporary controls with clear verification.
6. Review evidence on days 7, 14, and 21, then record the next decision.

What this review changes

A fatigue review should change the way the site approves work, not simply add another form. When overtime is treated as a schedule change with consequences for recovery, task allocation, and decision rights, leaders can act before reduced alertness reaches a critical barrier.

That practical discipline is consistent with the principle behind Safety Culture: From Theory to Practice. Culture becomes visible in the decisions that remain protected when the operation is under pressure, including the decision to stop extending a shift when the risk evidence no longer supports it.

For a broader view of sleep, shift design, and occupational exposure, compare this method with sleep-disorder patterns in shift workers. The two topics meet when a roster creates a predictable health burden that the safety system has not yet treated as an operational risk.