In the West, viewed over the long-term of decades, we have seen medical procedures become safer and more effective. Increased safety and effectiveness are often thought to have come exclusively as a result of improvements in medical technology. However, each improvement in medical technology brings with it its own new and interesting risks. Also, modern technical technology is not always available, or affordable in many medical situations. The reliance on improvements in technology to bring improvements in safety obscures much of the work done in understanding individual and organizational error, and in developing quality management systems. Significant improvements in safety and effectives can also result from better safety management systems and education.

In the related arena of aviation, much has been made of safety management systems. The FAA Air Transportation Oversight System that I have worked with evaluates the design of systems and the performance of those systems using six safety attributes:

• Procedures—documented methods to accomplish a process.
• Controls—checks and restraints designed into a process to ensure a desired result.
• Process measures—used to validate a process and identify problems or potential problems in order to correct them.
• Interfaces—interactions between processes that must be managed in order to ensure desired outcomes
• Responsibility—a clearly identifiable, qualified, and knowledgeable person who is accountable for the quality of a process.
• Authority—a clearly identifiable, qualified, and knowledgeable person who has the authority to set up and change a process.

Loosely translated to the medical arena, this means that the systems and procedures established to ensure safety should be well designed to achieve the desired end, and also possess built in controls (or defenses) to ensure that the systems and procedures operate as planned. The systems must be objectively checked to ensure that they are, in fact, in place and proving effective. Objective checking involves oversight by third parties as well as self-reporting of performance measure. Risk management is a third part of this. Safety can never mean freedom from all risk, but hazards and risks can, and should, be identified and managed appropriately. For more on safety management, the reader is referred to MIL-STD-882 as revised.

The previous paragraph focused on safety management, but that it not the whole story in operating a safe system. It is a feature of human nature that we focus on bad news – the accidents that were not prevented. These are the ones that make it on the network news, but there are every day dangerous situations that do not become accidents because of the quick thinking of person on the scene. It is likely that there are more of these accidents prevented by local heroism in the medical arena than there are in the aviation domain, but neither are they totally absent there.

To improve safety, therefore, we need to know more about how to design effective safety systems, and we need to know more about individual decision making. Acting as a backdrop to all of this are the actual cases that we are attempting to make safer. A flight in an airplane may appear fairly routine to most passengers, but in fact, NTSB accident reports bear out that some of the most dangerous parts of a flight are the initial ascent, and the descent. Similarly, while admitting the truth of Bjarne Stroustrup’s words that “Proof by Analogy is Fraud,” it would seem likely that some of the most dangerous parts of a medical case are the period of admission, where not all the factors impacting the case may be known, and the discharge, where a patient is left to carry out instructions regarding dosing and self-monitoring. Clearly there are other periods of increased risk, but I want to focus on just these two: Initial Admission, and interim discharge.