Reducing the Consequences of Human Error
April 22, 2015 - Kestrel Management
Unsafe employee behavior—from entry-level operators, plant management, and corporate personnel to contractors, suppliers, and consultants—represents a significant source of risk to operations.
Companies use a variety of controls to manage risk associated with unsafe employee behavior. These may include:
- Leadership and policies that establish and communicate goals and objectives
- Documented methods for risk identification, assessment, mitigation, and communication
- Planning and budgeting practices that are aligned with financial and risk management objectives
- Documented operating procedures and practices
- Sound employment practices for hiring qualified personnel with appropriate skills
- Effective training programs to enhance employee skills
- Auditing and verification of performance against regulations and desired performance standards
- Documented emergency response procedures to minimize hazards
Despite best efforts, however, unintended events can occur when there is a failure in one or more of these controls. The challenge is to know where to focus resources to reduce the frequency and minimize the consequences of human error. The answer lies in the relationship between human factors and error.
Accident Causal Theory
The majority of unintended events (including process safety, occupational safety, environmental, and quality) are associated with unsafe employee behavior. Unintended events are generally not the result of any single event, cause, or individual action. Rather, they are the end result of numerous contributing factors that lead to unsafe employee behavior.
In his model of accident causation (often referred to as the “Swiss cheese” model), James Reason describes four layers of human failure, with each contributing to the unintended event (Reason, 1990). Starting with the accident itself, the first layer is described as the Unsafe Act. The Unsafe Act includes the actions of the person/crew involved in the event. An example of an Unsafe Act would be failure to close a valve that led to a product release. Following the Unsafe Act layer are three layers of latent failure, including Preconditions, Supervisory Factors, and Organizational Influences. Latent failures are conditions that may have been present for some time prior to the Unsafe Act associated with the accident, but were undetected until the accident occurred.
According to Reason’s model, accidents occur when there is a failure in one or more of these protective layers (“a hole in the cheese”). This model provides a powerful visualization of how organizations try to prevent accidents and how, despite diligent efforts, they still occur. It does not, however, identify what the failures are. In other words, where are the holes and what is causing them.
Human Factors Analysis and Classification System (HFACS)
The Human Factors Analysis and Classification System (HFACS) provides additional detail to Reason’s model and helps to identify specific human factors that contribute to accidents (Shappell and Wiegmann, 2000). Similar to Reason’s model, HFACS is organized into layers, as described below.
Human Factors Analysis and Classification System (HFACS) Structure
Building on the previous example where a valve was left open, HFACS helps to determine why. Did the employee make an error or was it the result of a rule violation? It is possible that the employee followed a line-up checklist that was no longer valid. The end result would be a valve being left open and product being released. HFACS would consider this to be an “error”—the employee intended to do the correct thing but did not achieve the intended outcome. Conversely, this would be a “violation” if the employee was aware of the checklist and chose to not use it.
Analyzing only the Unsafe Act identifies a mere fraction of the factors leading to the event. It’s critical that latent failures are also considered:
- Preconditions for the Unsafe Act answers the question, “Why did the employee make an error or violate established protocol”?
- Supervisory Factors helps to answer whether supervision contributed to or failed to correct the Precondition.
- Organizational Influences assesses how the organizational processes and/or environment contributed to accident occurrence.
While HFACS helps to identify specific human factors associated with human error within operations, it still has two major drawbacks:
- It is not linked directly to specific procedures, work instructions, SOPs, etc. As a result, it does not dictate where to focus resources to make desired improvements.
- The human factors categories are relatively broad in HFACS, making it difficult to apply HFACS to incident investigation on a practical level.
Human Performance Reliability (HPR)
The Human Performance Reliability (HPR) process is intended to solve these drawbacks by helping companies to:
- Determine where to focus improvement efforts by identifying which controls need to be adjusted or developed.
- Characterize what needs improvement to reduce human error by identifying the factors within the controls that are contributing to the unintended events.
- Inventory – The first step is to inventory and document controls currently in place at the company.
- Review – HPR reviews should be conducted monthly using the Human Factors Integration Tool (HFITTM). The HFIT™ software includes a comprehensive set of questions based on the HFACS framework. Based on the responses to these questions, HFIT™ provides a report of behavioral and contributing factors, and identifies which controls are associated with the incident.
- Analyze – Data obtained from each review are aggregated and used to identify large-scale trends regarding the human factors contributing to incidents across facilities, business units, or even the entire organization. The analysis helps to prioritize corrective action based on which human factors are contributing to incidents most frequently.
- Improve – The HPR process helps identify where improvements are needed, follow through on corrective actions, and adjust/develop the controls associated with error. Improvement initiatives ensure that the human factors contributing to errors are addressed.
Immediate and Systemic Opportunities
The HPR process helps companies to:
- Integrate human factor analyses into their incident investigation processes
- Identify the underlying factors contributing to unintended events
- Link identified human behavior to specific operational controls that are missing or need improvement to reduce the likelihood of incidents
- Reveal immediate local and systemic improvement actions
- Identify system-wide opportunities for operational enhancements, improved reliability, and the prevention of unintended events
Submitted by: A.W. Armstrong