Hawaii under attack! Incoming missiles! People ran here and there. Representative Tulsi Gabbard, always a bastion of sober consideration, gathered her family in the bathroom. On Maui, Peter Shenkin slept, thinking it wasn't worth getting up for. His partner in gym ownership, Logan, out golfing when his party was flagged back to the clubhouse, objected. “It's probably an error,” quoth Logan, “and if it's not and we're going to be incinerated, at least let's see if we can get a few more holes in.” Golfers are rarely dissuaded from their chosen path.
In the end, of course, it was an error, the origin of which was depicted this way: one employee going off duty was presented on the computer screen with the choice of two buttons, one saying no attack and one saying yes attack. He chose the wrong one and the debacle unfolded. Understandable calls for his head have been unavailing – Pete says that in Hawaii, everyone is someone else's cousin and can be called a “good guy” – and according to the Honolulu Star-Advertiser, he has not been cooperating with the investigation. We'll see what happens.
Aside from the insight this event affords to the quality of Hawaiian governmental organization – one of the states where implementation of an Obamacare insurance exchange collapsed ignominiously to the surprise of few – it reminds me of my experience in running a pediatric practice. All practices are faced with the problem of how to maintain and improve quality. How can we do what we are supposed to do, what we want to do, what we need to do? How can we make sure all our prevention is done completely and correctly, how can we make sure our patients receive excellent customer service, how can we be sure we diagnose correctly, follow up correctly, remember everything we need to remember? It's a challenge not altogether different from the one that faced Hawaiian Civil Defense, at which they failed so miserably.
When you come down to it, there are two basic ways to deliver high quality and reliability. One we can call Professional Enhancement (PE). PE aims to improve the knowledge, skill, and motivation of the employee or the clinician. The other one we can call Systematic Reengineering (SR), where the focus is not on the individual but rather on the system. PE aims to put things into the brain of the individual. SR aims to put things into Standard Operating Procedures, which are outside the individual brain, but exist instead on routines, on paper, or in a computer.
In medicine, clinicians tend to think in terms of PE. We are all used to this, because this is what we were exposed to in all our years of schooling and clinical training. Remember this, there will be a test! Don't forget to put this in your differential diagnosis! Did you check to make sure the referral was completed? They are all over us, and we come to think we should be all over ourselves as we move into practice. That is our SOP – remember! It's not bad training per se, but it's incomplete.
Students of organizations take another tack. While they are aware of the requirements of knowledge, intelligence, and mindfulness of individuals in organizations, they tend to think not primarily in terms of PE, but rather in terms of SR. (Or sometimes not so aware, as this from W. Edwards Deming: “Ranking (of personnel) is a farce. Apparent performance is mostly attributable to the system that the individual works in, not to the individual himself.”i) SR comes from the context of QI in industry, where establishing regular procedures decreases variability, with the mantra “do-it-once-do-it-right.” The SR approach finds recurrent stereotypic situations within medicine’s endless complexity that can be choreographed for standard execution with low variation, and hence higher quality. SR introduces changes in how the system operates by introducing such elements as care plans, flow sheets, forms, and reminders, all of which could now theoretically be made so much easier with Electronic Medical Records. (In practice, of course, EMRs being the oft-reviled cumbersome obstacles that are best avoided, not so much – but I'm sure that's temporary. Pretty sure.) The goal of SR is for the clinical team to operate ensconced in a system where the easiest thing to do is also the right thing to do.
Those familiar with the work of Danny Kahneman will recognize the similarity of the distinction of PE and SR to his difference of thinking fast and thinking slow.ii If you do something routinely, it becomes a reflex and you don't have to think very hard about it, you just do it automatically and it's the right thing to do. As a matter of fact, thinking about it can lead to a problem which in sports is called, somewhat inelegantly, “the yips.”iii Having a behavior embedded in a system within an organization is the equivalent of thinking habitually, and therefore fast. When you have to cogitate about what the right thing is to do, when you have to search your memory bank, when there is nothing in the system to guide you and no button to push or menu to bring up, that's the equivalent of thinking slow, and is prone to error.
The key to management and organizational analysis is to realize that both approaches are applicable depending on the situation. For many situations – perhaps for all situations where one can think of a good system – SR is best. Clearly, for instance, narcotics control is better handled by a computerized check-in-check-out system than by education on the harms of addiction. In our own practice we approached the complex question of asthma, where I was bedeviled by clinicians opining “well, what I like to do is....” Instead of the extreme PE approach leading to everyone's taking his or own special path, we tried to have standard approaches, where one check mark led to a bevy of actions that were SOP for standard asthmatics. One check mark led to doing the right thing at the right time every time – at least for those who accepted the system. Docs are a tough crowd.
