Current Issue Pic Top Harvard Health Policy Review Archives Top
Current Issue Pic Middle About Us Fill Archives Bottom
Contact Us Fall 2000; Volume 1, Number 1
In Focus

Improving Patient Safety (cont.)
page 1 | page 2 | page 3

That brings us to the second system level. The first is the work system—the car itself. The second is the management system—the one that designs and creates the car in the first place, maintains it when needed, and, most importantly, redesigns and changes it when the old one is not good enough any more.

If we want to achieve new levels of patient safety, we must change at two levels. We must change the way we do our work—to a work process that is safer; and, we must change the way we manage that system of work—so that it, itself, can change and become safer.

The First Level of System: the System of Work—Safe Medication Designs
Now, at the first level—the Ferrari level—we know a lot. In my own organization, the Institute for Healthcare Improvement, we have formed and managed three separate national collaboratives of health care organizations trying to reduce error rates and to improve patient safety dramatically in less than one year.

Our Patient Safety Collaboratives have now involved over 150 organizations, and we have learned a great deal about the difference between a Windstar and a Ferrari, when it comes to improving patient safety. My close colleague, Tom Nolan, has been working with Harvard Professor Lucian Leape, other IHI Faculty, and the participating teams in the Patient Safety Collaboratives to help specify the design of a safe medication system, as one example of a system relevant to patient safety. I will describe very briefly a few of the ideas that they have come up with so far.

At our current state of knowledge, we divide the attributes of a safe medication system into three categories:

  1. Prevent errors when possible (make them less likely);
  2. Make errors more visible when they do occur; and
  3. Mitigate the effects of errors when they do occur and reach the patient.
Note, first, that all three elements must be present in the "Ferrari" system. The reason is quite simple: no system is perfect. The search for zero error rates is doomed from the start. Every new technology, even one whose sole purpose is to prevent errors, introduces its own, new forms of error. Human beings, even when they are very careful, have natural limits to their performance: limits of memory, fatigue, and vigilance, for example. All modern safety systems accept that some human errors will occur, and, even while they seek to reduce the intrinsic rate of error, they have the additional aims of making errors visible and of mitigating, blunting, and recovering from errors that do occur. The proper name of our quest for patient safety is not just to reduce errors, but to reduce errors and mitigate their effects.

Prevention of errors—the first level of defense—is most effective when it is informed by knowledge of the causes of errors in the first place. The most powerful cause is, at bottom, complexity. Complex systems break down more often than simple ones.

The statistics are quite simple. Imagine a system with, say, 25 elements, each of which functions properly—no errors㭟% of the time. If the errors in each element occur independently of each other, then the probability that the entire system of 25 elements will function correctly is (0.99)25 or about 0.78. With 50 elements, the probability is 0.61; with 100 elements, it is 0.37. Make the reliability of each element higher, say 0.999, and the overall success rates are 0.98 for 25 elements, 0.95 for 50 elements, and 0.90 for 100 elements. We can, indeed, improve the reliability of a system by perfecting its parts and hand-offs, but reducing complexity is even more powerful.

Using complexity, and its reduction, as design themes, here are some specific changes that have reduced errors in the IHI's Patient Safety Collaboratives.

Reducing the Probability of Medication Errors

  1. Use "Computerized Physician Order Entry" systems for medication systems to reduce handoffs, and to provide instant information on drug interactions, allergies, and prescribing errors;
  2. Limit hospital formularies to essential drugs and doses, with few duplications. Continually eliminate hazardous drugs as safer alternatives become available;
  3. Have clinical pharmacists work in patient care units and periodically join in rounds to smooth information flow;
  4. At change of shift, have nurses brief each other on circumstances that increase risks of error, such as unusual patients, unfamiliar attending physicians, or unusual chemotherapy regimens, and on ways to reduce those risks;
  5. Remove high-risk medications, such as concentrated electrolyte solutions, from patient care areas, and label high-risk drugs to indicate their danger;
  6. Remove or differentiate look-alike drugs and packages;
  7. Move certain tasks, such as calculating, drawing up, and admixing doses, to the pharmacy or to the drug manufacturer;
  8. Standardize equipment and supplies, such as intravenous pumps, across all units;
  9. Involve patients as active partners in checks, such as identifying themselves, assessing drug choices and doses, and reviewing allergy information.
After reducing the probability of error, the next line of defense is to make sure that errors are visible when they do occur. Modern, complex systems often make their own "state" invisible to the user, and so errors go unnoticed that, in a simpler system, would be obvious. This invisibility was one of the causal factors in the Three Mile Island nuclear power plant accident in the U.S. A modern approach to errors makes them noticeable as quickly as possible, so that people can provide a safety net to intercept them before they cause harm.

