I recently invited questions for Michael Balle, the author of the Shingo Prize winning book The Gold Mine: A Novel of Lean Turnaround. He has a new book (published by LEI, my employer), recently published, called “The Lean Manager.”
Here is the third Q&A. If you have a question for Michael, click here for more info on how to submit them. If you have a question in response to this post, please the comments feature. For previous Q&A, click the “Balle” link at the bottom of the post. Michael also has a new blog at the LEI website.
Q: Our Facility is looking at using Lean to address our Medication Errors. Wondering if anyone has done this and if they have any suggestions or comments on their experience?
Answer from Michael Balle:
Yes! Brilliant question. Because the flow stuff in lean is so impressive and, well, relatively simple, it's easy to forget that lean is first and foremost about improving value for the patient, and, in a hospital, the first step of this has to be avoiding medication or any treatment error for patients. This is indeed the right place to start, although not the easiest, but one should not shy away from this project and tackle head on. Lean in healthcare is about patients first, patients second and patients third. Patients!
Addressing medication errors is difficult because of the complex chain of events that leads to such an error. In the wards, there are sets of conditions that align unfavorably and produce an error. In all the bad cases I've had the unfortunate chance to see, which led to patient real danger or death, several conditions accumulated which led to a bad patient decision. The way I understand mediation errors from those I've seen, there are at least three elements to keep in mind to understand the problem:
- First, everyone wants to do the best for the patient. I've never seen any situation (I'm sure they exists but I feel they must be very rare) where what happens to the patient is indifferent to doctor or nurse.
- Second, there's an addition of adverse circumstances, for instance, an alcoholic patient being set up with a slow perfusion before surgery at a time of staff shortage and electricity failure – the patient died from the wrong drug being administered.
- Third, and this is strange, in most cases, someone notices something wrong, but then doesn't pick it up and lets the wrong action continue to dire outcomes.
So when these cases are addressed post hoc, it's always someone's action that's involved (usually nurse or consultant), But let's inquire more profoundly in this. What really happened is that in a difficult situation the person did not perform better than the system: they got trapped. Yes, because healthcare staff members have lives in their hands they should be vigilant at all times and smarter than the system, but overwork at the end of a shift, particularly with inexperienced staff this is a lot to ask, even from supermen and superwomen.
How can we understand such situations better? This is where lean concepts can be enormously helpful in understanding, and then improving the situation. Three core concepts that lean thinkers know are confirmation, overburden and unlevelness.
Confirmation is the notion of building in quality in every move. Doing any action should involve an immediate testing of this action to see whether it's done correctly or not. This means precise criteria of what distinguishes a good job from a bad job. This may sound obvious but in real life is far from simple. Take the plain jab. On the simple fact of cleaning (yes? No?), disinfecting (yes? No?) the patient's skin before taking a simple blood sample or an injection I've seen an incredible amount of local difference. There is certainly no consensus on the value of antiseptic cleansing on percutaneous injections, but it makes an intuitive sense and a lot of people do it. Now, any antiseptic needs a minimum contact time to be effective – but how long ? If you're in a hurry because of a long list of patients waiting, one could easily wipe with disinfectant before injection, and then puncture the skin without letting the few seconds necessary for the product to actually desinfect, thus performing badly an operation that might not be necessary in the first place. I am not in any case suggesting there is a one best way here because it completely depends of patient conditions and circumstances, what I'm trying to say is that many physicians and nurses can find themselves performing a repetitive act with consequences for patients with no clear confirmation standards.
The first step then to reduce medication errors and other adverse events is to have a clear list of the most frequent patient acts and to start reviewing the confirmation part of the process. Do is not enough. Do-and-test has too be set for every act. This is not an easy job, but this is the key to building quality into the process. Every employee can test their own work and whenever they have a doubt, call someone else immediately.
The second element to have in mind is overburden. In most tragic cases of errors, there was something else going on at the time that put staff in a specially stressful situation: too many patients, not enough staff, a special patient condition no one was particularly clear about. There are many forms of overburden: illegible doctor writing on prescription (the nurse has to be smarter than the system and interpret correctly), out-of-date products on the shelves (the nurse has to be smarter than the system and check the date of every material), different product and different dosages in the same presentation, sometimes written in small, illegible letters (the nurse has to be smarter than the system and know exactly what is what), and so on. Overburden is anything that creates unreasonable physical (get me these materials on the top shelf in an hurry please) or mental (can't you make the difference with two identical capsules with microscopic writing on them?) ergonomic burden.
