Human Error In Medicine Marilyn Sue Bogner

Why Human Error In Medicine Keeps Getting It Wrong

I spent years watching teams try to prevent mistakes using the same outdated methods that got us into trouble in the first place. The literature talks a lot about system factors, but the actual practice of applying those ideas is where things break down. Marilyn Sue Bogner's work on this is one of the more practical frameworks I've encountered, and it's also one of the most misunderstood. Bogner is a nursing researcher whose work focuses on how errors actually happen in clinical settings. She doesn't treat human error as a character flaw or a failure of attention. She treats it as a predictable outcome of systems designed by people who weren't there when things went wrong. Her research examines near-miss reporting, cognitive load in clinical decision-making, and why good professionals make bad calls under normal working conditions.

Understanding Human Error In Medicine Marilyn Sue Bogner

At the core of Bogner's approach is the distinction between lapses, mistakes, and violations, borrowed from broader human factors psychology but applied directly to clinical environments. A lapse is when you forget to do something you intended to do. A mistake is when you intend the wrong thing. A violation is when you knowingly break a rule or protocol. Most incident reports capture all three categories and treat them identically. That's the first mistake. What makes Bogner's framework different is that she spends more time studying near misses than actual harm events. The assumption is that for every injury that reaches a patient, there are dozens of errors that get caught before anything happens. The problem is that near-miss data is systematically underreported, and when it does get reported, the quality of the data is usually terrible because people report near misses the same way they report bad outcomes. They focus on themselves instead of the system. I worked with a medication administration system where nurses were logging near misses through a standard incident reporting tool. The volume was low, and the entries were thin. Something like twelve reports a month across a hundred-bed unit. That's not a reporting problem. That's a system design problem. I replaced the standard form with a structured near-miss log that asked three specific questions: what were you trying to do, what stopped you, and what would have happened if you hadn't stopped. The next month we had sixty-seven entries. The questions forced people to describe the system failure instead of their own failings.

The second insight most people miss is that Bogner's work doesn't actually say human error can be eliminated. She says it can be anticipated and contained. The difference matters because most institutions treat error reduction as a goal rather than a continuous process. You're going to have errors. The question is whether your system lets them accumulate into harm. There's a practical implication here that nobody wants to admit. Near-miss reporting systems tend to work best in environments where staff already trust that reporting won't be used against them. If your culture is punitive, Bogner's framework won't save you. You'll get the same low-volume, low-quality data regardless of how you structure the reporting tool. The workaround is to pair near-miss programs with explicit non-punishment language and then follow through on it. When I've seen this work, it typically takes three to six months of consistent non-retaliation before reporting rates climb to meaningful levels. Not because people suddenly care more about safety. Because they stop calculating the career risk of speaking up. One more counter-intuitive point. Bogner's research suggests that expertise doesn't reduce error rates the way most people assume. Experienced clinicians make different kinds of errors than novices. They rely more on pattern recognition and experience-based shortcuts, which means their mistakes tend to be systematic rather than random. A veteran nurse might miss a drug interaction because the presentation looks familiar, while a new grad would catch it by checking every field manually. Both are human errors. One is more dangerous because it's less likely to be flagged.

Training programs that focus on reducing novice errors through additional double-checks and protocols will miss the expert error problem entirely. The fix for that is different. It's decision aids, second-opinion requirements for high-risk decisions, and forcing pauses at specific points in the workflow. The problem is that these interventions feel inefficient to people who've developed speed through experience. They resist them. That resistance is itself a human error signal. If you're looking to implement anything based on this research, start with the near-miss data quality issue. Bad data will give you bad conclusions every time. Before you build dashboards or conduct root cause analyses, verify that your reporting mechanisms are actually capturing what they're supposed to. Spend a week reading every near-miss report your system generated and rate each one for usefulness on a simple scale. You'll probably find that less than half of them contain enough information to act on. That's not a staff problem. That's a tool problem. The research is available through nursing and patient safety journals. There's no single paper that covers everything. Bogner has published across several decades, and the near-miss work is particularly important but scattered. Look for her publications in the Journal of Advanced Nursing and works related to clinical decision-making and error reporting. The frameworks are more useful when you read them alongside the original human factors literature from Reason and Hollnagel rather than treating them as standalone theories.