A Surprising Study of Infant Mortality Rates: Evidence-Based Management Meets Evidence Medicine.

We just posted the summary of an intriguing study on www.evidence-basedmanagement.com. This longitudinal study examined the effects of increased collaboration in a sample of 23 neonatal intensive care units on infant mortality rates in a sample of 1061 patients (i.e., newborn babies that were sick enough to be intensive care – but those that did not live three days or were born with severe birth defects were excluded).  It also included a “control” sample of 21 units.  Three university researchers – Ingrid M. Nembhard, Anita Tucker, and Richard Bohmer (who is also a physician)  – worked with Jeffrey Horbar and Joseph Carpenter from the Vermont Oxford Network (a professional association for neonatal intensive care units). They implemented and evaluated interventions designed to increase the amount of collaboration among front-line staff (such as doctors and nurses) and between front-line staff and managers.

Their findings are intriguing. As proponents of the quality movement would predict, when there was greater input from non-physicians in developing treatment plans and more communication among all members of the units, infant mortality rates were lower.  The logic here is that communication (and the related permission to speak-up when a higher-status person is doing something wrong) enables people to make fewer errors and that greater communication enables superior practices to spread more quickly and completely. Similarly, these researchers also found that when front-line workers collaborated more on making process improvements in the units, mortality rates were also lower. That meant selecting projects for improvement, and then using practices during the project such as soliciting staff ideas, educating the staff, using pilot runs and dry runs, and applying the Plan-Do-Study-Act problem solving cycle

BUT the big surprise is that collaboration wasn’t all good.  One kind of collaboration was linked to higher mortality rates.  When front-line employees became more involved in unit governance -- doing things like being involved in decisions about who was hired and fired, the creation of new positions, scheduling, and budget allocation decisions –  mortality rates WENT UP.

Pretty scary, huh?  Perhaps asking employees to participate in management decisions isn’t such a good idea – at least in neonatal intensive care units.  The authors speculate that collaboration may slow decision implementation or that the decisions that are made may be worse because too many compromises are made because there are so many more “voices” driving the decisions . I would also speculate that the staff who were involved in those decisions might have been distracted from their jobs – taking care of sick little babies – and that in some cases (although they were given lots of information) they may have been given a greater voice in decisions that they lacked expertise about. These are just speculations, but just as sham participation is a bad idea, it may also be that authentic participation also has  drawbacks (On the other hand, as the authors emphasize, participation and collaboration in governance has been shown to improve performance in other settings.)

Of course, more research is needed to see if these findings hold in other neonatal intensive care units, let alone in other settings. But it seems to me that distracting from their primary jobs – especially when their primary jobs entail working with very tiny and very sick babies -- might be suspect. And given that, like taking care of sick babies, management work requires skill and experience, so involving people in the process who lack such skills and experience might be unwise at times.

Participation and collaboration are loaded words, and especially in a democracy like ours, we tend to automatically think that more is always better. But on closer inspection, we know this isn’t always the case. To take an extreme example,  I don’t want the pilot who flies my next plane to invite the flight attendants and the passenger to help him or her make decisions about how to fly the plane.

Comments

[Shaun Kieran]

Fascinating. That last part about how the mortality rate went up with increased collaboration definitely needs to be explored and understood.

Much of my own work has been in hospitals, and I see two general principles underlying why collaboration like that described is so good.

First there's the obvious Hawthorne Effect which is that when people know for sure they're being observed, their behavior is different (mostly better.)

Secondly, dovetailing with the awareness that everyone is watching, is the confidence that mistakes are a) routinely uncovered and b) not punished, but corrected. Everyone’s covering everybody’s back.

The working assumption is that everyone wants to get it right, that uncovering and pointing out mistakes is not pulling rank, harassment, or otherwise personal, but a reflection of an authentic focus on the task at hand, the mission.

When gotten right, this “cultural” atmosphere is energizing and contagious.

[Chip Overclock]

Folks interested in software development processes have looked closely at open-source projects. At least some of the most successful ones seem to suggest the need of a "benevolent dictator" who can bring debate to a close and make a final decision.

It's probably not a "once paradigm fits all" situation; the right decision model might even be domain specific.

James Surowiecki never claimed that "The Wisdom of Crowds" was universally applicable.

[Wally Bock]

Excellent post, Bob, and an incredibly important point. Participation and collaboration shouldn't mean "sharing ignorance." When that happens, there's trouble. But participation by informed players and collaboration among stakeholders with diverse informed perspectives can be very powerful indeed.