“The only real mistake is the one from which we learn nothing.”  — Henry Ford

I know a young woman whose relationships are in a constant state of turmoil. She keeps picking the wrong guys and each relationship ends in heartbreak. “Why do I keep making the same mistake over and over?”

Why indeed? I once cautioned that the only thing more suspect than fashion advice from an engineer is relationship advice from an engineer, so I won’t even pretend to offer relationship advice.  However, I am willing to state, without hesitation, that we know how to be safe by studying the mistakes of the past, and that our only hope for improving safety is to learn from those mistakes and then to apply those lessons.

So, it should make perfect sense that one of OSHA’s requirements for process hazard analysis is that “[t]he process hazard analysis shall address [t]he identification of any previous incident which had a likely potential for catastrophic consequences in the workplace.” [ From the Process Safety Management (PSM) standard, 29 CFR 1910.119(e)(3)(ii) ]

What Previous Incidents?

The PSM standard doesn’t give much advice on what previous incidents are supposed to be addressed. The PSM element on incident investigation, in 29 CFR 1910.119(m)(1) requires that we “investigate each incident which resulted in, or could reasonably have resulted in, a catastrophic release of highly hazardous chemical in the workplace.”  Catastrophic releases of highly hazardous chemicals are those that “may result in toxic, fire or explosion hazards.”

So, an incident where someone slipped on an icy sidewalk and was knocked unconscious, while recordable, is not the type of previous incident the PSM standard is telling us to address.

How Far Back?

Again, we can turn to the PSM element on incident investigation.  In 29 CFR 1910.119(m)(7), the PSM standard requires that “[i]ncident investigation reports shall be retained for five years.” It is no coincidence that the requirement for updating and revalidating a process hazard analysis is “at least every five (5) years”. [ 29 CFR 1910.119(e)(6) ]. OSHA wants the incidents that have occurred since the last PHA addressed in the current PHA. Presumably, the incidents before that were addressed in the previous PHAs.  If they weren’t, then there is no time like the present.

Just the Incidents from the Process Being Reviewed?

At a minimum, a PHA of a process should address all incidents in that process since the last PHA that required an investigation in accordance with the PSM element on incident investigation.

The purpose of the PSM standard and the PHA, however, “is to prevent unwanted releases of hazardous chemicals especially into locations which could expose employees and others to serious hazards.” If there are lessons to learn from elsewhere in the facility, or for that matter, from elsewhere in the industry, they should be incorporated as well. Groucho Marx said, “Learn from the mistakes of others. You can never live long enough to make them all yourself.”

How to ‘Address Previous Incidents’?

The point of addressing previous incidents during a PHA is not to relitigate the conclusions and recommendations of the incident investigation that followed the incident.  No, the point is to make sure that the PHA team is aware of the incident so that similar hazards don’t go unrecognized during the PHA. For that to work, the team must be aware of the previous incidents prior to reviewing segments or nodes of the process with similar hazards.

Ideally, then, the PHA will begin with a review of previous incidents that are relevant. In addition, PHA team members will share their experience and knowledge with the team when a particular hazard is encountered.

Documenting that Previous Incidents were Addressed

As with any effort that involves compliance with a regulation, it is not enough to do the thing. The regulation also implicitly requires documentation as evidence of compliance, and in the specific case of the PSM standard, evidence that will withstand the scrutiny of the 3-year compliance audit required in 29 CFR 1910.119(o).

OSHA gives no indication of what it expects in terms of documenting that previous incidents were addressed. Not in the standard itself, nor in the nonmandatory Appendix C, Compliance Guidelines and Recommendations for Process Safety Management, nor in any of the hundreds of standard letters of interpretation that OSHA has issued. The closest thing to guidance comes from the standard letter of interpretation to Baruch A. Fellner, dated June 22, 1998, where OSHA says that there should be “a review of all incident investigation reports required by 1910.119(m) to assure all affected PSI, procedures, training, etc., have been updated to reflect recommendations set forth in the subject reports and that appropriate information from those reports have been incorporated into the PHAs.

Typically, we have taken this to mean that if there is a formal incident report log at the facility, it should be incorporated by reference in the PHA report.  If there is not a formal incident report log, or the log is too broad, then a list of relevant incidents should be compiled for inclusion in the PHA report.

Finally, in the PHA worksheets themselves, any reference to previous incidents—either in the identification of causes or in the description of consequences—should be explicitly called out. The PHA report should then note that this was done.

Remembering the Past

Addressing previous incidents is not just a requirement of the PSM standard in the conduct of PHAs. It is essential to improving safety. George Santayana, famously said “Those who cannot remember the past are condemned to repeat it.” Addressing previous incidents is a way of remembering the past, so that we don’t repeat it. It needn’t be complicated, but it needs to be done. Any of us can survive the heartbreak of doomed relationships, but the heartbreak of catastrophic releases is much harder to survive.


  • Mike Schmidt

    With a career in the CPI that began in 1977 with Union Carbide, Mike was profoundly impacted by the 1984 tragedy in Bhopal and has been working on process safety ever since.