You can't use a shovel to make a tent...at least not easily anyway. Keep this in mind when you evaluate the use of control charts in healthcare!
First, using control charts in healthcare is not as simple as it seems. Sure--we all agree that control charting, used properly, can help separate signal from noise when it comes to making changes and monitoring systems. A proper control chart can help prevent type 1 and type 2 errors in performance improvement!
But, ut oh, we need to be very careful when we use the technique in our service industry called healthcare, and here are at least two reasons why. (Disclaimer: I love the use of control charts in healthcare...so let's love them enough to carefully use them right!)
- A control chart does NOT tell you whether a process is making an output that is acceptable.
Consider the recent write-up in Trauma-News.com. (Click the link at the bottom of this entry for the full write up.) Using a p-chart to evaluate performance on VAPs (Ventilator Associated Pneumonias) may sound just great. But ask yourself this: what is the acceptable number of VAPs in your system?
In other words, what is the Voice of the Customer (VOC) in your process, and how many VAPs would they accept? It's a special problem in healthcare to decide who the customer is exactly (more on that here) but, no matter who it is, no third party payor is willing to pay for a VAP and no patient wants one!
So now consider the beautiful control chart. Pretend, for example, you have a perfectly beautiful p chart with NO sign of special cause variation--everything is in control. Amazing! Nothing fails a Shewhart rule or Westinghouse rule or whichever set you want to use to interpret. You'd say "Awesome!". You may even look away from the process because, heck, it looks pretty good and you're a busy person!
...and you'd be wrong.
That's because any VAP is unacceptable. The VOC says so. And this is why Lean Six Sigma (and other methods) teach that we should not apply a control chart to anything until we know the Voice of the Customer and until AFTER we have worked to improve the system. The bottom line is the chart could tell us that we have routine variation (everything is fine), and yet the process is completely unacceptable.
Would you want a process that reliably produces a VAP every month or so? No way! The VOC says "nope". But wait, the control chart said that's normal variation! Now you see the issue.
Refine the process, improve the variation, learn the VOC and then slap a control chart on there to maintain your improvements if you want. But don't fail by using a chart that says statistically things are fine when the VOC tells you they just aren't acceptable.
Sure, there are other ways to use control charts. But you may be lead astray unless you're careful.
So it is with many healthcare endpoints: the control chart misleads staff into thinking everything is OK because, applied improperly, it shows routine variation...of an unacceptable process that still makes VAPs! A statistically in control process that's unacceptable? No good!
- The wrong control chart can be relatively insensitive to significant events.
One classic problem with p charts in particular is that they may be relatively insensitive to rare events. Have one VAP a month? Well, that's pretty rare given how many patients you see. The p chart will likely NOT detect any issue in that healthcare system because it can't. P charts are notorious for this insensitivity to rare events in certain situations.
Click here for a write up about how the attribute charts like the p chart miss rare events (by showing things are ok when they aren't) and so don't help with their detection and correction.
Click here for another write up about using the g chart in healthcare to monitor those rare events.
Bottom line: we need to be very careful about applying control charts in healthcare. Done incorrectly, they lead us astray to both type 1 and type 2 errors.
All of that said, using the tools of statistical process control in healthcare should be applauded--nice work to the Trauma-news.com team for highlighting these useful tools. As healthcare develops and matures in these techniques, let's continue refining and implementing their use!
SPC is the methodology of choice...Because it helps us distinguish between common cause variation (expected, normal variation in data) and special cause variation (which represents a statistically significant change in process). In other words, SPC lets us separate the signal from the noise. A chart for all seasons SPC encompasses dozens of tools and processes, but there is one tool that I turn to more frequently than any other — the P chart. A P chart is a type of control chart that uses proportions to determine whether variation in data is significant. You can use a P chart retrospectively to determine whether a PI initiative was successful. You can also use P charts to monitor patient care or system issues on a concurrent, continuous basis.