Warning: You May Be Asking The Wrong Questions About Quality

By:  David Kashmer, MD MBA FACS (@DavidKashmer, LinkedIn profile here.)

 

A recent blog entry by a colleague of mine (see it here) focuses on the importance of taking care of critically ill patients at difficult hours in challenging situations. This struck a real chord with me, as my focus is on how to construct healthcare services for high quality outcomes. Building a service for high quality means, in part, that service needs to perform similarly under often varying conditions. It needs to be both resilient and robust.  Why is it that many hospital services aren’t constructed to be able to perform at this level, as our colleague asked? What does it take to create a robust service that performs well twenty four hours each day rather than just during the easier times? This entry contends that how and which questions we ask about quality can set our systems up for erratic performance down the line.

 

It’s No Easy Task To Design For Quality

 

First, it’s important to realize that creating a service that performs at difficult hours is very challenging. My colleague’s entry gave me a reason to write up some tips and techniques to make the job easier.

 

One of the biggest keys is actually data collection. That is, first, the team must focus on its data rather its gut or how it feels things are. Remember, this is because most service industries perform at a rate of one defect in every thousand opportunities. My claim is that level of defect, well known to pervade service industries, is very challenging to feel on a day to day basis. Staff often think things are going great (or at least very well) because 999 times out of 1000 things do, in fact, go fairly well. It lulls one to sleep.  However, in high stakes fields like healthcare, we’ve demonstrated that a defect level of 1 in 1000 opportunities is completely unacceptable. For more on the reasons why, look here.

 

How Data Are Collected And When They Are Collected Is Key

 

Just as important as being willing to look at (and respond to) data is making sure that the data are being collected in a way that’s non-pejorative. One of the barriers in healthcare, as we’ve previously discussed (here and also here), is that we often equate data collection with an out-to-get-you mentality. (Maybe it’s related to tort law and malpractice issues–who knows.) It’s important that the data be collected regarding the team as a whole and not be personally assignable in any way. This key to the process improvement system creates a process that is non-pejorative, and lessens the chance that the process improvement system will be used as a weapon by one provider or staffer against another. It helps ensure that people will participate and be willing to collect, review, and respond to data.

 

Another important idea concerns collecting data from the entirety of the system. Often, when we see process improvement done, it’s focused on the hours from which it’s easy to get data. Those are the good ol’ nine to five hours. However, as my colleague points out here, trauma (for example) is “a disease of nights and weekends”. So, for high stakes, challenging services lines like trauma, it’s even more important to sample performance during nights, weekends, and other times from which it is generally more difficult to get data.  Those are often the key times!

 

It is important to realize how the answers we get to questions about quality are often framed by the nature of the data we collect. Let me say it again:  in healthcare, we need to collect data from those hard-to-get times because that’s when issues line up and the bad things have a higher risk of occurrence.  Remember the advice to collect data directly from the process whenever possible rather than sampling cleaned data from hospital databases–prospective data taken right from the action is often a lot more valuable.

 

Design In A Way To Respond To Surges In Volume

 

Next, for services like trauma, acute care surgery, and hospitalist medicine, one of the important concepts is surge capacity. How well does the system absorb large influxes of patients? Sometimes patients come in as multiple trauma activations at once. Three to four patients (or more) may show up at one time followed by long stretches in between of one patient arrival at time. Therefore, the ability to flex up to provide the same level of care when three or four patients come in together is every bit as key as performing well when one patient at a time enters the system. The capacity to be resilient and robust must be consciously designed into the system.  Usually a failure mode effects analysis (more on that FMEA technique here) will reveal how focus should be placed directly on events such as multiple trauma activations.

 

The Questions We Ask Determine The Systems We Build

 

So, at the end of the day, my colleague’s question about why things aren’t designed in healthcare to run consistently over 24 hours is based, in large part, on the nature of the questions we pose to our systems. The questions we ask with data, and how we ask them, leads us to design systems that are more apt to work well from 9AM to 5PM.  The first of these counts is we often don’t collect data or respond to it for fear of personal reprisals. Second, if we do collect data, we often don’t collect the data to completely embody the robustness of the system. A third major challenge is that our data often don’t reflect the surge capacity in these high stakes situations where multiple patients may enter the system at once. Concentrating on each of these important ideas allows us to create a high quality, robust system that has a better chance of humming along consistently for 24 hours a day.

 

Knowing Obligates Us To Act

 

Once we know these important facts, if we are serious about quality and robustness in our system, we become obligated to progressively improve our care during those times when weaknesses in our system line up–especially nights and weekends.  We’ve learned to avoid framing certain questions about quality in the wrong way.  When we collect data from the full house of the system, we become obligated to respond to it, as a team, in a meaningful way for the good of our patients.