People make mistakes all the time. To err is one of the well-recognized characteristics of being human. Alexander Pope is credited with having written, “To err is human, to forgive divine.” In our highly complex society, we might rightly expand on Pope to say, “To err is human; to really mess up takes a system with a convoluted user interface.” Unfortunately, in medical technology, as well as in other safety-critical systems, mistakes can lead to serious, often catastrophic consequences requiring much more than divine forgiveness.

As the interaction between humans and their machines became more complex, the science of human factors evolved to study the ways humans relate to the world around them. Human factors engineering (HFE) is a branch of engineering devoted to applying the science of human factors to the design, development, and deployment of systems and services. HFE is sometimes referred to as usability engineering and internationally as ergonomics. One objective of HFE is to improve both operational performance and safety.

An aspect of the HFE process is the development of a taxonomy, or system for classifying interactions between humans and machines. Understanding the taxonomy is important to appreciating how mistakes in using a medical device can contribute to medical error.

Medical errors of all kinds are often cited as a significant cause of death in the United States.¹ Accurate numbers are hard to come by because they depend on a reporting process that is frequently criticized for underreporting of adverse events. Also complicating the situation is uncertainty in ascertaining the root cause of error. However, there is general agreement that medical error is a serious problem in the United States. In the publication To Err Is Human: Building a Safer Health System, the Institutes of Medicine issued a call for reform in the healthcare system to address what it considered the nation’s “epidemic of medical errors.”

Taking up the call, FDA has increased its emphasis on usability in the design of medical devices, including those intended to be used for in vitro diagnosis. Having analyzed the event reports in the Medical Device Reporting system, FDA concluded that device use error is a significant but addressable cause of serious consequences for patients (illness, injury, or death). FDA sees that many use errors are not just random human error, but are often induced by the design of the device and its labeling. Through its Human Factors Program, FDA has strongly encouraged manufacturers to embrace the principles of HFE. FDA defines HFE as “the science and the methods used to make devices easier and safer to use.”² HFE takes into account how users interact with a device. It sets out a process for identifying the issues that may prevent the device from being used as intended.

For some reason, the formal application of HFE in the medical device sector has tended to lag behind some other sectors. From a standards point of view, some of the first formal work in this area was done by the Association for Medical Instrumentation (AAMI’s) Human Factors Engineering Committee. AAMI/HE48, Human factors engineering guidelines and preferred practices for the design of medical devices, was published in 1993. The focus of this document is on the ergonomic aspects of medical device design. The AAMI committee followed this up with the publication in 2001 of AAMI/HE74, Human factors design process for medical devices. The AAMI committee is close to publishing its opus on HFE for medical devices, AAMI/HE 75, Human factors engineering—Design of medical devices.

Using the AAMI work as a springboard, the International Electrotechnical Commission (IEC) and the International Standards Organization (ISO) jointly published IEC 62366, Medical devices—Application of usability engineering to medical devices, in 2007. IEC 62366 describes a usability engineering process and provides guidance on how to implement and execute that process to enhance the safety of medical devices.

Figure 1. (Click here to enlarge) Categories of foreseeable user action. In this figure, an action can result from a user choosing to do something or failing to do something. See Annex C of IEC 62366:2007 for lists of potential use errors and abnormal use or their causes. Nescient is used in the context of a lack of awareness of the adverse consequences of a skill-based action. Source: Figure B.1, Categories of foreseeable user action, from IEC 62366:2007. Copyright IEC, Geneva, Switzerland, and used by the author with permission.

At the center of the HFE process described in IEC 62366 is understanding how humans behave when faced with a task that requires the user and the medical device to interact. This is particularly important when the interaction occurs in an emergency or other stressful situation, or when the user is fatigued or uses the device infrequently.

This understanding enables the Human Factors (HF) Engineer to classify the causes of various types of errors that might be made using the device. Several standards offer systems, or taxonomies, for classifying error types. Figure 1 shows a taxonomy that was adapted for medical devices from work done by James Reason.³ This taxonomy is described in Annex B of IEC 62366:2007.

The understanding of the taxonomy begins with understanding the definition of the term use error. Use error is defined in IEC 62366 as an “act or omission of an act that results in a different medical device response than intended by the manufacturer or expected by the user.” There is a subtle but important distinction in the choice of the term “use error” over the more commonly used terms of “user error” or “human error.” It recognizes that not all errors arise because of carelessness or inattention on the part of the user. The term is intended to be “blame neutral.” It recognizes that an error may be the direct result of a user interface design that did not properly take into account the capabilities and limitations of the human beings who were required to interact with the medical device.

User actions (or inactions) can be broadly classified into those that are foreseeable and those that are not foreseeable. The interface designer can only deal with those things that are foreseeable. One of the roles of HF engineers in the design process is to apply their knowledge and the tools at their command to identify the ways things can potentially go wrong so that the interface designers can address them. Tools such as cognitive walk-throughs, contextual inquiry, functional and heuristic analyses, rapid prototyping, and testing in simulated environments and in the field help the HF engineer anticipate what the user will do when faced with carrying out a particular task.

The taxonomy in Figure 1 is based on classifying user actions. However, it can also be used to classify user inaction—the situation where the user should have done something but took no action.

