Petri Dishes and Healthcare DFSS

There's a Reason for Small Planned Experiments

August 16, 2009

I wonder what German bacteriologist Julius Richard Petri would think about the current legislative initiative to re‐engineer the U.S. healthcare system? Petri is credited with inventing the Petri dish, a small glass container that allows scientists to experiment by culturing microbes in a controlled, safe environment.

As a scientist, Petri might ask "Is there an experimental plan to run small trials of the planned changes to determine whether or not they will yield the desired result?" Sadly, the answer is apparently NO. Politicians have been rushing to approve legislation that has not even been fully reviewed, let alone tested. This is not a commentary on healthcare policy, or the objectives of healthcare reform — it's a debate about process.

In the Lean Six Sigma world, we teach experimentation to confirm theories, followed by pilots to confirm the experimental data prior to product launch. Even relatively simple process improvement activities can go seriously awry without a clear understanding of the Voice of the Customer, or an understanding of the underlying value‐creating transfer function — quantified relationships between inputs and output.

I was reminded of the risks lately when an organization in our industry launched a major new IT system, only to discover upon launch that the system was rejected by customers. The IT project managers had not run any trials or focus groups or planned experiments of any sort, and the failure yielded a very painful lesson. There was no lack of good intentions on the part of the company, just as there is no lack of good intentions on the part of our politicians. But in both cases, the process was(is) utterly inadequate.

From my perspective as a Lean Six Sigma practitioner, I'm horrified that our national leaders are skipping from a poorly drafted Charter and an incomplete Define phase directly to the Improve phase, without bothering to spend any time at all in Measure or Analyze. This is not a typical "quick hit" corrective action that is inexpensive, painless to implement, and easily reversed if it doesn't work.

Healthcare represents approximately $2.5 billion of annual U.S. economic activity, and is probably the most complex of all value streams. What are the odds that a system of such size and complexity can be restructured effectively without running any experiments? The politicians involved are either completely ignorant or incredibly arrogant, or some combination of the two.

People who wouldn't bake a cake without tasting the batter first are willing to redesign the most complicated system on the planet without any pilots. It doesn't make any sense. If this is a good process, then we should encourage the pharmaceutical companies to bring all their high‐potential drugs to market without any clinical trials.

Jim Womack, President of the Lean Enterprise Institute, recently commented on the "Mind of the Lean Manager," and noted: "…the lean manager believes that all problem solving is about experimentation by means of Plan Do Check Act. No one can know the answer before experiments are conducted and the many experiments that fail will yield valuable learning that can be applied to the next round of experiments."

So where is the plan to conduct experiments on intended revisions to the healthcare system? What are the odds that politicians with no experience in data‐driven process improvement know all the answers before running any trials, pilots, or experiments? Perhaps we should try to persuade them to slow down and do it right.