Circulation 60,475 • Volume 15, No. 4 • Winter 2000

Critical Incidents, Anesthesia Safety And Record Keeping

Jeffrey B. Cooper, PhD

This APSF program is dedicated to consideration of automated record keeping for perioperative care. This is not a new issue, but maturation of the technology and forces in the arena of patient safety in all of healthcare may be creating an environment that is now ready and willing to accept this concept. In this introduction I address the basic questions of why we need record keeping and why the time is ripe for it. This isn’t an exhaustive review of record keeping—it’s more of a pep talk for the presentations to follow, which will address what data need to be captured, how they will be captured, what technological issues remain to be resolved and how data will be used.

To start, I ask you to consider if anesthesia is now safe enough. Should we stop trying to make it safer? There are still no credible data to answer these questions. We still don’t know what are the rates of mortality and serious morbidity related to anesthesia. We perceive that malpractice rates have dropped more in anesthesia than in other specialties. There are some data to suggest that. There are other specialties that have done well according to this measure, but probably not as well as anesthesia. There are other data from which people argue that, for healthy, elective patients, the rate of perioperative mortality where anesthesia was the primary cause is on the order of 1:200,000. And, the high profile anesthesia accidents that make the news seem to be relatively infrequent. Maybe things aren’t so bad. So, what’s the problem?

To begin, the 1:200,000 figure is probably not accurate. I don’t think we know what it really is. And, even if 1:200,000 healthy elective patients died directly from anesthesia, that would be too much. Further, if you work in the safety business, you know it’s a never-ending battle. When things are going well, somebody wants to cut some corners—safety is usually the first thing to go. Or, when things aren’t going well for the bottom line, safety measures are the first things to take the heat. That’s because lack of safety is hard to see unless there is a catastrophic event. The production pressure in anesthesia is real and increasing—anyone working in a hospital knows that. There are increasing numbers of anesthetics being administered outside of well-regulated facilities. This is further challenging the safety systems that have evolved in anesthesia. And, inside the hospital walls, anesthesia in out-of-the way places is increasing. Many locations are not well designed for anesthesia and are far away from the assistance needed in an emergency.

So, considering the continuing rate of adverse outcomes and the relentless pressure working against safety, there is still a problem. Yet, we still do not have the means to learn from events that would help to identify solutions. Automated record keeping or the complete, electronic perioperative medical record, whatever name is used, is one element of the solution. People have been working on this for decades. Companies have come and gone already. While the technology to do it is here, the problem is not yet solved completely.

We also have to look at safety in the context of what’s happening in the rest of healthcare. Anesthesia isn’t an island. The Institute of Medicine Report, “To Err is Human”, released in November, 1999, placed a spotlight on patient safety for all of healthcare.1 While some of us have been working in patient safety in anesthesia for as long ago as 25 years and the APSF was founded 15 years ago, the rest of healthcare got the wakeup call from this blockbuster IOM report. Now, each domain in healthcare, not just the medical specialties, must learn as did anesthesia about how to implement patient safety in its sphere of influence. What can be learned from anesthesia’s experience? While there still remains much to do to further improve patient safety in anesthesia and perioperative care, much has been learned that can be used to help other domains address their patient safety issues. Anesthesia can teach how to use technology wisely toward safety, how to gather and use qualitative data from incident studies and how to change a culture towards one that is merely process driven to one that is safety driven.

My own introduction to patient safety was in the 1970’s, when we studied critical incidents in a group of hospitals in Boston. We learned qualitatively, by gathering information from self reports of critical events. Yet, perhaps what we learned most from those studies is how much we don’t know. While people can give reports about events, there’s so much information missing because memory is incomplete and often inaccurate. While our studies in the 70’s helped draw the big picture of error in anesthesia, there’s much more that was missing. Beyond that, I draw from my experience from investigating specific events over the past 20 years, learning in the process how often we were missing critical information to understand what was the cause of an adverse outcome or a close call. Certainly, anyone who has tried to use a written record to reconstruct an event knows how much information is missing. As important, on the occasions where people have actually had some electronic record to look at, it becomes clear how much additional information is added to any investigation of a critical event. From our experience in qualitative data, and what I hope will be our experience in quantitative data, anesthesia has a lot to teach the healthcare world that is now prepared to take on the challenge of patient safety. This is a great time for those who work in anesthesia to take the next step in developing the next learning tools and to teach other specialties how to do it. This is a pivotal moment for patient safety and for using the electronic record as a tool for progress.

There are two good safety arguments for why we need the so-called “black box”, automated record, flight recorder, data management system, the electronic anesthesia record, or whatever you want to call a system that records information automatically.

1. We need it to learn what happened and why after near misses and accidents. An automated record will provide information data that is usually not available or not accurate from either a paper record or from memory. There are several studies that show this to be so. Inevitably, someone will bring up the argument that this is dangerous data to have. Yet, more often than not, the data will defend cases for the clinicians involved. Having the data and really knowing what happened will be more useful for the people providing the care. This just should no longer be an argument against these systems.

2. The electronic medical record within the perioperative process will provide a database of physiologic and outcome data that can be used for hypothesis testing about global safety issues and for comparisons between institutions. It will also allow “data mining”, i.e., looking for answers to questions posed retrospectively. Whether data mining will be useful is questionable, but, if available, there are those who will use it in that way, and perhaps something good will come of it.

There is also a financial justification for automated record keeping. While some have calculated a return on the investment (ROI) for such systems, I’m not sure we can really do that with any degree of certainty. And, it likely depends a great deal on local conditions. But, hospitals have already made the commitment to electronic medical records and drug order entry. An automated anesthesia record is just an extension of that. Because the perioperative process is one of the most acute components of hospital care and health care, it’s only natural to have the electronic medical record in the operating room and in those areas connected to it in the patient care process. And, that’s much easier and more effective to do now because network infrastructures in hospitals have now matured. Five or ten years ago, there was not a robust, diffuse network to connect with in most hospitals. It wasn’t possible to download easily from labs or demographic data. That is now a fairly solved problem. It’s one critical barrier that has been removed.

The question of adopting electronic record keeping has an answer based more on concepts of sound process management: to run a modern factory, which hospitals are in many respects, you have to keep track of the product and its parts, etc. That means keeping track of where patients are and what’s happening to them throughout the entire period of care. It’s the best way to keep the production moving smoothly. While that’s not as good a reason as safety, it is such reasoning that is driving the automated record and is going to make it something that hospitals must have. It’s just time to get on with the business of doing it. It still won’t be easy. There are barriers that remain and problems to solve. The technology is not yet optimal, but it’s good enough for everyone to get started. We’ll learn more about the specific issues in the presentations to follow.


1. Kohn LT, Corrigan JM, Donaldson MS. To Err is Human: Building a Safer Health System. Washington DC: National Academy Press. 1999. ISBN 0-309-06837-1. Available at: