The culture of safety

Our understanding of human error and medical errors was greatly enhanced by the work of Professor James Reason, whose seminal book, Human Error, provided both a taxonomy of human errors as well as a deeper understanding as to how they occurred, and why evaluation of the systems in which the errors occur are so important if we are to prevent future errors.

Figure 1: Factors that lead to clinical error (Click to enlarge)
Figure 1: Factors that lead to clinical error (Click to enlarge)
Often events far removed from the point of care may have a profound influence on the events at the point of care. We often term the events at the point of care as events at the “sharp end” of care, and those far removed at the “blunt end” of care (Figure 1). For example, in a hospital setting where physicians and nurses have very busy, stressful clinical shifts, if an administrator significantly lengthens the duration of their shifts, this action at the “blunt end” of care will frequently increase the likelihood that the health care providers will become more error prone and make errors at the “sharp end” of care.

There are many points at which defective systems of care can influence the likelihood of errors. Examples of systems of care adversely affecting patient safety include poor lighting in areas where providers write orders, or failure to utilize computerized physician order entry systems (CPOE). CPOE systems have been well-documented to reduce certain types of errors in care, particularly regarding medication prescribing and administration.

Systems issues in handling medical information in diabetes care

Many medical systems have converted, or are in the process of converting their clinical patient data from paper-based systems to electronic health records (EHR) and attempting, with different levels of success, to move key relevant information from provider to provider and storing selected data from large populations in increasingly sophisticated clinical registries. Modern diabetes care benefits greatly from efforts to share relevant clinical data rapidly. Data that is needed at the point-of-care must be accurate and timely, or serious errors and injuries may result. In modern diabetes clinics, however, flaws in providing the most accurate, complete, and timely information at the point of care are often most evident during transitions of care.

A transition of care occurs every time a person with diabetes moves from one clinical setting to another, such as from an intensive care unit to a more general hospital ward, or from the hospital to home, or from one hospital unit to another. Often, the essential information needed to continue the care gets truncated or completely lost, and errors in care result. Transitions of care also occur when key members of the clinical team change, even if the location of care is the same, and the changing of shifts for both physicians and nurses may also create a potential defective “hand-off” of information leading to error because of the defect in the transition of care.

Systems of care need to be designed so that special attention is paid to transitions of care, which, overall, is one of the times when errors are most likely. Check lists, medication reconciliation, and the development of transition-of-care documents are especially important in the care of patients with diabetes, and are a feature of the safer systems of care.

Electronic Health Records (EHR) and Computerized Physician Order Entry (CPOE)

Although electronic health records have been, in general, a major technological advance which is a most helpful tool in diabetes care, there are many problems which have not yet been satisfactorily corrected. The most fundamental need is that the data in the EHR must be correct, and yet, less attention is given to making sure the data sources are accurate and inputted correctly. Both historical information and data from a physician examination are complex and highly subject to errors by both patient and provider, particularly regarding details on family history, history of comorbid conditions, examination of the foot, etc. Yet, once faulty data is entered into an EHR, they may be recopied many times and the errors are propagated without correction. Medication reconciliation is particularly high risk for error.

Laboratory data also may be inaccurate, not only because of data entry issues but also because laboratory values done with different instrumentation may not be comparable, yet may be uncritically compared to other laboratory values with very different ranges of normal. Data entry issues in some EHR systems with poor oversight can result in a large fraction of values being incorrect, even misleading.

The timely availability of key data is very important in diabetes care. A glucose level reported 30 minutes too late may not be relevant anymore. Particularly during transitions of care, the key data must be transmitted both timely and accurately, and the interfaces that allow for sharing of EHR data must be carefully designed and maintained.

Computerized physician order entry (CPOE) has the promise to reduce errors of transcription and transfer of orders. Many errors due to the ambiguity of written orders can be eliminated. But CPOE is vulnerable to other problems as well. The provider’s view of the orders for the patient and their stage of completion is not always optimal, and the sheer amount of detail may obscure key points in the planned orders. CPOE’s which contain less intuitive, more complex order sequences may result in clinician errors. The clinicians who use CPOE systems should regularly review the CPOE for usability and train personnel well as to how to deal with any more difficult properties of the system they use. Programmers should work closely with clinicians to optimize their system.

