Why we need to consider employee sentiment in conjunction with work design signals when assessing psychosocial risks at work.
On 27 July, two aircrafts came close to colliding at Sydney Airport. The investigation is still under way, but reporting from Reuters on 4 September revealed something that deserves attention far beyond the aviation sector.
Five of the ten air traffic controllers on duty that day had been classified as high-fatigue risks. Two of the three controllers directly involved were in that group.
The rating did not prove that anyone was fatigued. It was based on workload and scheduling conditions that increase the likelihood of fatigue, including extended shifts and overtime.
That distinction matters.
It would be wrong to claim that fatigue alone caused the incident. However, this is a pertinent example for employers to take note of.
Too often, organisations treat psychosocial risk as something that becomes visible when a worker reports stress, completes a survey, takes leave or makes a complaint. Those are important sources of evidence. But they are not the only factors you should be looking at.
Sometimes the work itself is already telling us that conditions are changing.
Understanding the patterns of risk
Safe Work Australia defines high job demands as a psychosocial hazard when demanding work is severe, prolonged or frequent. Its guidance asks employers to consider how long, how often and how severely workers are exposed, as well as the design of work and the systems used to organise it.
Controls must also be monitored and reviewed to see whether they are actually working.
This creates a practical measurement problem.
A survey can tell us how people experience work. It cannot always tell us, with enough detail, what changed in the work that produced that experience.
For example, a worker may report feeling overloaded, but the organisation still needs to understand whether the cause is sustained overtime, too many simultaneous demands, inadequate staffing, poor recovery time, constant interruptions, excessive meetings or some combination of these.
The reverse is also true. Operational data cannot tell us how a person feels. A calendar full of meetings is not a diagnosis. Working late does not mean someone is burnt out. A heavy workload does not prove psychological harm.
But those patterns can still be evidence.
The useful distinction is between worker evidence and work evidence.
Worker evidence includes surveys, consultation, interviews, complaints and reported experience. It tells us how work is being experienced by the people doing it.
Work evidence includes observable conditions: overtime, shift length, staffing levels, recovery opportunities, work intensity, scheduling pressure, meeting load, after-hours activity or repeated periods of unusually high demand. It tells us what is changing in the system around the worker.
Neither is enough on its own.
The mistake is to ask one type of evidence to do the job of the other.
Look for conditions that create exposure
This is why a recent AHRI article on task observation is timely. It argues that observing how work is actually performed can provide a more reliable view of work intensity than relying only on surveys.
The broader lesson is that psychosocial risk management should not start and finish with asking people how they feel. It also requires looking closely at the conditions that create the exposure.
For HR and safety leaders, this changes the questions worth asking.
Instead of only asking, “Are people stressed?”, ask: “What has changed in the way work is organised?”
Instead of only asking, “Did our survey score improve?”, ask: “Did the exposure we were trying to control actually decrease?”
If the issue was excessive workload, did overtime fall? Did staffing improve? Did periods of sustained demand become shorter? Did people get more recovery time?
If the problem was meeting overload, did meeting hours and back-to-back scheduling change? If the concern was after-hours work, did it reduce after the intervention?
These are not replacements for worker consultation. They make consultation better because the organisation has something concrete to investigate with employees rather than starting from a vague average score.
The Sydney Airport case is an extreme example because the consequences of failure can be immediate and visible. Most workplaces do not operate with that level of safety sensitivity. But the management problem is familiar.
Looking for early warning signals
Risk rarely appears from nowhere.
Before an absence trend, grievance, burnout case or serious incident, there may already be small changes in workload, staffing, recovery time, coordination and working hours. None of those signals proves harm. However, together, over time, they may be enough to justify a closer look.
That is the point of early warning.
You’re not looking to diagnose the worker before they speak or to replace surveys with surveillance. It’s about noticing when the conditions of work are moving in the wrong direction, while there is still time to ask why and do something about it.
🧰 HR’s career resource toolkit
- Course: Explore AHRI’s Psychosocial Code of Practice course.
- Report: Managing and Minimising Psychosocual Risks in Australian Workplaces (October, 2025, AHRI).
- Podcast: Redesigning work for the age of AI with Dr Sandra Parker
Disclosure: Sten Kreisberg works on SignalTrue, a workplace analytics product focused on observable work patterns. This article is written as an editorial contribution and does not claim that work-pattern data can diagnose individual psychological states.
