Part 5: Blind Spots Can Kill Patients-Simplicating, Satisficing, and Bounded Rationality | Joe Bormel, M.D. | Healthcare Blogs Skip to content Skip to navigation

Part 5: Blind Spots Can Kill Patients-Simplicating, Satisficing, and Bounded Rationality

February 28, 2011
| Reprints
Unlike skills and knowledge, it's either impractical or impossible to train talent

Aviation examples and metaphors serve healthcare well. Patients “ spiral downward and crash.” They can be on “ glide paths” with expected destinations at predictable times, and operating a health system with a limited number of runways, spelled “ beds, doctors, nurses, and competencies,” has eerie parallels to air traffic control.

In a legendary quote, Dr. Lawrence Weed said, “The difference between physicians and pilots is that pilots realize that they have to get into the airplane with the people they care for.” For the record, I personally know many physicians who realize this, but the point is still an important one.

The most basic and often unspoken, if not overlooked, aspect of management, both in aviation and in healthcare, is Goldratt's law of three constraints. The idea is that almost every system can be characterized by three high level constraints. If you fail to recognize or ignore even one of these constraints, you will crash. In aviation, I'm told the three constraints are air speed, altitude and fuel level. If you run out of any of the three, no matter how well the other two are managed, the game, and possibly your life, is over. For projects, the law of three constraints nets out to scope, time and resources. You probably get the universality of this point.

In the first four parts of this series we focused solely on what amounts to the Meaningful Use constraint. While important and a necessary precondition for HCIT-enabled improvement, it is only one of three constraints. With no disrespect to the teachings of Avedis Donabedian, structure, process and outcomes focuses aren't enough to improve our healthcare system. The challenge is attending simultaneously to three constraints. In our case, those three constraints are

1) achieving Meaningful Use through building the HITECH required infrastructure,

2) ensuring adequately talented people are on board relative to the full range of necessary individual competencies, and

3) getting an organization to actually organize itself and be healthy. The first and second of these are a connection to the earlier post on reducing cognitive errors.

To borrow a chapter from Jim Collins’ book, Good to Great: Why Some Companies Make The Leap... And Others Don’t, the first of our three constraints is the equivalent of “ getting the right folks on the bus.” This is the first place where cognitive abilities and related executive functions become essential concerns. You must hire for and prune for these talents.

Unlike skills and knowledge, it's either impractical or impossible to train talent. The flip side of this, and perhaps the more positive, is that for most of us our talent, including the talent for cognitive processes, increases in dramatic ways every five to ten years. If you've been in the same role for that long and are doing the same things in the same ways, opportunity may be knocking. Maturity is important. It doesn't stop growing.

Second, the important role of the organization is often poorly understood and even more poorly executed. Organizations (aka Governance) serve a critical role in structuring work and people’s roles to get work done. This was much less so 50 years ago when economists thought that markets alone were the most efficient " decision makers" regarding where resources should be invested. The work of Nobel Prize winning economist Herbert Simon was one of the first places where explicit discussions and considerations are found.

This, interestingly enough, just like airspeed, altitude and fuel level, interact with cognitive abilities. In the same way that it’s impractical or impossible for humans to choose each move in a chess game by exhaustively considering all downstream consequences of all possible moves, organizations serve a similar function in reducing the complexity of decision making. Organizations are often incapable of exhaustive planning, but can organize and delegate a path to meeting their goals. A process that Simon terms “bounded rationality.” No matter how well an organization hires, or how intelligent, balanced and capable the staff may be, if the organization is actually disorganized (or worse yet, toxic), no individual or staff can collectively dig its way out of the grave. Note that I appreciate I'm preaching to a choir of my readers here.




Dr. Bormel,
Thank you for addressing a "deep" subject that most bloggers would avoid. Your willingness to go where few are willing is most appreciated.

Now for my question. It appears that you're advocating that government entities can, and perhaps should, define things such as Meaningful Use, but that legislating such is likely to be counter productive. Is that correct?

In fact, I think that over-involvement by government can, as it has so often in the past, lead to mandates that inhibit, rather than promote advancemets.

What are your thoughts, and are we on the same track?

Doc Benjamin

Thanks Doc Benjamin, for the kind words and recognition. Yes, this five-part series ended up taking us deeper than I foresaw.

In my last post (HIMSS2011), I quoted Dr Mostashari using the term "Simplicating." Astute readers will appreciate the similarities with Herbert Simon's "Satisficing."  I've been teased mercilessly since that post in a good hearted way, publicly and privately.  The word, "simplicate" truly resonated with a lot of people.

That brings me to your question about government legislation.  A wise friend and ONC insider confided that there is strong evidence that neither an all-market, or private sector approache, nor legislative approaches will result in the HCIT we all want for our care.  The fact that many of us have multiple vested economic interests makes the equation considerably more complicated.

Although we're probably on the same track, I'll summarize my high-level take:

1) We need payment reform.  The current FFS doesn't motivate the right behavior or delivery system.

2) Healthcare economics is a classic Wicked Problem. More correctly, it's a Wicked Predicament. I've written about this before, as have Richard Farson and others.  "Solutions" cannot solve a predicament.  They should only be expected to improve (and change) the problems.  That can take us to a better, higher value system, but not a total fix.

In a chat last week with a recent Ivy-league professor of Economics, I ask him about Herbert Simon's work, impact and relevance to Healthcare Informatics. In short, he said that Simon's work, including satisficing, is recognized as a critical and relevant framework. The problem is that it doesn't lead to making predictions based on mathematical models. Therefore, it's applicability to academics is limited. In the non-academic real world, very few executives are committed to modeling, planning and writing.