When driving down the road, we glance in the rearview mirror to see what’s happening behind us — but the majority of the time we concentrate on what’s in front of us. However, when it comes to process safety metrics, we still focus on the lag metrics, what happened yesterday, rather than the lead metrics, information that can help us manage our future. This is like driving while mostly looking in the rearview mirror. At some point, our not observing information in front of us may lead to an accident.
In the early 2000s, the Center for Chemical Process Safety (CCPS)  and the American Petroleum Institute  worked to achieve some commonality in process safety metrics. This led to clearly defined lagging metrics, which we now call Tier 1 and Tier 2 losses of containment. The two groups then developed some options for leading metrics, known as Tier 3 and Tier 4. Additionally in 2010, CCPS published the results of a survey of leading indicators used by industry .
Following on from this work, the members of the IChemE Safety Centre (ISC) launched a project to fully define a suite of leading process safety metrics. Defining the metrics in detail would enable benchmarking on the leading side of safety, not just the lagging side. This project culminated in a published guidance document titled “Lead Process Safety Metrics — Selecting, Tracking and Learning” . (Get a free copy.) A number of companies contributed to this document. I particularly would like to acknowledge: Quadrant Energy, EnVizTec, MMI Engineering, Orica, Simon Casey Risk and Safety Consultant, Todd Corporation, and Woodside.
Several prominent process safety incidents, such as the 2005 explosion at BP’s Texas City, Texas, refinery, have underscored the need for operating companies to track and understand process safety metrics just as they do occupational safety ones. Indeed, the tracking of process safety metrics is vital to help us understand the state of our facilities and systems — and to alert us to impending issues. Importantly, while lagging process safety metrics enable us to compare current to historical data to monitor improvement, they won’t necessarily predict future loss-of-control events. In contrast, leading metrics are proactive and afford the opportunity to manage potential safety issues. However, lead metrics of course aren’t an absolute predictor of process safety.
The ISC project started with reviewing the lead metrics reported by each member company to look for commonality. Then, the team winnowed these metrics based on their apparent value for driving decisions or actions in the organization. Recording and analyzing data takes time and resources. If the data analysis doesn’t result in useful decisions or actions, it’s a wasted effort. The team next undertook a second screening based on “ease of collection” to further identify the most important metrics and reduce the list to a manageable number.
A Dual Approach
Lead metrics fall into two types. The first identifies positive situations, such as work being completed to schedule. This is related to the concept of “Safety-II”  where the focus is on understanding how a thing goes right in a quest to replicate it, as opposed to understanding how a thing goes wrong in a quest to prevent reoccurrence. The premise is that things go right far more often, so there’s a greater source of information to learn from. The focus for this metric is to trend toward 100% compliance. However, achieving 100% isn’t reason to relax but instead cause to explore whether the metric is accurate and not misleading about the health of the system.
The second type of lead metric is akin to holes forming in the so-called Swiss cheese barriers . This metric measures failure of a barrier — so for it to be considered a lead metric, other barriers must have prevented the consequence from occurring. It can be thought of as a measure of the barrier’s weakness. The focus of this metric is to drive the failure toward zero. However, a zero or very low result should prompt analysis to ensure the metric is accurate.