Introduction 

The need to “learn” from clinical incidents was a motivating principle behind the global adoption of patient safety reporting systems in the early 2000s. Of course, just knowing what failures were occurring was not going to be enough - we also needed to ably answer the ‘why’ and ‘how’ questions, and thereafter implement meaningful strategies to manage risk. The first generation of patient safety models (like Reason’s Swiss Cheese Model) and investigation/analysis tools (like RCA) came to be adopted without much debate. Once in place, however, these models and practices became gradually enmeshed with healthcare’s core beliefs about safety and harm, what types of lessons could be learnt from failure events and what actions should logically follow. Unfortunately, the dividends from these early efforts, certainly in terms of its impact on the rates of harm, were uninspiring to say the least.

Mitchell and colleagues, in a 2016 paper in BMJ Quality and Safety¹ , invited prominent patient safety leaders (mostly from the United States of America) to share their reflections on the movement’s journey, fifteen years after the publication of the seminal ‘To Err is Human’ report. Among other observations, respondents voiced broad disappointment at the lack of progress against patient safety targets, yet expressed almost universal agreement that advancing our capacity to learn from adverse events remained a central focus. In other words, despite our best efforts, we were failing at our core aim of learning from harm. Another decade has nearly passed since then, and yet the story is the much the same. So what is the problem?

One theory is that the issue is not with our model of learning per se, but that the solutions to the problem are less obvious than initially anticipated. Thomas Edison is said to have once remarked on his quest to invent a commercially viable incandescent lightbulb: “I have not failed. I have just found 10,000 ways that won’t work!” While persistence is certainly a helpful quality when dealing with complex problems, we might also wish to learn from Edison’s capacity to harness and apply new thinking when needed. Thirty years into a journey of learning from clinical incidents, perhaps the time is ripe for a rethink.

In this issue of The Human Stream, we continue our exploration of ‘learning’ in the context of patient safety. This issue is the first of two parts on the topic of learning from harm (before we move on to learning from success). This issue highlights how the established model of patient safety naturally predisposes us to think about failure in predictable ways, in turn leading us repeatedly down low-yield paths of learning. In the next issue we look at alternative ways to learn from failure and strategies to enhance our capacity to learn higher-yield lessons to benefit patient safety.

Six barriers to effective learning

In my work with safety professionals and healthcare leaders, I am always attentive to how colleagues seem to make sense of the world around them and the various challenges they have to routinely navigate. What practitioners say, what they emphasise and focus on can offer a window into much deeper patterns of thought, sometimes revealing underlying assumptions that they themselves might not be aware of.

In many instances, it can be helpful to surface such assumptions, especially when they get in the way of true learning. Below is an attempt to compile a short list of barriers (or traps) that I see teams commonly struggle with.