Photo by Artur Matosyan | Unsplash
As the CEO of the Institute for Healthcare Improvement, it will surprise no one to learn that I believe in improvement science. I’ve spent the last 20 years studying and using the Model for Improvement and other improvement tools and methods. I’ve come to see that the quality sciences have the potential to create change in almost any system.
In fact, I would argue that improvement science can and should be used to address some of the biggest challenges facing health care today, including the COVID-19 pandemic and systemic racism. To come to this realization, I first had to have some of my deepest beliefs about improvement science and methodologies shaken to their core.
Conventional wisdom sometimes holds improvement science as almost above investigation or question. This is risky as it shifts improvement to dogma and away from science. Speaking from experience, improvement science is often seen as inherently apolitical and free from bias. And because we intend our improvements to result in better care for all and we do not consciously discriminate against anyone, we assume these improvements lead to equitable care.
I was compelled to rethink these assumptions about five years ago during a review of several major IHI projects. Evaluated by multiple different measures, these were some of IHI’s most successful improvement efforts. They got results and improved outcomes at scale. During the review, we asked a simple question: Did all populations (defined by race, gender, ethnicity, and language) experience the benefit equally?
The answer for most projects was that we didn’t know since we had not stratified the data by population. We simply did not know whether all populations were experiencing improvements, or if the improvements were further exacerbating underlying inequities. For projects with stratified data, there were some indications that improvements might not have benefited more disadvantaged or marginalized populations. We came to the disturbing realization that some projects improved care for those who were already better off, but not necessarily for those who needed our help most of all.
The experience was eye-opening. At IHI, we pride ourselves on improving the performance of systems. It felt like a big miss to recognize that we had neglected this fundamental feature — unjust variation across populations — of how systems are built and organized. We learned that without consciously designing for equity, we run the risk of reinforcing inequitable systems. It can happen to the best of organizations.
For example, a few years ago, a health system in the Midwest reviewed its data on colorectal cancer screening. The data showed that their overall population-level results were getting progressively better. However, prior to working with IHI, the health system hadn’t stratified the data to examine which patient populations were being helped by their improvement efforts.
When the health system looked more closely at the stratified data, there was steady growth in screenings among their White, non-Hispanic population. But screenings among their Hispanic and Black populations were staying the same or declining overall.
In other words, the data showed that the systematic application of improvement methodology was indeed improving the median performance of the system for colorectal cancer screening rates, but most of the change occurred in only one population segment. Their Latinx and African American patients were not experiencing the full benefits of this organization’s years of improvement work. In fact, the application of the science had widened the disparity for this outcome measure.
While this was not intentional, it was also not accidental. It was the result of a legacy system that has bias baked in. When you study your work and you find inequities — as this health system and IHI did — you see that improvement science, like all sciences, is not impartial.
Like many others in the last few months, the killing of George Floyd has compelled me to reexamine some of my most fundamental beliefs. I’ve been studying the origins of the science of improvement to better understand whether there are ways in which it has been historically deployed to maintain the status quo more often than to change it.
All sciences, including the science of improvement, can be leveraged by those in power for self-interest. For me, this understanding is both threatening and liberating because the methods that can be passively used to reinforce systems can also be used to systematically dismantle injustice. I now see that the science of improvement should be part of the dialogue about social change and improving human circumstance.
Consider the current pandemic, for example. With its clinical, financial, and operational challenges, COVID-19 has highlighted the racism structured into our systems. We've seen Black, Latinx, and indigenous populations endure a disproportionate impact from the virus. It’s logical to use the science of improvement to challenge the systems that produced the current state because this science is designed to eradicate variation — and health inequities result from systematic, unjust, and unwarranted variation.
At its heart, the science of improvement is a science of knowledge and learning. It can speed our ability to detect where systems are failing, and bring resources and better solutions to the most challenged and weakest parts of systems. It will take time, energy, and disciplined application, but I believe that the improvement methods we have developed over decades can help us learn our way into a more just and fundamentally different system.
Editor’s note: Look for more each month from IHI President and CEO Kedar Mate, MD, (@KedarMate) on improvement science, social justice, leadership, and improving health and health care worldwide.
You may also be interested in:
Innovation During a Public Health Crisis
HealthLeaders Media: New IHI Chief Executive: 'There Is No Quality Without Equity'