There are many other points at which systems of care need to be specially designed to meet the needs of our patients with diabetes. For example, if a hospital does not have a robust plan to maintain the accuracy of the point-of-care (POC) glucose meters, then over a period of weeks or months deterioration of the accuracy of the meters used in the hospital may result in inaccurate glucose meters being used in critical situations. Also, patients with diabetes who are provided POC glucose meters with displays that are too small or too dimly lit to be seen well will also be at risk for errors in self-care because they misread the display on their meter and make incorrect decisions.

“Culture of Safety”

The term “culture of safety” is used to describe a clinical setting where a collection of individuals providing care work together in a collaborative way in order to promote patient safety. There are abundant data which show the importance of well-functioning nonhierarchical teams, from both medical systems of care and many other systems where teamwork is key, as for example, from aviation safety studies and other high-risk settings such as nuclear power plants. How well people work with each other, how frequently they take a nonjudgmental approach to safety issues, how regularly and systematically they check each other’s work, all these parameters are valid measures with predictive value as to how effective the team will be in reducing errors, both in routine and in emergent situations. For diabetic patients, since the central therapy is often insulin, which is considered a high-alert medication with a narrow margin of safety, it is most important to develop a “culture of safety”. Also, it is important to remember that a “culture of safety” is nearly always more effective when the administration of the inpatient or outpatient facility strongly supports such an effort.

Swiss-Cheese Model of a safe system

Figure 2: Conceptual model for safer systems (Click to enlarge)
Figure 2: Conceptual model for safer systems (Click to enlarge)
Reason had developed a conceptual model (Figure 2) which has been widely used and modified to illustrate the concept of developing redundancy in the safety measures. He points out that nearly all plans to prevent errors have some latent flaws, but when the defenses against errors are placed in sequence, it reduces the likelihood that an injurious error will occur.

Reason has used the term “error traps” for systems of care where the vulnerabilities of the system are so powerful that it can be expected that capable providers of care, placed in the same situation, will often make the same mistake as the last person. Organizations in which “error traps” occur are more commonly ones where there is not a “culture of safety”, but rather, more hierarchical systems where people who make errors are punished severely and education of medical and nursing staff regarding patient safety issues is discouraged.

In general, whether the system is a diabetic clinic, a hospital ward, a nuclear reactor or an airline, if error prevention is not taken seriously, latent errors are more likely to persist, “error traps” to develop, and catastrophic errors are more apt to occur and recur.

At the end of the day, a systemic approach to error reduction is a powerful way to improve patient safety. Continuous examination and re-examination of relevant data is key. Hospital or clinician data on glucose values, inspection of all adverse events, attention to education of patients, family, care providers, and others indirectly involved in the care of the patient, and creating and maintaining a “culture of safety” are all key to improving patient safety and reducing injurious medical errors.

Figure 3: Culture of safety and the scope of awareness (click to enlarge)
Figure 3: Culture of safety and the scope of awareness (click to enlarge)
Figure 4: Culture of safety and the scope of awareness in the system under stress. (Click to enlarge)
Figure 4: Culture of safety and the scope of awareness in the system under stress. (Click to enlarge)
In the two diagrams (Figure 3 & 4), the concept of scope of awareness is introduced. No one can possibly see all of the direct and indirect consequences of a diagnostic or therapeutic action. The importance of a culture of safety can be seen in the first diagram. In the second diagram one can see how the effect of a system under stress, often in an effort to increase “productivity” or “efficiency”, may lead to a diminished scope of awareness and an increased risk for injurious medical errors.


Reason J. Human Error. Cambridge, UK: Cambridge University Press; 1990.

Hellman R. Patient safety and inpatient glycemic control: Translating concepts into action. Endocr Pract. 2006;12(Suppl 3):49-55

Bogner MS. Human Error in Medicine”. Hillsdale, NJ: L. Erlbaum Associates;1994

Hellman R. A systems approach to reducing errors in insulin therapy in the inpatient setting. Endocr Pract. 2004;10(Suppl 2):100-108.

Reason J. The Human Contribution: Unsafe Acts, Accidents and Heroic Recoveries. Farnham, Surrey, UK: Ashgate;2008

Bates DW et al.The Impact of Computerized Physician Order Entry on Medication Error Prevention. JAMIA. 1999;6:313-321

Schiff GD et al. Computerised physician order entry-related medication errors: analysis of reported errors and vulnerability testing of current systems. BMJ Qual Saf. 2015;24:264-271

IOM (Institute of Medicine). Health IT and patient safety: building safer systems for better care. Washington DC: The National Academies Press, 2012.


Nobody has commented on this article

Commenting is only available for registered Diapedia users. Please log in or register first.