Safety science is not good at tracking small changes and noticing when things are slipping inexorably but slowly. When something bad happens, it is sometimes because the reality on the ground no longer matches the audit instrument.
The “Swiss cheese” model would be more realistic if the cheese slices were melting, the holes were getting bigger, and the holes were starting to line up but nobody noticed. I think safety audits and evaluations need to incorporate a random component in addition to the systematic survey. Unpredictability and random snooping can uncover a completely unexpected deviation that would be missed if the safety professional is fixated on checking a box on the checklist. This component has to be truly random. If you don’t know what you don’t know, you won’t find it looking in the place you expect to see it.
Tee L. Guidotti is Vice President for HSE/Sustainability at Medical Advisory Services, Rockville, Maryland. His website is www.teeguidotti.com.