The People Margin

Healthcare leaders should face the inconvenient truth that humans may corrupt precision workflows

I hope you’re having a great week. I stumbled on this quote in my Obsidian database (my second brain) and got to thinking how this fits in with healthcare’s growing focus on process and quality. As leaders we have to think realistically about both the power and limitations of efficiency.

This is written in the ‘short form’ spirit of my old blog — a riff focused on a tension that I see in my world. I’d love your thoughts on this if you have a moment to comment.

---

Healthcare processes are key in helping us provide consistent quality care to patients. However, people are complex and often unpredictable, which means that data and algorithms can never fully account for the nuances of human behavior.

I love this concept of the people margin. It comes from Automattic engineer Mike Shelton in 2018.

Data can be precise, specific, absolute and is meant to represent the actions and behaviors of people and things. Yet, people themselves can be imprecise, abstract, non-linear, and unpredictable. I call this the people margin – data’s margin of error when applied to everyday life. Context matters. We intuitively modify our behaviors based on numerous inputs. These modifications often can’t be explained with data alone. Only when we apply context to our product based on actual people’s stories, can we create real experiences.

As he suggests, we all “encounter situations where we’d be unable to move the needle despite our data telling us we were right.“ Shelton wasn’t talking about healthcare but he could have been.

Industrialized healthcare positions patients as units in clinical workflows. Shaped around the legacy of industrialization, a model of care delivery has emerged based on conformity, standardization, and interoperability - for process, physician and patient. Of course, fixed processes are critical for many functions in a hospital, but may not work so well in others. As Shelton suggests,

Take people, for example. Trying to create a process involving humans as a fixed variable sounds reasonable, but it won’t work for long. We are unpredictable and no two of us are alike. We’re emotional, weird and inconsistent. We change and adapt on a dime. We modify our behavior based on complex inputs — mood, social situation — and these modifications may not always be easy to explain, quantify or predict.

All facts that corrupt the algorithm.

When designing health system processes it would be a good idea to consider the people margin and the context of the patient. And recognize humans for the messy wildcard that they are.

Photo by Timon Studler on Unsplash