What last year’s numbers cannot tell you

Most safety reports measure the past with great precision. The trouble is that the past is not the thing you are trying to change.

There are two families of measure, and it helps to keep them apart. Lagging indicators count harm that has already happened: injuries, days lost, claims paid. They are accurate, they matter, and they are entirely in the rear-view mirror. Leading indicators try to point forwards: near misses reported, inspections completed, corrective actions closed on time, training done before the job rather than after the incident.

A quiet year is not the same as a safe one

Here is the problem with steering by lagging numbers alone. A year with no serious injuries can mean you are genuinely safe. It can also mean you were lucky and were not looking closely. From the injury count by itself, you cannot tell which.

This is not a hypothetical worry. Regulators have found, more than once, that a workplace with an excellent personal-injury record was sitting on serious underlying risk that the headline numbers simply did not show. People were not slipping and spraining, so everyone relaxed, while the larger hazard built quietly in the background. Both the UK Health and Safety Executive and the OECD now urge organisations to watch leading indicators precisely because the lagging ones can flatter you right up until the day they do not.

Lagging indicators tell you how last year went. Leading indicators hint at how next year might. Only one of those is still changeable.

The measure that becomes a target

There is a second trap, and it has a name. Goodhart’s law says that when a measure becomes a target, it stops being a good measure. Lean too hard on the injury figure and you may not reduce injuries. You may simply reduce the reporting of them, which is worse, because now you have less harm on paper and the same risk in the building.

This is the quiet case for treating a rising number of reports as a good sign rather than a bad one. If people are logging more near misses, more observations, more small concerns, the most likely explanation is not that the place got more dangerous. It is that you can finally see it.

A small set you actually act on

Leading indicators are messier than lagging ones, and the temptation is to drown in them. Resist it. Pick a few you will genuinely act on: are actions being closed on time, are inspections happening, is reporting holding up or quietly falling away. Watch the trend more than the number. A figure that only ever gets reported, and never changes a decision, is not an indicator. It is wallpaper.

References

1. Health and Safety Executive, Developing process safety indicators (HSG254). hse.gov.uk.
2. OECD, Guidance on Safety Performance Indicators (second edition, 2008). oecd.org.

Shannon Hamming writes about how the working world keeps itself safe and honest: incident reporting, quality, ESG, compliance and the sensible use of technology in safety-critical work. Useful first, interesting always.