If the user intended to take an action, the result will fall into one of the following three categories.

Correct use. This is operation of the medical device in accordance with the instructions for use or in accordance with generally accepted practice for those medical devices provided without instructions for use. In Figure 1, this is shown as “normal use.” However, it is important to remember that a use error can occur while a device is being used as intended. If a user error occurs even though the user is trying to use the device as intended, it is not correct use in this taxonomy.

Abnormal use. This is an intentional act or intentional omission of an act by the user as a result of conduct that is beyond any further reasonable risk control by the manufacturer of the medical device. Typically, abnormal uses are associated with use scenarios where there is no effective risk control measure that can prevent the use scenario. An example would be the use of an automated analyzer without checking calibration in violation of obvious warnings on the screen that calibration is to be checked. This and other examples that are based on actual events that were determined at the time to be instances of abnormal use are provided in Annex C of IEC 62366:2007. These examples were in turn taken from a paper on reporting of use errors prepared by Study Group 2 of the Global Harmonization Task Force (GHTF)⁴.

Abnormal use is often thought of in terms of a malevolent action or irresponsible use. However, it need not be. Abnormal use is simply using a medical device for a purpose or in a way other than those intended by its manufacturer, often in violation of clear warnings or contraindications. It may or may not have adverse consequences. Abnormal use is not considered a use error because the user understands he or she is using the device in a way not intended by the manufacturer.

Mistake. A mistake is a failure of judgment or the inferential process leading to an incorrect decision about what action to take. Mistakes can arise from applying the wrong operating principles or procedures when making a decision or from nescient error arising from a lack of understanding of the adverse consequence of a particular course of action. Mistakes differ from abnormal use in that with a mistake the results are different than those expected by the user. For example, the user takes a well-intentioned shortcut on procedure, thereby omitting important steps. It is not obvious that the shortcut is hazardous. In the taxonomy shown in Figure 1, a mistake is a type of use error.

If the user did not intend to take a specific action, the result will fall into one of two categories.

Slip. A slip is a failure in the execution of an action sequence. A slip is a potentially observable externalized action not as planned (e.g., a slip of the tongue). For example, the user intends to press one button on a control panel but presses the one next to it instead. That would be a slip in this taxonomy. If the user intended to press the wrong button thinking that it was the correct action to take, it would be a mistake. If the user pressed the wrong button knowing full well that it was the wrong button, then the action would be abnormal use in this taxonomy.

Lapse. A lapse is generally reserved for more covert error forms and often involves a failure of memory, for example, forgetting to do something like cleaning an instrument before using it or confusing the meaning of an alarm signal. Lapses may not manifest themselves in immediately observable behavior because they frequently involved a failure to take a particular action, although they can have an observable effect on the outcome. A lapse may only be apparent to the person who experiences it.

Slips and lapses are errors that result from some failure in execution regardless of whether or not the plan being followed was adequate to achieve the intended purpose.

In the taxonomy in Figure 1, slips, lapses, and mistakes are the elements of use error. Although they may seem very similar on the surface, they arise from different sources. It is important to understand those sources when determining how to deal with them in the design of the user interface.

The following are some examples of use error for which the cause has been classified as a slip, lapse, or mistake.

•  The user misreads the value on a glucose meter by interpreting 2.2 to be 22 mg/dL. This is a mistake caused by a display that does not make the decimal point easy to read.

•  The user takes an incorrect dose of insulin from an adjustable delivery dose insulin pen after reading the small digital LCD display upside down. This is a mistake because the device design does not give adequate orientation information to properly read the display.

•  The user skips a step in loading reagents into a laboratory diagnostic system causing the blood chemistry results not to be produced and error messages to be generated. This is a lapse, because the user omitted a planned item.

•   A user is unable to get a blood gas reading from a handheld analyzer, because he or she put the blood sample into the wrong channel on the test cartridge before inserting it into the analyzer. This is a slip due to a reversal in the selection of a target channel for the blood sample.

In a comprehensive risk management process, the manufacturer must identify the hazards connected with the medical device, evaluate the associated risks, and control those risks throughout the medical device’s life cycle. This includes any risks related to the user interface. Applying HFE helps the manufacturer identify potential use errors so that appropriate and effective control measures can be designed. Classifying potential use errors as slips, lapses, or mistakes can provide insight into the ways that these error types may be controlled. Application of this taxonomy provides a lens through which various use-related problems can be viewed.

References

1. “How Common Are Medical Mistakes?” acessed online at http://www.wrongdiagnosis.com/mistakes/common.htm. 

2. “About Human Factors,” accessed online at http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/PostmarketRequirements/HumanFactors/ucm119185.htm#2 (Rockville, MD: U.S. FDA, May 13, 2009). 

3. J Reason, Human Error (Cambridge, England: Cambridge University Press, 1990). 

4. GHTF SG2N31R8:2003, Global Harmonization Task Force (GHTF), Study Group 2 (SG2), Medical Devices: Post Market Surveillance: Proposal for Reporting of Use Errors with Medical Devices by their Manufacturer or Authorized Representative.

Charles Sidebottom is the director, coporate standards, for Medtronic Inc. (Minneapolis). He can be reached at charles.sidebottom@medtronic.com.