Human Factors: The PHA Requirement Hiding in Plain Sight

Throughout my career, I have interviewed for numerous safety, EHS, and Process Safety Management positions. I have never once been asked about Human Factors in safety or PSM.

Of all the PHAs I have participated in, I have encountered only one existing PHA that included a meaningful Human Factors analysis. Even then, Human Factors was not initially included in the revalidation scope. I had to raise the issue while the project scope was being developed.

That is remarkable when you consider that Human Factors is not an industry best practice that companies can choose to adopt. It is an explicit PHA requirement.

Human Factors, like Facility Siting, appears under 29 CFR 1910.119(e)(3), but it rarely receives the attention its regulatory status warrants. Also like Facility Siting, OSHA does not define the term or provide much context within the standard itself. Employers are given a requirement without a clear explanation of its intended scope.

As a result, Human Factors is often reduced to a generic question at the end of a PHA checklist:

“Have Human Factors been considered?”

The team answers yes, points to training, procedures, and administrative controls and moves on.

That is not a Human Factors analysis.

What Are Human Factors?

Human Factors examines the relationship between people and the systems in which they work. It considers how equipment, controls, displays, alarms, procedures, staffing, schedules, communication, training, supervision, and the physical work environment influence a person’s ability to perform a task correctly.

In practical terms, Human Factors asks whether the work system has been designed around realistic human capabilities and limitations. It considers questions such as:

• Can the operator clearly identify the correct valve?

• Can employees reach the equipment they are expected to operate?

• Does the control system present the information needed to recognize an abnormal condition?

• Are alarms understandable, prioritized, and actionable?

• Does the operator have enough time to complete the required response?

• Can the procedure actually be followed under field conditions?

• Are employees working schedules that create fatigue?

• Is one person expected to monitor too many processes or complete too many simultaneous tasks?

• Can a technician perform the task while wearing the required PPE?

• Are labels, signs, displays, and controls consistent?

• Can an employee recognize and recover from an error before it causes a release?

These are not merely personnel-management questions. They are process-safety questions because human actions are often part of the system used to prevent, detect, control, or mitigate a catastrophic release.

One of my favorite examples of Human Factors in Process Safety is the nuisance alarm. Every plant has at least one. Operators grow so accustomed to acknowledging it that they eventually do so without even looking at the control panel. After weeks or months, the response becomes muscle memory.

Eventually, someone corrects the field condition. The process returns to normal, and the alarm is once again functioning as intended. But to the operators, it is still the same old nuisance alarm they have acknowledged hundreds of times. So, when it activates again, they do not stop to determine what has changed. They acknowledge it and move on.

Only this time, the alarm was real.

Congratulations! Your facility is on the evening news.

Human Factors Is Not the Same as Operator Error

One of the greatest obstacles to a meaningful Human Factors analysis is the phrase “operator error.”

A PHA team identifies a scenario in which an operator could open the wrong valve, skip a procedural step, enter an incorrect setpoint, or fail to respond to an alarm. The cause is recorded as operator error, training is listed as the safeguard, and the team considers the scenario complete.

But operator error is not a root cause. It is the beginning of the analysis.

If an employee can open the wrong valve, the team should ask why. Where is the gap in the system that allowed the potential for an employee to open the wrong valve?

Are two valves identical? Are they located beside each other? Are the labels difficult to read? Does the piping arrangement create confusion? Does the procedure use equipment numbers that do not match the field labels? Is the task performed infrequently? Is the area poorly lit? Is the employee wearing gloves that make the controls difficult to operate? Is there an independent verification step? Could the equipment be designed so the incorrect action is physically prevented?

Telling employees to “be more careful” does not correct any of those conditions.

People make mistakes. They become distracted, misread information, forget steps, misunderstand instructions, and make decisions based on incomplete or misleading information. A well-designed process anticipates these possibilities and prevents a single foreseeable error from becoming a catastrophic event.

Why Human Factors Gets Missed During PHAs

Traditional PHA methodologies tend to focus on equipment failures and process deviations.

The team is comfortable discussing a pump failure, blocked line, high temperature, loss of cooling, control-valve malfunction, or relief-device discharge. These scenarios fit naturally into HAZOP guidewords and process nodes.

Human performance is more difficult to represent. Employee fatigue, stress, mental and emotional well-being, physical ailments, operating environment, injuries, medications, process knowledge and experience are all Human Factors variables that are overlooked during the PHA.

