We are in fact looking at patient de-identified, aggregate data, for well accepted health quality measures, like “immunization rates in children.”

No nose sticking is perceived - I think what you are saying is that an even-handed approach should be taken in looking at data, to do the five why’s - that it is just a map. Yes?

I didn’t mean to be sticking my nose into “medical judgment”, I was thinking more about “management judgment.”

The fear of overreacting from a management standpoint comes from giving people “pats on the back” or “kicks in the butt” when things have “gotten better” or “gotten worse” when it wasn’t statistically true or valid.

We have a very proficient statistical analysis shop here. At the same time, there’s a balance between getting lost in what the data means and acting on the data. The history of the medical profession has typically been steeped in the former (questioning the data and what it means) rather than the latter (intervening to improve health).

The results achieved in our nation as a whole demonstrate that.

Statistical analysis is one component. Medical judgement, and the role of physician leadership, is just as important at this stage.

We are not using the wall to establish causality. Each patient is unique and has their own story. The wall is a guide in deciding areas of emphasis, PDCA style. It’s a journey to the goal, a healthy community, rather than great measures.

Another question, are you using Statistical Process Control to analyze the data and trends? Deming taught (and Donald Wheeler’s book “Understanding Variation is wonderful for this topic) that reacting to or trying to explain every up and down in a metric is actually “tampering” with the system. Certain ups and downs and “trends” are really just normal variation in the process and can’t really be explained by a “special cause” explanation.