A key lean activity is to systematically spot these overburdening situations and work with kaizen groups to 1) preferably eliminate them or 2) build the appropriate countermeasures. For instance, it might be very hard to fight the drug companies about their products' conditioning – so when different products and dosages have exactly the same appearance, it is possible to kaizen the storage to minimize the chances of injecting the wrong product or dosage, even in a panic. So, the lean technique here is “5S”, on in Michael E. Parker's more user friendly formulation: SAFE-T:
- Separate what is not needed
- Arrange in order
- Finish with inspection
- Everyone should understand the rules
- Test constantly to confirm the system is working
The point of conducting these frequent activities is to eliminate cases of overburden one by one through kaizen.
Thirdly, unlevelness is also a critical contributor to accidents. This is not the actual deficiencies in the delivery system, but this is overburden we deliberately create through our scheduling system (not deliberate in the sense of mean, but in the sense of scheduling deliberately meaning well without understanding the consequences). For instance, scheduling all patient intervention at the beginning of the day thinking that patients prefer to get it over and done with creates a massive resource bottleneck and so on.
Finally, the most disturbing aspect of the real bad cases of medication error I've seen is that the patient himself or herself starts reacting. In one case, I've been part of investigating the death of an elderly patient in the recovery room after an operation because the wrong blood was being transfused – and the patient knew it and said so! She realized that it wasn't here name on the label, and somehow no one listened to her enough to check (there had been a mix up at the blood bank and there were two labels – horrific). Because she was confused coming out of anesthesia, she didn't make much sense, didn't do a scene, the staff were shorthanded and when one nurse senior enough actually realized it was too late. So what kind of lessons can we draw from such tragic situations (yes, it hurts):
First, I believe we all need to acknowledge that errors do happen and we have to face to the fact that we need to understand why and how they happen. This is difficult because contrarily to industrial processes, the errors I've seen in hospitals never have twice the same pathway – so it's important to constantly get staff to think deeply about this and start deploying systematic confirmation. Keeping people focused on confirmation in tehir busy jobs requires a regular briefing on “what we did wrong, today, this week: Let's think deeply about it.”
Second, although everyone is very busy, it is key to continuously teach nurses and doctors to look at the patient and see. Exceptional physicians and nurses can catch errors before it's too late because they can tell something is not right with the patient – they're too white, or too blue or that wound doesn't look and smell right and so on. This is a teachable skill, but it needs dedication and focus. This is exactly the judgment that Toyota tries to impart to all their line operators about every part (does the sound of the car door closing feel right?)
To do so, the only known technique is what is usually known as “stop-at-defect” but is really call whenever you have a doubt. The lean technique of “andon” is about management reactivity, certainly, but more importantly about staff training. The key thing here is never to make the decision “I have a doubt but let's see how this develop” alone. Wards must be structured so that whenever a nurse of a physician has a doubt about a patient, they should immediately be encouraged to share this doubt with someone else and never be left alone with a problem.
So, no easy answers. In terms of reducing medication errors (I have not adressed wrong diagnostic which is something else again), lean has a number of leads to attack this core problem:
- First, confirmation and immediate stop-think-check at every non standard patient reaction. This is a key effort to be conducted by physician and ward managers, to make sure that every staff in contact with the patient is clear on their confirmation criteria on theirr main activities so they can test whether the action they just performed is good work or not.
- Second, reducing overburden. By conducting simple “5S” and kaizen type activities one can identify all the “traps” in a ward, all the situations where the consultant or the nurse have to be smarter than the system to actually do the work well (nobody told you? We keep those in the top left-hand drawer under key because someone use to steal them?). Through repeated 5S activities in the ward, patient standardization of basic acts and kaizen, the overburden elements will reduce, and with them the risk of adverse events for patients.
- Third, learning to limit unlevelness – having a critical look at our own scheduling procedures and figuring out how this really affect both patients and staff. Nurses, typically batch operations, and then rest. They have periods of frenzied activity, and then lulls. This is a deliberate scheduling decision which can be questioned continuously.
- Fourthly, and probably more importantly, constantly remind healthcare staff to look, see and listen to the patient when they are doing any activity (have you ever heard porters discussing their weekends while they push the patient trolley around?). Most people who go into healthcare choose this profession because they want to help patients heal – but it's a harrowing job, with a lot of day-to-day suffering to deal with, and the management system needs to constantly remind everyone of why they are actually there: help patients get better and return to their normal lives as soon as possible, with the least pain and hassle.
No easy answer, no magic wand. Still lean is probably the most appropriate management method to deal with intractable issues such as medication errors. With some hard work, patience and persistence, it will deliver spectacular results as staff themselves transforms their own practice and working environments.
Don't want to miss a post or podcast? Subscribe to get notified about posts via email daily or weekly.