The P&ID shows valves, instruments, piping, and equipment. It does not show workload, fatigue, confusing displays, poor communication, weak supervision, or an operator’s dependence on an outdated procedure.

Human Factors also gets missed because training and procedures are often accepted as universal safeguards. When a scenario involves a human action, the team may automatically record “trained operator” or “procedure in place” without evaluating whether those safeguards are reliable under the conditions of the scenario.

A procedure cannot compensate for every design weakness. Training cannot guarantee that a person will correctly interpret misleading information. Experience cannot create more response time when an upset develops faster than a person can react.

Another problem is that the right people may not be involved. A PHA team can understand the engineering design without fully understanding how work is actually performed. Operators, maintenance technicians, laboratory personnel, loaders, and contractors often know where confusion, workarounds, distractions, and error-likely situations exist.

Their participation is essential to a credible Human Factors review.

Human Factors Are Present in Every Part of the Process

Human Factors should not be treated as a separate issue that is discussed once at the end of the PHA. Human interaction should be considered throughout the analysis.

Control System Design

Control systems should help operators understand what the process is doing and where it is going.

The PHA should consider:

• Whether critical information appears on the same screen.

• Whether displays clearly distinguish normal and abnormal conditions.

• Whether trends are available to show developing problems.

• Whether control direction and equipment status are obvious.

• Whether important information is buried in multiple screens.

• Whether operators can identify failed or unreliable instruments.

• Whether the system provides feedback after an operator takes action.

An operator cannot make a good decision if the system provides incomplete, confusing, delayed, or inaccurate information.

Alarm Management

An alarm is only an effective safeguard if the operator notices it, understands it, and has enough time to respond.

The team should evaluate:

• How many alarms may activate during an upset.

• Whether alarms are prioritized.

• Whether alarm messages tell the operator what condition exists.

• Whether the expected response is clear.

• Whether nuisance alarms have conditioned operators to ignore them.

• Whether the response requires several actions in a limited period.

• Whether the operator is responsible for other processes at the same time.

Listing “alarm and operator response” as a safeguard is not enough. The PHA must consider whether that response is realistic.

Procedures

A procedure may be technically accurate and still be difficult to use.

Human Factors considerations include:

• Whether the procedure matches the actual equipment and control system.

• Whether steps are written in the correct sequence.

• Whether warnings appear before the hazardous step.

• Whether equipment identifiers match field labels.

• Whether the procedure clearly identifies decision points.

• Whether abnormal conditions and stop-work criteria are included.

• Whether employees must rely on memory while moving between locations.

• Whether the procedure is usable while wearing required PPE.

The best way to evaluate a procedure is to observe a qualified employee using it under actual or realistically simulated conditions.

Equipment Layout and Identification

Poor layout can create opportunities for errors even when the equipment is mechanically sound.

The PHA should consider whether:

• Similar valves or connections can be confused.

• Incompatible materials can be connected.

• Equipment labels are visible and durable.

• Controls operate in a consistent direction.

• Emergency isolation points are accessible.

• Indicators can be seen from the location where adjustments are made.

• Sampling and maintenance tasks require awkward positioning.

• Employees must stand in a hazardous location to operate equipment.

The design should make the correct action clear and the incorrect action difficult.

Workload and Staffing

Human reliability decreases when employees are overloaded, rushed, fatigued, or forced to divide their attention among competing priorities.

The PHA should evaluate:

• The number of processes assigned to each operator.

• The number of simultaneous tasks during startup, shutdown, or upset conditions.

• Whether adequate staffing is available for abnormal operations.

• Overtime and extended work schedules.

• The quality of shift turnover.

• The availability of supervision and technical support.

• Whether production demands conflict with safe operating decisions.

Normal staffing may be adequate during steady-state operation but completely inadequate during a process upset.

Maintenance and Nonroutine Work

Human Factors applies to maintenance personnel as much as it applies to operators.

Maintenance tasks may involve:

• Breaking containment.

• Removing instruments from service.

• Installing blinds.

• Bypassing alarms or interlocks.

• Reassembling equipment.

• Restoring valves to their proper positions.

• Coordinating work across operations, maintenance, and contractors.

The PHA should evaluate how the facility prevents incorrect isolation, improper reassembly, premature return to service, and failures in communication.

What Major Incidents Teach Us About Human Factors

The 2005 BP Texas City refinery explosion provides a powerful example of why Human Factors cannot be reduced to operator error.

The Chemical Safety Board’s investigation found that operators made errors during the startup of the isomerization unit. However, those errors occurred within a system containing misleading instrumentation, malfunctioning alarms, inadequate control displays, abnormal startup practices, reduced staffing, insufficient training, and severe fatigue.

