Operating with Data: Surgeon David Kashmer Helps Patients with Statistics and Scalpel

Dr. David Kashmer

Everyone agrees that health care should be delivered as efficiently—and effectively—as possible. The medical field has yet to adopt the data-driven quality improvement (QI) methods as rigorously as many other industries. But Dr. David Kashmer, the Chief of Surgery for Signature Healthcare in Brockton, MA, consistently applies QI methods to trauma and acute care surgery, and helps fellow practitioners learn to do the same.

When this Lean Six Sigma Black Belt first brought the idea of analyzing data into the operating theater, he encountered some resistance. “I kept hearing, ‘This guy is nuts,’” Kashmer recalls. “’What’s he even talking about?’” But 10 years later, his skeptical colleagues have become true believers, and Kashmer knows that even simple statistical analysis can yield big improvements in patient outcomes. He’s got the data to prove it. We asked Kashmer to share some of his insights gleaned from a decade of challenging misconceptions, extracting meaning from data, and getting results using Minitab Statistical Software.

Minitab: What triggered your appreciation of QI and its applications in healthcare?

Kashmer: While learning about statistical process control, I really came to see it as a different tool, and I realized that this power could be applied to what we did every day in surgery. What really resonated with me was that a lot of the tools we’re looking for in healthcare already exist, it’s just that we typically don’t know about them. It’s not what we’re taught in medical school, so we’re sort of reinventing what’s already been done.

Minitab: What QI tools/techniques have you incorporated into surgery?

Kashmer: I’ve used multiple regression and different quality tools in Minitab to make a direct impact on how we improve patient care. I use them routinely to obtain quality outcomes. For example, I’ll look at what factors are significantly associated with how long patients stay in the emergency department, or whether patients get wound infections. I’m able to use our data and our population to make meaningful quality improvements, and I’ve been doing that for more than a decade now. Measuring the normality of a data set with tools like the Anderson-Darling test helps our team choose the right statistical analysis—so we can confirm that our changes have yielded real improvements rather than relying on what our intuition tells us was successful.

Minitab: How do you explain the general reluctance to implement QI techniques and statistical control tools?

Kashmer: Generally in healthcare—especially with physicians who are undereducated about statistical tools—the perceived value of these tools is far less than their actual value. The tools aren’t typically applied for process improvement in healthcare because they seem too complex— you have to learn all the techniques, know which tool to use in a given situation and with which data, and even know how to set up certain tables correctly. But the tools are simple with Minitab, whatever it might be: developing a final regression equation, finding the R squared, or validating a new measurement tool in a Gage R&R study.

The Assistant’s interactive decision tree (above) guides you to the right statistical tool by posing a series of questions about the type of data you’re working with and the objective of your analysis.

Information about how to set up, collect, and enter your data eliminates guesswork and ensures a successful analysis.

Advanced tools that examine a population to show statistical significance are very powerful, but the language of health care is different from the language of statistics. This hasn’t been accounted for in the healthcare decisions that are based on the data we see on an individual basis, so the expectations are low. Often in Surgery, process improvement is driven by individual cases. Some data are used, yet important ideas such as how to calculate adequate sample size and whether some percentage moving up or down has real meaning are rarely considered. When surgeons think about these ideas, the language of the motivation for using statistics—to guard against type 1 and type 2 errors—is lost on us. We focus more on what we think will help an individual patient in a particular situation. But when we learn how statistics can help us to avoid making a change when nothing was wrong with the patient, or to avoid thinking there wasn’t a problem when there was one…well, that’s when these techniques become much more powerful and interesting.

Minitab: Have the benefits of applying QI techniques met your expectations?

Kashmer: The tools in Minitab and statistical process control have been exactly as valuable as I thought they would be, meaning very valuable. For example, we can make a change in practice and then use data to see if there’s been a significant improvement, or if there hasn’t been a significant improvement in a particular metric. People don’t typically do that. Often, we’ll make changes and if the percentage looks a little higher, we celebrate. If it looks a little lower when we wanted it to be higher, then we’re upset. If it’s green on the dashboard, then we love it. If it’s red on the dashboard, we hate it. But using the statistical process control tools in Minitab and other packages can help you determine which improvements are truly significant, and can guard against type 1 and type 2 errors. Right now, many things in healthcare do not get that kind of evaluation—so we end up making a ton of changes because we want to do a good job for patients, but end up chasing our tails. Other times we do not make enough changes, and that’s arguably even worse. Tools like Minitab help you avoid uncertainty about whether an improvement has worked and whether you should do more.

Minitab: How did you overcome your colleagues’ reluctance to use statistical methods and how have you been incorporating the techniques into surgery?

Kashmer: I had the opportunity to do some turnarounds where different systems and outputs performed really well in different sections of Surgery. I developed a team, built a shared knowledge base with the team, and focused on visible, early improvements. Again, at first, people really thought I was crazy—but when they saw the results from using statistical process control rather than typical improvement methods, they understood and began to appreciate their value.

Minitab: So how did you get traction?

Kashmer: As we did Lean and Six Sigma, we made changes based on our data analysis with Minitab and saw significant improvements. The group helped everyone on our team learn what those statistics meant, what they guard against, and how they help us—it was a combination for success. As we produced even better outcomes, more people started to buy in. They were receptive to results, and the fact that doing it this way often keeps the team together rather than hunting for who “messed up” in a given situation. They see things in the system, and they see the distribution of the data they all share together. I really try not to ever separate things out by individual providers, and that seems to be a useful part of it also.

Minitab: What obstacles have you encountered while employing statistical tools in surgery and implementing changes?

Kashmer: Making decisions with data is challenging since it doesn’t resonate with everyone. Putting a human face on data and using it to tell a story that people can feel is key when talking about the true performance of our system.

Minitab: Are there any specific tools or methods that help you tell the story?

Kashmer: The Minitab Assistant is really useful for providing visuals for different tools. I may understand why an analysis is different and what it means, but it’s useful to walk a group through it, and the Assistant explains exactly why Minitab is asking for something, and shows them why so that we can move on to the next step. That really adds value, having the software walk you through, say, a gage R&R analysis and bring the team to an understanding. They know why you’re using the tool and how you got the output. I’ve found it to be very useful.

The measurement systems analysis method called Gage R&R allows you to assess and validate measurement tools. The Assistant provides requirements, assumptions, and guidelines (above) that ensure an effective analysis.

Minitab: What advice would you give to someone who is trying to deliver an idea using data?

Kashmer: Keep it clean and keep the message simple and human, meaning know the measures of variation and the risk in a system. Pare the output down to the simplest part of the message—show the distribution and put a human message on it. Leverage the work a tool like Minitab does for you—it highlights, makes distributions and graphic outputs look beautiful, and gives you exactly what you need in a useful way.

Minitab: Where have you seen the biggest impacts of QI?

Kashmer: With continuous data. The level of development for most hospitals is to give you discrete data, such as percentages, but Minitab shows you the distribution of continuous variables, such as time, and these allow you to tell a much more robust story. For example, using the percentage above a certain time indicates the ER’s or the trauma surgery patients’ length of stay, but for critically ill patients, that percentage may not be as valuable as looking at the distribution. Minitab is really good about not just showing the distribution, but it can easily highlight important aspects as people walk through their analysis, making it much easier for hospitals to process data in a way that extracts meaning better. I’d love to see more hospitals using rather than shying away from continuous data.

Minitab: Do any major success stories come to mind?

Kashmer: I can highlight two great examples. We felt we had an issue with trauma patients in the emergency department, but the median time for a trauma patient looked great, so the group couldn’t figure out why we had an issue. So we used Minitab to see the distribution, and it was a nonnormal distribution that was much different than just a bell curve. We saw that the median was actually a bit misleading—it didn’t tell the whole story, and that highlighted the problem nicely: the distribution revealed a tail of patients who were a lot worse when they stayed in the emergency department for over six hours, so we knew to focus on this long tail instead of on the median. Looking at the data this way let us see something we didn’t see before.

In another case, a hospital was so busy it didn’t have any beds, and it kept going on what’s called “diversion”. We didn’t want to divert sick patients away from the hospital, so we were trying to figure out why this kept happening. The first issue was that it would take us six hours to decide whether to even go on diversion. Meanwhile, the place was in chaos, so we collected data on each of the things that we thought might contribute, and one of them included which emergency department provider was on duty. We put it all into a regression model and found that the only significant factors associated with going on diversion were lack of staff and ICU beds. That analysis allowed the organization to concentrate its energy on those two factors instead of the 15 that they were trying to go after. When those two issues got fixed, sure enough, the diversion factor disappeared—it just went away.

Minitab: Do you anticipate any future trends for QI in healthcare?

Kashmer: Quality is being focused on more and more by the third-party payers, who reimburse hospitals and physicians for their care. Healthcare is just starting to learn how to improve quality. We’re kind of casting out in the dark for tools that already exist in statistical process control. Of course, that requires education and specialized knowledge that we as physicians typically don’t have, and that’s part of why these tools are not used as often as they could be. It’s been fascinating to learn about the applications of Lean and Six Sigma and then learn the software that can be associated with it, like Minitab. Once you’ve been doing this for 10 years, the benefits become very obvious. You can’t help but wonder why you haven’t been doing it all along.

Minitab: Do you have any closing thoughts about statistical process control and QI as it applies to healthcare?

Kashmer: Providers should know that quality is already here—it’s already in many of your third-party payer’s contract, and that has a huge impact on reimbursements. These robust tools show us how our changes to systems perform, and how we can improve them. But most important, once you know there is a set of tools out there that can help people, you have a responsibility to use them—understanding your data has a huge impact on your patients at the end of the day.

By using this site you agree to the use of cookies for analytics and personalized content in accordance with our Policy.

OK