Here's a brutal statistic: A recent article from the American Hospital Association found that nearly 90% of patients who switched providers said they did so because the organization was "difficult to do business with."
They didn't leave because of clinical outcomes. They left because of friction. When nearly 90% of your patient churn is caused by friction, that's not a leak, it's a flood. But what does "difficult to do business with" actually mean? It could mean:
- Being asked the same medical history questions they just answered online
- Being offered an earlier appointment when the new appointment date would be too early for insurance or clinical reasons
- Not being able to find the podiatry department, resulting in a long and painful walk
Would you have guessed those were driving your churn? I’m guessing not, because many of us who work in health IT or consumer experience leadership are not normal patients. We know how to work the system, how the sausage gets made, how things really work under the hood. We shouldn't assume that we know what the average patient needs. If we want to reduce patient churn, we should stop designing healthcare experiences based on our insider knowledge and start listening to what patients are actually telling us every day.
What does listening to the patient look like?
I recently had a surprising conversation with an administrator of a 50-bed rural hospital. When a bad storm knocked out the satellite signal during a big college football game, the team knew exactly how to react. Why? Because their administrators had learned through patient rounding how much those games mattered to patients. So, they quickly scrambled devices with streaming services to patients' rooms. Problem solved = happy patients. The administrator also pays attention to small preferences — Coke vs. Pepsi — and makes sure the team remembers. These seemingly small gestures, uncovered simply by asking patients what they need, really add up when a person is already facing acute healthcare challenges.
Would you have guessed any of that? Would it have shown up in your satisfaction survey? This is what companies in restaurants, hospitality, and retail mean when they talk about customer obsession: paying attention to what customers need (or want), not what you assume they need.
The scale problem
Here's the challenge: most health systems can't have administrators personally round on every patient. A 50-bed rural hospital might be able to do this, but a system with 100,000 patient encounters a month simply cannot. Patient surveys can be useful, but that's asking patients to do yet another thing.
But here's what executives can miss: patients are already telling you all of this. Every single day. In call centers, at registration desks, through patient portal messages. Your health system generates thousands of these interactions where patients explicitly describe their friction points. QA staff might be spot-checking a handful of calls, but the rest of these complaints, questions, and frustrations just disappear. You're not facing a data collection problem. You're facing a listening problem.
The value is there. One executive I spoke with invested in a comprehensive manual review of three months of call center logs and discovered something striking. A significant number of calls weren't about complex medical issues or billing disputes. They were simple questions about basic appointment information: "Where do I park?" "Which entrance do I use?" "What floor is the cardiology department on?" These weren't questions that required a trained call center representative. They were navigation problems that could be solved with better wayfinding, clearer pre-appointment communication, or a simple text message the morning of the visit. The system was spending valuable resources answering the same basic questions hundreds of times a day. But they didn't realize it until they really dug deeper.
How today’s technology can help
Here's what I hear when I share the stories of the rural hospital or the comprehensive call log review: "I can't do that at scale." But this isn’t impossible. The trick is to realize that patients are already telling you what improvements to make!
Most health systems have patient communication data — call center logs, portal messages, chat transcripts — scattered across multiple systems, and once combined, it’s an absolute gold mine. Once you have connected these disparate sources into a unified view, a properly trained AI can separate the actionable insights from the noise.
What a QA team does manually with a small sample of calls can now be done across every interaction. When you can analyze six months of patient communications and surface the actual friction points (not just the ones you assumed existed), you can build interventions that address tangible problems. Solving these quickly leads to improved patient experiences, less churn, and reduced workloads for staff. Win-win-win.
The rural hospital administrator who rounds on patients is modeling what all of us should be doing: listening with enough humility to be surprised by what patients actually need —and then acting on it. The question is whether the rest of us are willing to build the systems that let us do the same, and do so at scale.
If you're wrestling with how to start listening at scale in your organization, let's talk about what your data are already telling you.
