The way I look at it is I start with Reason’s Swiss Cheese model. The H&S system consists of slices of cheese. There are ‘holes’ in the cheese and when the ‘holes’ line up you get an accident/incident. Lagging indicators shown up where failures in the system occur ie accident/ incident stats and the like, so you are looking at the ‘holes’. For leading indicators you are looking at the quality of the ‘cheese’ around the holes so you are looking things like; training completed, audited carried out and H&S actions closed out. The only problem with leading indicators is that you must make sure that you are measuring ‘performance’ ie what you are doing to improve H&S not simple ‘activity’; doing stuff just fill in the KPIs. I have had arguments with people who want to simply use the number of risk assessments produced. It is possible to fill files with risk assessments for opening and shutting windows and carrying cups of tea but this does not contribute much to improving the organisation’s H&S performance.

Recent (brief) discussion on this topic: http://forum.iosh.co.uk/...spx?g=posts&t=109592