Inspection Process Best Practices

The quickest way to making the entire process ineffective is to fail to look at what is collected or act on what is found.

• By Cary Usrey
• Aug 01, 2012

At the start of the safety regulations for construction, OSHA calls for "frequent and regular inspections of the job sites, materials, and equipment to be made by competent persons designated by the employers." Unfortunately, OSHA does not go on to state how this should be done or what to do with the observations collected during this process.

As a result, many see the act of doing the inspection as a finite process with no post-action required. For example, I once asked a forklift operator to see his inspection records. The operator beamed and gladly handed me a stack of forklift inspections. During my review, I noticed that each inspection for the last few weeks had the same hazard noted: The brakes were requiring more effort to stop the vehicle. I asked the operator about this and he said that it had been reported, yet nobody had gotten back to him about it. But the job went on, day after day, with a clear hazard noted but not corrected.

I often refer to this problem mentality as "whack-a-mole" –- the carnival game where plastic moles pop up and the participants strike them hard with a hammer. Similarly, observers are going through the motions of collecting information, but the information is not tracked such that it can be used to perform trend analysis on the findings. This continues until something undesired occurs, such as an accident, and during the investigation you discover, "That's the way we've always done it!" This article will introduce a process where observation data can be used to both predict and prevent work site injuries by focusing on the leading indicators most likely to cause harm.

This process is called Plan-Do-Study-Act (PDSA), a continuous improvement cycle made popular by Dr. W. Edwards Deming, whom many consider to be the father of modern quality control. However, it is much more complex than simply applying this model to your existing process. You must have an organization supporting it with commitment from management to look at the data and provide necessary resources to quickly act on the findings. In addition, everyone involved must be properly trained so consistent and meaningful data is collected and acted upon.

Plan
The first step of the PDSA is to define a plan or purpose for the inspection process. Is the inspection process done simply to satisfy the obligation of a regulation? If so, you can quit reading now and continue doing what you were doing. However, if the objective is to identify weaknesses in your processes so you can make proactive changes likely to prevent incidents, then read on. Next, you should define expectations around the process. OSHA recommends these components in its Challenge program, which guides participants through a three-stage process to implement an effective system to prevent fatalities, injuries and illnesses.

Components of a documented system for routinely scheduled self-inspections:

• Tool or checklist
• Inspection schedule
• Training for members of the inspection team(s)
• Recording of findings
• Responsibility for abatement
• Tracking of identified hazards for timely correction

You also must develop a plan for using your data, not just collecting it. How will the findings be communicated and shared? Will they be corrected in a timely manner and tracked to completion? As trends develop, how will action plans be developed and implemented to prevent reoccurrence, as shown in the example of the forklift hazard findings?

Do
The next step of PDSA is to collect work site safety observations. It takes data to fuel your trend analysis, so keep in mind that the more employees involved, the more observations that can be collected, allowing more hazards to be identified and addressed. This should be done by all employees in the field, not just safety and health staff.

It is also necessary to address cultural bias to ensure you have a "just culture," as Dr. Sidney Dekker, professor of human factors and systems safety and director of research at the Lund University Center for Complexity and Systems Thinking, refers to it. He suggests it is essential that the observation reporting, especially significant hazards, be non-punitive and protected. In other words, don't shoot the messenger! This often occurs from a misperception that finding unsafe observations is a reflection of how well one does the job and then will reflect poorly on the observer or project team. It also can occur if a senior manager reacts poorly to the discoveries, such as reprimanding the observer or failing to act on the data collected.

Study
Most companies stop once the inspection is completed and the initial hazards are both discovered and fixed (the moles are whacked). This is a major error that will prevent meaningful improvements in the safety process. For example, an observer discovers poor housekeeping around a drywall contractor's work area on Monday. The observer takes note and asks the foreman of the drywall contractor to clean it up. An argument ensues -- whether cleaning up should take place as they go versus cleaning up at the end of the task, which could be hours or even days later. The observer insists and the foreman directs workers to clean up the mess. A day or two passes and the observer returns to check on the drywall contractor. Of course, he finds poor housekeeping again. He failed to address the causal factors associated with this issue: What was leading the contractor to do what he did? In this case, the contractor is paid to hang drywall, not clean it up. It is in his best interest to push productivity over safety. If he has gotten away with it in the past, then what would prevent him from continuing it in the future?

If you are not conducting trend analysis on the observations you collect, these scenarios could be happening quite often. Ideally, you should be looking at trends and leading metrics on a number of fronts, for instance:

• Are inspections taking place?
• What is the quality of the inspections submitted?
• What are the top hazards identified by hazard category?
• Which contractors were observed?
• Who are the most at-risk contractors?
• Which projects/locations were observed?
• What are the most at-risk projects?
• What recurring trends are developing?

Remember that in the study phase, the data will not make sense if you have not done your due diligence in the planning phase by developing clear expectations and effectively communicating them. You are looking for deviations to your plan (e.g., not following safety procedures or not conducting inspections at a regular frequency) so that you can take action and address gaps in your safety process. By tracking and trending this information, you can turn collected data into actionable information.

Act
The final and most important step in driving improvement is to provide necessary action and feedback on the process. Observations are the beginning, not the end, of the inspection process. In their book "Managing the Unexpected," Karl E. Weick and Kathleen M. Sutcliffe noted the following: "Error is pervasive. What is not pervasive are well-developed skills to detect and contain these errors in their early stages." The development of these skills is necessary to change how things are done.

Going back to the example of the drywall contractor, what can be done to break the cycle of poor housekeeping? It becomes readily apparent that an action or imposed consequence is necessary in order to change the outcome. This could be a meeting with senior managers, levying a fine for failure to act, or hiring a laborer to clean up and charging the contractor. Whatever change is done, the results of the action should be measured with observations going forward. If the housekeeping improves, the action was sufficient. If housekeeping remains poor, the action plan was inadequate and will require a different approach.

In addition to action plans to address deficiencies, both positive and negative feedback from management is needed to sustain the process successfully. Monitor the employees conducting inspections and let them know the quality of their observations and whether they are focusing on the right things. Provide the resources needed to close out at-risk findings quickly and effectively. Drive accountability to ensure the entire process is followed. The quickest way to making the entire process ineffective is to fail to look at what is collected or act on what is found.

Inputs such as behaviors and conditions (who, what), coupled with an expected process throughout (how), will yield predictable and measurable outputs (what). These leading indicators -– all easily obtained by performing worksite observations -– can tell you the true state of your safety program. It is up to you to determine the "why" in behaviors and take action accordingly to effect desired changes. By following the Plan-Do-Study-Act process, you can begin to predict where work site injuries are likely to occur and gain the confidence necessary to proactively prevent unintended consequences from an out-of-control process.

This article originally appeared in the August 2012 issue of Occupational Health & Safety.