Tuesday, March 15, 2016

RMP NPRM: Accident Investigations Overview

This is part of a continuing series of blog posts about the EPA’s recently published notice of proposed rulemaking (NPRM) for revisions of their Risk Management Program. Earlier posts in this series include:

Accident Investigation Overview

The preamble to the NPRM describes current requirements for accident investigations in the RMP program. This includes both investigations specifically required for ‘catastrophic release’ incidents and to support process hazard analysis (PHA) program requirements for addressing previous incidents.

The NPRM is proposing the following changes that would affect the requirements for accident investigations:

Modifying the definition of catastrophic release in §68.3 to be identical to the description of accidental releases required to be reported under the accident history reporting requirements in §68.42;
Adding a definition for ‘root cause’ to §68.3;
Requiring a root cause analysis during incident investigation requirements under §68.60 (Program 2) and §68.81 (Program 3) to ensure that facilities determine the underlying causes of an incident to reduce or eliminate the potential for additional accidents resulting from deficiencies of the same process safety management system;
Requiring the owner or operator to address findings from all incident investigations required under §§ 68.60 and 68.81, in their hazard review {§68.50(a)(2)} or PHA {§68.67(c)(2)} respectively;
Clarifying that incident investigations are required even if the process involving the regulated substance is destroyed or decommissioned following or as the result of an incident;
Requiring that facility owners or operators complete an incident investigation report within 12 months of an incident that resulted in, or could reasonably have resulted in, a catastrophic release;
Requiring the inclusion in the RMP accident history under §68.42 of information on root causes analyzed as part of an incident investigation;
Modifying the on-line reporting system for RMPs (RMP*eSubmit) to incorporate an appropriate list of root cause categories for RMP facility incident investigations of RMP reportable accidents; and
Requiring in §68.195(a)(2) that the root cause categories be submitted in the RMP within 12 months of the release;

EPA Feedback

At the end of each of the discussions about the proposed changes described above, the EPA has also listed a number of questions upon which it would like specific feedback. Sometimes these questions refer to potential future rulemakings, but mostly they deal with alternative solutions to the problems discussed that the EPA has been considering. Comments about those alternative solutions are important in deciding how the EPA will move forward with the final rule.

Root Cause Analysis

The term ‘root cause’ would be defined in §68.3 as: “a fundamental, underlying, system-related reason why an incident occurred that identifies a correctable failure(s) in management systems”.

The changes being made to §68.60 (Program 2) and §68.81 (Program 3) in regards to root cause analysis are identical. In paragraph (d)(7) in each section the requirements for an incident report, including a root cause analysis are specified as:

“The factors that contributed to the incident including the initiating event, direct and indirect contributing factors, and root causes. Root causes shall be determined by conducting an analysis for each incident using a recognized method;”

The EPA is not requiring a specific root cause analysis method, but the only method specifically mentioned in the preamble is the “Guidelines for Investigating Chemical Process Incidents” from the Center for Chemical Process Safety (CCPS). The preamble does note that OSHA is working on developing a fact sheet on existing resources that explain how to conduct root cause analyses.


Over the last twenty plus years in the chemical process industry I have taken part in a number of accident and near miss investigations. In the early years of my career those investigations concentrated on the proximate cause of the incident and how to fix that specific problem. As the industry has matured there has been more of an emphasis in most major chemical companies at looking deeper into the causes of accidents to identify the root cause.

Proximate causes of accidents are the easiest to identify and correct, but they only serve to protect against that specific version of the problem. Identifying root causes and correcting those takes much more time and resources to complete the identification process. Fixing those root causes is also a much more difficult task to complete, but success helps to prevent whole classes of future incidents not just the re-occurrence of the incident under investigation.

The only drawback that I can see of requiring root cause analysis for all required accident investigations is that a properly conducted analysis almost requires someone who has been specifically trained in root cause analysis. Too often I have seen inexperienced investigation leaders guide an investigation team to a pre-determined cause that did not get to the ultimate systemic problem that was the true root cause of the issue. Still, even those poorly conducted root cause analysis investigations did a better job of preventing future accidents than did the old-style proximate cause investigations that may still be too prevalent in many small to medium sized chemical manufacturing facilities.

No comments:

/* Use this with templates/template-twocol.html */