Not too many months ago an elevator at a construction site in Stockholm crashed to the ground, killing five people. It turns out big parts of what was supposed to keep the elevator together were missing and the company doing to the inspections(who had cleared the elevator) were not allowed to make other inspections for that company.
This brings up an interesting point, that humans are prone to identify things that stay the same. This happens when people get hit by a train, they passed the railroad a millions times before and there was no train so they started assuming that there was never a train on the railroad. Until there was a train and they got hit by it. There was a scandal here in Gothenburg a few years ago where the local hospital had cleared peoples’ tissue samples as being non-cancerous and it turned out that it was cancer.
So whether you never have to recognize the issue(in the case of the train) or just rarely have to recognize the issue it’s easy to not do a proper inspection. I therefore argue that you should introduce known errors for inspectors to catch to keep them on their toes and to normalize that issues are reported. This doesn’t always work, for instance the FIU bridge collapse probably would still have happened as everyone was wholly set on everyone being on track so many issues were simply disregarded. Similarly Chernobyl could have been prevented by even the most basic willingness to learn from the many warnings that preceeded it.
But it’s still an improvement that deals with the inherent weakness of our tendency to assume that “it hasn’t happened yet” means “it’s never going to happen” and our tendency not to check or report things properly. For the likes of elevators you can introduce non-dangerous faults before inspection but this make inspectors only pay attention to aspects that are formally required but not actually dangerous. I think it’s more appropriate to inform everyone on-site that the elevator is not usably since it has not been inspected, that the elevator is clearly marked as “not in use”, which (possibly dangerous) errors will be introduced and when inspection will happen.
This is of course slightly dangerous itself, we can’t rule out someone using an elevator clearly marked as “not in use” and injuring themselves. But this way all inspections ought to be much more robust so even if we people do end up getting injured once in a great while we should have improved safety quite a lot. This is harder for some areas, we can even stay in the construction industry. How do you introduce serious errors in the concrete construction of a floor in a building? You’re pretty much going to have to rebuild things afterwards, which is very expensive. For most sectors however we should be able to find suitable tests to improve the quality and rigor of inspections.
In the case of the tissue samples at our main hospital I think known good and known bad samples should be part of the sum total samples everyone goes through. If any sample is given a conclusion that differs from that which has been recorded it should trigger a warning and the pathologist needs to explain his methodology to the other pathologists. Perhaps the sample was incorrectly labelled from the beginning! Either way people stay on their toes and thinks “how would I explain my conclusion in this case”.
Safety culture
An interesting subject is an organization’s dedication to safety. DuPont has been a trail-blazer in this regard(not without exception) and I think BP is on the other side of that scale. Only an organization that actually wants to achieve high safety even if it costs some money(possibly saving some as well) is going to be helped by introducing errors for inspectors to find. I mentioned Chernobyl before as an example where people could easily find problems but no one wanted to find any so it was only when disaster struck and the entire world noticed that errors became widely known. Trying to solve safety issues stemming from a disinterest is safety by adding more bureaucracy is a fool’s errand. We see this in cases like Williams Olefins’ issues and the aforementioned BP explosion at Texas City refinery, where people check boxes and skip things that are trying to stop them from doing something dangerous. Even the most basic steps towards finding errors are skipped so why would they change how inspections are done to find more errors in the first place?
For this reason accidents like Chernobyl, the FIU bridge collapse and two trains colliding are very useful to us. You can easily create an organization geared towards group-think, ignoring issues and warning signs and using safety overrides to improve efficiency but accidents show us the reality of what we’re dealing with. In the case of an organization that has a safety culture these accidents show us what we’re missing and in other cases it demonstrates the lack of safety culture.
Cost
I really like watching USCSB videos about their investigations but they are entirely focused on safety and they don’t seem to balance that with costs. If all organizations in the US that is in some way involved in processing really had a safety culture and really implemented Process Safety Management, Management of Change and Process Hazard Analysis, how much process industry would they have? If the answer is 90% then I think we should go with what the USCSB is saying. If it’s 50% then maybe not. I don’t see any indication that the USCSB ask these questions and I doubt these questions are within their remit.
But they are necessary for people to ask. We don’t actually improve safety by imposing PSM on US companies that then outsource all that stuff to South East Asia where safety isn’t even attempted. That moves the problem – admittedly – but it actually makes the problem worse. When an organization ignores safety – more or less – we really need to ask ourselves if maybe they know something we don’t. That’s not always going to be true but we can’t always assume that PSM and Safety Culture is some blanket solution that can be applied everywhere to get rid of accidents. Sometimes you just move the accidents.
This is a hard argument to make so I assume BP used to write something like “Safety is of the utmost importance in everything we do” but let every facility work that out themselves. At no point was process safety something that BP demanded of facilities but budget cuts were something they followed up closely. So we can’t take some random statement at face value, even those who are disinterested in safety will claim to be all about safety. We have to look at what they actually do and then we need to understand why. If they are ignoring process safety to make a profit of 5% become a profit of 8% then that’s a great situation to nail people to the wall. If they ignore safety because otherwise their work is done by people in Indonesia instead, then we might want to let things be.
Either way, this is a decision that a democracy needs to make. We can’t afford saying “Safety first” like it’s dogmatic because we can actually makes things less safe in our pursuit of making things more safe. We can’t afford to lose some industry and we shouldn’t pretend like that isn’t on the table. These are the big issues that politicians aren’t inclined to touch because it’s basically a set of bad choices we have. “More safety to save money” is a subset of all safety and not the totality of it. For most examples safety is about making hard choices.