For thousands of years, doctors treated virtually every skin ailment by ‘letting’ or draining the blood of the patient. Leeches are really good at doing this as they quite literally drink up the allegedly ‘poisoned’ blood that is being removed. Of course, by the late 1800s, science had advanced to the point where it was realized that this was nonsense, and so leeches fell out of favour in the world of medicine.
But that same scientific revolution saw the development of drugs like heroin and cocaine to cure everything from schizophrenia through to children’s cough. With doctors prescribing these drugs left right and centre, and worldwide epidemic of drug-addiction misery was spawned.
We can’t look back at that period of medical discovery with any sort of mitigating explanations. We can’t say that the doctors prescribed these awful, non-effective and collaterally damaging substances due to ignorance. And the reason why is because there was absolutely no evidence to suggest that these new, popular, trendy and heavily marketed drugs had any beneficial effect for the ailments they were being prescribed for.
So why did fleets of highly trained doctors do this?
Experts and the "need to do something."
Unfortunately, doctors are still prescribing useless (and potentially harmful) remedies to all sorts of modern ailments. A really interesting paper looked at why this is still happening. And lots of reasons were identified that draw lots of parallels to the reliability, quality and maintenance engineers responsible for keeping machines just as ‘healthy’ as doctors want their patients to be.
And perhaps the most significant reason is for the ‘need to do something.’ When something ‘bad’ happens, it is psychologically difficult for many to simply do nothing. That means we rush to do something … anything … in response. We can often see the risk of errors in the act of commission (doing something wrong) is much smaller or less severe than the risk of errors in the act of omission (not doing something when we should).
So that means when we are faced with an unexpected failure, outage or offline event, many of us launch straight into a mandated ‘overhaul,’ ‘servicing’ or ‘preventive maintenance’ action. The problem is that this is little more than a hope that these actions will resolve the problem.
And hope is not a strategy.
Unnecessary maintenance is always, really, really bad.
A study of Land Rover 110 vehicle post servicing failure rates showed that there was absolutely no evidence of wear out in any vehicle for at least 25 000 miles or 40 000 km after the vehicles were serviced or subjected to preventive maintenance.
But … the study found something way more important than a lack of wear out. The failure rate was steadily decreasing for these tens of thousands of miles or kilometres after servicing. What this meant was that the servicing actions were actually ‘spiking’ the failure rates in a completely unnecessary way.
The mandated servicing interval was 6 000 miles or 10 000 km, even though the data showed that the failure rate was still decreasing for tens of thousands of miles or km after that servicing was supposed to be conducted. So preventive maintenance was occurring so frequently that statistically, all failures were based on these failure rate spikes.
This is what we call ‘maintenance induced failures.’
Doctors and ‘pathophysiological models.’
A ‘pathophysiological model’ is a model of how diseases change the human body. Doctors need to understand this model to work out what remedy they prescribe. But in many cases, the entire medical profession holds on to outdated pathophysiological models even when evidence shows otherwise.
For example, James Lind was a Royal Navy physician who conducted clinical trials that showed that citrus fruit cured scurvy. Scurvy was a prevalent disease that decimated the crew of long-distance seafaring vessels due to a lack of vitamin C. Lind and everyone else in the medical professions did not know what vitamins were. Lind doubted his own results, and it took another 50 years for the Royal Navy to start using lemon juice to prevent scurvy.
Even Robert Falcon Scott during his ill-fated Antarctic expedition (150 years after Lind’s clinical trial) believed that mythical ‘ptomaine poisoning’ was behind scurvy. This was reinforced by his observation that eating the meat of Antarctic seals (which were high in vitamin C) cured scurvy.
And the same condition of ignorance afflicts today’s reliability, maintenance and quality engineers. For example, there are lots of ideas about what fraction of components wear out, and how frequently they fail when they do. There are outdated and poorly researched statistics that are decades old but still quoted to this day that downplay the prevalence of wear-out failure mechanisms in today’s machines.
Why are wear-out failure mechanisms so important for any discussion on servicing or preventive maintenance? Wear-out failure mechanisms involve the accumulation of damage where things like lubricants gather contaminants, fatigue cracks slowly grow, and brake pads wear away. Preventive maintenance, overhaul and servicing are simply intended to remove damage, so by definition their focus is only on things that wear-out.
Which is why the fact that the Land Rover 110 failure rate never increased for 25 000 miles or 40 000 km after servicing is very important. It shows that we should not even think about servicing during this period. If there is no evidence of an increase in failure rate, there is no suggestion that anything is wearing out, and by extension no justification for servicing or preventive maintenance.
In fact, when we prescribe unnecessary servicing, we simply spike up that failure rate without any corresponding decrease in wear out failure mechanisms.
Doctors, rituals, and mystique.
The article quoted above described how there are things that doctors traditionally and routinely do … mainly because they are traditional and routine. Be it peer pressure, lack of critical thinking, or just stress, blood tests are traditional and routine before many medical procedures even though the results aren’t reviewed or influence any decision thereafter.
And anyone who has worked as part of a design team, manufacturing shift or maintenance crew know that there are things that are done simply because ‘they have always been done.’
And again, psychology comes into play. There is less perceived risk when people follow what has been traditionally and routinely done before. It is actually a legal defence! So when we do something different to what has always been done, it becomes visible and the focus of any perceived unfavourable outcomes thereafter.
And it also happens that organizations who normalize ‘doing what has always been done’ are not critically thinking about what actually needs to be done. These are the organizations that look for blame and scapegoats when something goes wrong, because they don’t have the engineering or technical expertise to correctly diagnose what did go wrong. Who betide the person who ‘did something different’ when something bad happens.
Preventative maintenance isn't all bad- it comes down to critical thinking.
Which is why not all engineers are good engineers.
The ability to pass tests to get a degree or certification does not always correlate with demonstrating the ability to critically think. Tests are often a measure of people’s ability to memorize facts and processes. They can’t always measure the ability for a mind to think critically in a curious way. They certainly can’t measure the moral conviction an engineer, designer, manufacturing technician or maintainer has to follow what their critical thinking suggests should be done (especially when this is different to what has ‘always’ been done).
And this is why virtually every machine, system, vehicle or device that is subjected to servicing or preventive maintenance is over maintained. Over maintaining drastically reduces reliability. But this is rarely taught or emphasized.
Just as was the case for scurvy, there is ample evidence out there in terms of data in your organization or conclusions made in peer reviewed articles that shows that while servicing is very important, we tend to do it way too much.
And it comes down to psychology. The design team behind a brand-new machine wants to protect their ‘baby.’ So in their minds, there is no risk in over-maintaining (remembering they themselves are not the maintenance crew), and a huge risk in under-maintaining (we are force fed this idea that everything is about to fail all the time unless we do something). And as a result, we have ‘laminated’ servicing intervals that never get reviewed or changed, over-maintenance is normalized, and ad hoc servicing or overhaul is done as knee jerk reaction to ‘bad’ events in the vain hope that all will be made good as a result.
Maintenance regimes must instead be based on a thorough understanding of how YOUR machine or product wears out. And the benefit of this knowledge is that you will know which part of your machine wears out the fastest, meaning that potentially small components or subsystems could be made more robust, allowing servicing and preventive maintenance to become even less frequent.
And of course, plenty of maintenance regimes furiously remove the ‘wrong’ damage from machines, mainly because ‘that is the way it has always been done.’ This wastes time, money, generates a false sense of security and a culture of scapegoating when something inevitably goes wrong.
So, for your machines and equipment, are you prescribing heroin when you should simply be drinking lemon juice?
Comments