Operators had reportedly worked 12-hour shifts for 29 or more consecutive days. A level transmitter indicated that the tower level was falling when it was actually rising. Important flow information was not presented together on the control display. The unit lacked instrumentation that could have helped operators understand the actual condition inside the tower.

Calling the event operator error would have concealed the conditions that made the errors more likely and made recovery more difficult.

Similar lessons appear in other CSB investigations. The 2016 MGPI chemical release occurred when sulfuric acid was unloaded into a sodium hypochlorite tank, producing chlorine gas. The investigation emphasized the need to evaluate how drivers and facility operators interacted with unloading equipment and to design transfer systems that reduce the likelihood of an incorrect connection.

The recurring lesson is straightforward: if one foreseeable human action can cause a catastrophic release, the system needs stronger protection than an expectation that nobody will ever make a mistake.

How Human Factors Should Be Addressed in a PHA

A separate Human Factors checklist can be useful, but it should supplement the PHA rather than replace Human Factors analysis within each node.

Whenever a person initiates, controls, detects, prevents, or responds to a scenario, the team should ask:

• What information tells the person that action is required?

• Is that information clear, reliable, and available at the right time?

• What specific action must be taken?

• How much time is available?

• Can the action be completed from a safe location?

• Is the person likely to be performing other tasks?

• Does the procedure accurately describe the required response?

• Has the employee practiced the response?

• What happens if the action is delayed, omitted, or performed incorrectly?

• Is there an independent layer of protection if the human action fails?

The answers should be documented. If the team cannot explain why a human-dependent safeguard is reliable, it should not receive unquestioned credit.

Better Human Factors Recommendations

Weak Human Factors recommendations usually focus on the employee:

• Retrain the operator.

• Remind employees to follow the procedure.

• Discuss the event at the next safety meeting.

• Add another warning to the procedure.

These actions may have value, but they often leave the original error-likely condition unchanged.

Stronger recommendations address the system:

• Install uniquely keyed or dedicated connections for incompatible materials.

• Automate a time-critical shutdown action.

• Add an independent high-level shutdown.

• Rationalize alarms and eliminate nuisance alarms.

• Relocate critical information to the primary operator display.

• Improve labels and equipment identification.

• Redesign controls to provide clear feedback.

• Simplify the task or procedure.

• Improve lighting or equipment access.

• Require independent verification for critical lineups.

• Adjust staffing for startups, shutdowns, and other high-workload operations.

• Establish fatigue controls for extended work schedules.

The objective is not to remove people from every process. It is to design the process so that people have the information, time, tools, environment, and support necessary to succeed.

Human Factors Must Be Managed Between PHAs

A five-year PHA revalidation is not enough to keep a Human Factors analysis current.

Human performance can be affected by changes to:

• Staffing levels.

• Shift schedules.

• Control-system graphics.

• Alarm configurations.

• Operating procedures.

• Equipment layout.

• Job responsibilities.

• Training programs.

• PPE requirements.

• Contractor use.

• Production rates.

• Organizational structure.

The facility’s management-of-change and organizational change processes should consider whether a proposed change creates new workload, communication, interface, or error-recovery concerns.

Incident investigations should also look beyond what an employee did. They should examine why the action made sense at the time, what information was available, what competing demands existed, and what system changes could prevent recurrence.

The Question Every PHA Team Should Ask

Human Factors is not about proving that employees are capable of making mistakes. That has already been established.

The real question is: “Has the process been designed so that a foreseeable human mistake will not become a catastrophic event?”

That question changes the focus from blaming individuals to improving the work system.

After years of safety and PSM interviews, I have never been asked how I evaluate Human Factors. After participating in numerous PHAs, I have seen the subject meaningfully addressed only once in the initial PHA to be revalidated.

That should concern us.

Human Factors has been part of OSHA’s PHA requirements since the PSM standard was issued. It deserves more than a checkbox, a reference to training, or a recommendation telling someone to be more careful. It deserves the same structured analysis given to equipment failures, process deviations, and failed safeguards.

People are part of every process, and they are the most unreliable part, not because they aren’t perfect, but because our systems are not perfect. A PHA that does not seriously evaluate how people interact with that process is not evaluating the entire system.

Appalachian EHS can help facilities strengthen PHA scopes, evaluate human-dependent safeguards, identify error-likely conditions, and incorporate meaningful Human Factors analysis into PSM programs.