Data answers “what happened.” It rarely answers “what should we do about it.”
Data Tells You What Happened
Healthcare generates more data than most industries. Patient outcomes, operational throughput, financial performance, quality metrics, patient satisfaction scores: all of it gets captured, reported, and reviewed. Most health systems have more data than they have time to act on.
Eric Cecava came up through business intelligence before he ran operations, which means he’s spent a lot of time on both sides of that relationship. He’s built the dashboards. He’s also been the person expected to make decisions from them. The gap between those two roles is larger than most people realize from either side.
Data tells you what happened. It can tell you where performance declined, which units are underperforming, how a patient population shifted. What it can’t tell you is why those things happened, or what the right response is. Those questions require judgment, and judgment requires something that data doesn’t contain.
The Question Data Can’t Answer
The question data can’t answer is: what is this situation actually about? A drop in patient satisfaction scores might be about staffing. It might be about a specific unit’s culture. It might be about a process breakdown that nobody has named yet. The data tells you something changed. It doesn’t tell you what changed, or whether the thing that changed is the thing you should fix.
Eric Cecava saw this repeatedly in operational reviews. A metric would decline, and the first instinct was to address the metric directly: change the survey, retrain the staff on patient communication, add a step to the discharge process. Sometimes that worked. More often, the metric was downstream of something structural that the data couldn’t surface.
The structural thing was usually a conversation that hadn’t happened, a decision that had been deferred, or an accountability gap that everyone knew about and nobody had addressed. Data doesn’t show you what people aren’t talking about. You find that out by being in the room.
How to Hold Both
The approach Eric Cecava developed over time was to use data to identify where to look and to use conversation to understand what he was looking at. The data points to the unit, the time period, the patient population where something is off. The conversation tells you what’s actually happening there.
This requires a different relationship with data than most organizations have. Instead of treating the data as the answer, you treat it as the question. The answer comes from the operational work: walking the floor, talking to the people doing the work, asking what’s in the way.
Data-driven decision-making is not wrong. It’s incomplete. The word “driven” implies that data is doing the steering, when the most it can do is point the direction. Someone still has to decide whether to turn.
What That Looks Like in Practice
In practice, it means building time into the operational rhythm for sense-making, not just reporting. It means creating space where the people closest to the work can tell you what the numbers aren’t capturing. It means resisting the pressure to act on a metric before you understand what’s behind it.
Some of the most expensive operational decisions Eric Cecava has seen were made in response to a data point, fast, without the conversational work that would have clarified what the data actually meant. The fix addressed the symptom. The underlying problem continued, and showed up again six months later under a different metric.
Data is necessary. It’s not sufficient. The decision you can’t make with data alone is usually the most important one.