Some say that SR is inevitably superioriv, but it seems to me that this would be true primarily for stereotypic situations – immunizations, for instance. PE needs to be the primary approach of choice for human relations, complex diagnoses, unique situations, and judgments as members of adaptive organizations. Empathy is better achieved by education and reinforcement than by signs on the wall saying “We Care.” As clinicians we understand very well the non-SR behaviors of being reflective, seeking the hidden meaning behind patient visits, and continually trying to detect disease early from small clues.v Clinicians remember when a stray piece of medical knowledge led to a clever diagnosis, or when even in the midst of a busy schedule they were patient with a patient and then made an important diagnosis. We recall the meaning we bring to some patients’ lives, and vice-versa. This is the core of traditional medical quality. Silicon Valley has a lot to add to medicine – a lot! – but the day is far off when these qualities will be replaced.
To be practical, if you are involved in managing an organization, when faced with a quality issue, the first question to ask is often, can we put this into a systematic routine? If so, that's the way to go, with the usual problems of implementation, including operator acceptance and understanding. Sometimes, however, as with empathy, the solution will not be SR, but rather PE. And likewise, if you look hard enough, I think that very often the solutions will be a mixture. If your problem is friendly reception at the front desk, part of the solution will be education and practice on how to be friendly, and some of the solution will be rewriting the job description and finding good ways of measurement and feedback. Most times, it seems that an artful mix is required.
Which brings us back to the missile attack. What if this error had been made in a hospital? My friend Paul Levy, former CEO of Beth Israel Deaconess Hospital in Boston, is fond of recounting his intelligent approach to error. If a nurse makes a dosing error, for instance, and he or she is called in to recount the circumstances in the process of seeking the root cause of the error, he likes to ask: “Did you intend to do what you did?” It would be unusual for the answer to be “Yes.” His inference, then, is that most often one should seek the error in the system that led the nurse to give the wrong dose. It can be something as simple as two vials of different strengths that look virtually the same sitting side by side in the refrigerator, leaving it to the nurse to make a difficult distinction as to which is which. Fair enough – systems problem. Amazon, King of Systems, fixes this by storing similar looking items at a distance from one another. I often differ with Paul, however, in leaving it at that and exonerating the perpetrator. Yes, it is expected that people will make mistakes, but it is also fair to ask how often that particular person has made mistakes, whether sloppiness in endemic to that person, whether mis-assignments are habitually made by that supervisor, etc. To me, it's always a mix.
So if the non-cooperating nuclear missile attack declarer was confused by similar looking buttons sitting side by side – basically, an attractive nuisance – that's a systems problem. If there is no second check necessary in sending out the attack notice, that's a systems problem. If there is no way to expeditiously correct an error in something less than the 38 minutes it took in Hawaii while Tulsi Gabbard was cowering in her bathroom, when the governor knew within 2 minutes that it was an error but couldn't tweet out a correction because he forgot his Twitter password, that's a systems problem (not to mention a Governor problem.)vi Instead of calling for the head of the careless employee, it's the head of the system designer that should be called for. Differing with Paul, I'm always looking for incompetence, and in this case the probability that the job was awarded to someone with a connection, this being Hawaii.
Or, if the employee, despite his inevitably being somebody's protected cousin, if he continues to be uncooperative, it's entirely possible that he could suffer not from the original act, but because of obstruction of justice. Yes, it's a system, but in the end, systems are manned by people, and everyone has to take responsibility.
It will be interesting to see what happens, especially as the one-term Governor is being challenged. It will be interesting to see if anyone is implicated, and if anyone suffers anything, and if an “A team” is brought in to fix the system, à la Obamacare, and if people then wonder why there wasn't an A Team in the first instance.
But in the end it was only the public who suffered. And poor Tulsi in the bathroom. No heads will likely roll, cousins will keep their jobs, and another contract will be let to a putative A Team, headed by someone else's cousin. Hawaii....
iDeming WE, The New Economics for Industry, Government, and Education. 1994, MIT Press, Cambridge, Massachusetts.
ivShojania KG, and Grimshaw JM. Evidence-Based Quality Improvement: The State of the Science. Health Affairs 2005; 24:138-150.
vMiller WL, McDaniel RR, Jr., Crabtree BF, et al., Practice Jazz: Understanding Variation in Family Practices Using Complexity Science. J of Family Practice 2001; 50: 872-878.