Chasing the Rabbit: Official Blog by Author Steven Spear

Incident Reporting Systems (IRSs) have been energetically engaged by hospitals seeking to emulate the aviation industry’s record of safety.  According to safety expert, Dr. Bob Wachter, they cost too much and accomplish too little.  His complaint is backed by sound systems thinking.  IRSs gather and process data that is delayed and aggregated.  While useful for seeing trends and identifying hotspots, such data is not useful for diagnosis and treatment.  By the time there is a response, the conditions that caused the problems may have disappeared.  What is needed is real time, nested problem seeing and problems solving so systems can maintain their stability and responsiveness without overloading some central safety function.

Incident Reporting Systems (IRSs) have become a mainstay of patient safety and quality of care efforts in many hospitals.  According to Dr. Bob Wachter, leader in the patient safety and hospitalist efforts, these IRSs are most often unwieldy, too costly, and largely ineffective.

His critique is wholly consistent with sound systems thinking.

Any complex system-one composed of many ‘components,’ connected in intertwined, multiple, and non-linear ways depend on closed loop feedback and control to maintain stability and responsiveness.  Critical factors in designing these controls are fine-grained, frequent, real time assessment and adjustment of the process.

This is true for situations as diverse as managing blood sugar levels or controlling the flight of inherently unstable fighter jet.

In the case of blood sugar, there are mechanisms working within cells, at the level of tissues, organs, and systems adjusting your metabolism based on what you’ve eaten and when relative to what you are doing.  That is why, be it after eating sugar backed donuts or after several hours without a meal, you can function pretty well, more or less.

In the case of an agile plane like the F-22, there are countless sensors detecting changes in pitch, roll, and yaw, making up to 1 billion adjustments per second in engine thrust and flight control inclination.

Of course, shut off those controls, and you’ve got one heck of a mess.  The diabetic is stuck with self monitoring and crude blood sugar control.  The pilot has to eject from a plane with the computer disabled.  No ‘flying by the seat of one’s pants’ as the old movies would have it.

In the case of Incident Reporting Systems, events happen, they are reported on a delay, information is gathered, and investigation occurs long after the conditions that provoked the incident have changed.  While useful, perhaps, for identifying hot spots-epidemiologists do this to identify disease outbreaks, and sophisticated police departments do this to recognize areas in need of extra policing, they are not an adequate substitute for the real time identification, investigation, and addressing of departures, deviations, and disruptions from targets and norms.

If ISRs are useful for trend identification but not for bona fide problem solving, why might they be so popular?

Here is one possible explanation.

When the health care sector was first chided for the abysmal quality and safety offered to patients, the commercial aviation industry was held up as a paragon of virtue.  After all, flying on an airliner is just about one of the safest things one can do.

When the question was asked why flying is so safe, people undoubtedly noted certain behaviors: accident investigation by the National Transportation Safety Board as the last line of defense but an active reporting system on ‘close calls.’  The theory and practice being that if you see problems that almost have consequence, then you can avoid problems that do have consequence.

Missing, probably was all the micro, real time problem solving and organizational learning that is going on in sub systems and sub sub systems of the entire aviation system.  There is the problem solving and learning that goes on within crews, the cross crew learning that occurs on routes, within aircraft types (e.g., among all USAirways crews flying Boeing 737s), or within airlines.

Likewise the learning networks among air traffic controllers and in maintenance and gate and ground operations.

So, what is missing?  The nested control and learning loops within disciplines, in units, on floors, across service lines and so forth that address local problems locally, thereby not overloading some centralized group with every problem all of the time.

Related posts:

1. The Basic Science of High Velocity Systems: Principles for generating and sustaining improvement and innovation
2. High velocity innovation and food safety…
3. Asking what quality initiatives get sacrificed under budget pressure asks the wrong question…
4. Who Was Caring for Mary–Revisited…
5. Womack’s ‘Beyond Toyota’ is wrong challenge…’beyond lean’ is…