Making Medication Errors Visible

  • In order-entry systems, provide screens to check against rules, such as dosage limits, allergy checks, and drug interaction alerts;
  • Have a pharmacist review all orders before dispensing;
  • Use bar coding to detect errors at the point of administration;
  • Explain medications and their purposes to patients before administration;
  • Use "hearback" for oral orders and instructions;
  • Use proper monitoring systems—such as for laboratory tests and vital signs—for patients receiving hazardous drugs;
  • Use double-checks sparingly. When using double-checks, make the process serious and truly independent. The third line of defense, after prevention and visibility, is error-mitigation, the process of reversing or recovering from an error that actually does reach the patient. Once error rates are very low, it becomes increasingly difficult to make additional reductions in rate, and improving recovery processes becomes a more and more important component of the complete safety system.

    Mitigating the Effects of Medication Errors that Reach the Patient

    1. Keep antidotes for high-risk drugs closely on hand and updated at the point of administration;
    2. Standardize and train through simulation in procedures for rapid response to adverse events, such as anaphylaxis;
    3. Avoid giving potentially lethal medication in bolus form;
    4. Assure that, in failures, equipment defaults to the least harmful modes. For example, intravenous administration devices should default in failure to cut-offs, not to free-flowing mode.
    By careful, shared, ongoing study both of safety-improvement efforts in health care systems, and by learning from successful designs from human factors engineers and cognitive psychologists in other industries, anyone interested can add nearly endlessly to the list of interesting and plausible ideas for redesigns that improve safety. The above list is only a suggestion of some places to start. What we really need over the next few years is to clarify such designs for various hazardous systems in health care, such as medication administration, for organizations of different levels of size, wealth, and task. What, for example, is the best known safe medication system for a nursing home? A small community hospital? A large teaching hospital? If we define and address this as a serious research agenda in health care for as few as five years, we make tremendous progress.

    The Second Level of System: the System that Designs and Manages the System of Work—a Study from NASA
    Sadly, however, design of the work system is only half the battle. There is a system above the work system that must be equally carefully designed—the management system—and, compared to the work system, it may need even more overhauling in American health care if we are to reach the safety goals we ought to.

    I learned this lesson early in my journey into quality as my obsession. It came to me first from a man named Guy Cohen, at the U.S. National Aeronautics and Space Administration (NASA).

    I am sure that you recall the accident involving the Space Shuttle Challenger in 1986. The loss of life was, of course, tragic, but behind the tragedy were serious managerial and system breakdowns that led to that disaster. Those breakdowns occurred in an organization, NASA, that, only five years earlier, had been at the very pinnacle in the world in the management of quality, respected by all, and breaking new ground in the total management of quality at the world-class level. When President John F. Kennedy first set traveling to the moon as a national goal, NASA commissioned a consultant report on the probability of success. The estimated chances of success were less than 2%. NASA threw the report out, and managed to a level of reliability and excellence in a complex organization never before seen on the planet.

    Guy Cohen was part of that story. By the time I phoned him, he was about to retire as the Director of Quality, Safety, and Reliability at NASA. I called him first when I was the Vice President for Quality-of-Care Measurement at the Harvard Community Health Plan, searching for hints about how to help improve care faster than I had been able to. To my surprise, Guy immediately understood what I was asking, and 48 hours later, he was in my office in Boston, alone with me, with about 200 overheads and what turned out to be a five-hour one-to-one teaching session on how NASA, at its height, made quality real.

    page 1 | page 2 | page 3

  • Subscribe
    Fall 2000, Volume 1, Number 1
    Table of Contents
    Editor's Note
    Features: Election 2000
    Health Highlights
    In Focus
    Glossary of Health Care Terms

    about us | links | contact us | subscribe | epihc