To Measure Outcomes,Sweat the Details

If there is ever hope of delivering optimal care, we need to fine-tune all the different inputs that contribute to the final product. In our case, the final product is the best surgical outcome as we understand it scientifically and as our patients experience it. Sounds simple, but analyzing outcomes requires a bit of discipline to submit to a process for a specified period (usually months), substantial time investment, manpower allocation, a measure of humility and plenty of transparency. As most clinicians are already stretched for resources and time, it is easy to see how issues that do not immediately impact the short-term care of the patient often wait, and wait…and wait.

The time-tested DMAIC acronym of implementing quality and process improvement, often attributed to the mathematician/statistician W. Edwards Deming (1900–1993), remains a reliable way to think about our patient outcomes. It is not new to anyone in project management and manufacturing, but please indulge me with an ounce of patience as I briefly explain the DMAIC cycle.

• Define. Any problem should be clearly articulated. Is there a specific complication you are seeing more frequently? Is there a recurring patient complaint? This requires sharpening your questions and goals. A quest to reduce all wound infections is much more vague than asking if maintaining patient body temperature above 36°C during surgery reduces wound infections. At this stage, you should also determine the goal and timeline you have to reach your goal. This requires discipline to work realistically within your limitations. One cannot expect to change global healthcare expenditures based on a 6-month timeline and a US$1,000 budget. I would add to this step the need to deconstruct your entire process or operation into smaller sequential steps such as process maps or flow charts. By doing this, you can better locate where your variances occur. In practice, this is probably the most painstaking (but critical) part of DMAIC.

• Measure. Collecting data on what you have defined as the problem helps you take stock of how you are doing right now. It establishes the baseline to which any future improvements are to be compared. The more specifically you define your question or goal, the more precise and meaningful your measurements will be. Once you know what you are measuring, this step is relatively straightforward, albeit it may require time and manpower to capture the data. Again, knowing that you have a 6% wound infection rate is less meaningful than knowing that 60% of your wound infections occur in patients whose intraoperative body temperature was below 36°C.

• Analyze. This stage will usually require simpler quantitative tools such as descriptive statistics, histograms, Pareto charts, or some sophisticated statistics tools, as well as qualitative assessment for root-cause analysis such as the fishbone diagram or asking the “5-Whys.” The analysis helps detect areas of high standard deviation, a surrogate for variance, and informs the team about prioritizing resources for implementing changes.

• Improve. This stage allows the team to design process improvements and implement changes. Again, effective implementation requires simple directions for those who are following them, and making sure what you are capable of accomplishing does not outsize your most important resources—time and manpower. This action stage is often exciting and the zenith of any improvement process, but it is best to temper any enthusiasm because it is still a test, and any progress still requires validation.

• Control. This term is not as intuitive because it refers to steering your process so it does not veer off course. However, steps inherent in this stage include monitoring your new implementations, making adjustments, documenting them, and measuring again.

Improvement processes require the dedication of administration, involvement of all stakeholders including patients, and almost always need to be a line item in the budget because there is a cost. There are clear benefits to being systematic about process improvements. First, it can be transferred to other systems within your practice or organization. One of the clear advantages of using this quality improvement exercise in our bariatric program was seeing the skills attained transferred to our liver and colon surgery service lines. Being able to use the same language between teams also allows for other groups to test your findings, improve their process, and maybe improve yours further. Herein lies a major difference between manufacturing and healthcare. Whereas manufacturing best practices can easily be used for competitive advantage, I believe most healthcare systems are eager to share their best practices to anyone willing to listen.

In this issue, several groups are indeed sharing their best practices and findings of their own process improvements. Kamal et al. examine what it takes to get a new program going with early results, which really is the D-M aspect of DMAIC. Grimaldi and Aarts go directly to the patients and ask them questions in order to determine the drivers for patient choices after surgery. Chlysta et al. report outcomes of their venous thrombotic embolism prevention implementation. Hernandez's group has had great experience with certain fibrin sealants, and tests one more to see if outcomes improve. Rothwell and Kow* report on whether implementing a presurgical low calorie diet affects outcomes after adjustable banding surgery. These latter two studies are important reminders that noninferiority or even negative findings are equally important for patient care.

Finally, I draw your attention to a Proof of Concept paper from Chun et al. from the Republic of South Korea who have been working on ways to measure gastric volume for sleeve gastrectomies and other gastric reductive procedures. Obviously, they have defined the problem to be the variable gastric sizes found after surgery and hope to standardize functional gastric volumes for weight loss.