Medical Device Link.

Originally Published MDDI July 2002

HUMAN FACTORS

Some medical errors can be prevented by incorporating human factors considerations into a product's design and development.

Christine Engelke and Daniel Olivier

Do medical mistakes result from human error or from poor human factors design?

The answer to this controversial question varies depending on who is being asked. Manufacturers have historically identified user error as the cause of many problems that FDA and consumer groups have attributed to poor device design.

Whatever the cause, medical errors are on the rise. This worrisome trend is a result of increasingly complex medicine coupled with a heightened demand for products with greater functionality.

According to FDA, medical errors take the lives of up to 100,000 Americans annually. Such mistakes also injure 1.3 million people, account for more than 3 million hospital admissions, and increase the nation's hospitalization bill by as much as $17 billion each year.¹

As a result, human factors engineering is receiving more and more attention from both FDA and the customers who demand products that are easier—and safer—to use.

In fact, medical device problems attributed to inadequate human factors considerations have been growing for many years. Table I provides a listing of some of these device problems.² Although some may seem trivial, the consequences in many of these cases have been catastrophic.

HUMAN ERROR VERSUS DESIGN ERROR

The popular six-sigma management strategy argues that processes, not people, fail; therefore, poor design is the culprit when use errors occur.³ Many manufacturers are now beginning to accept the responsibility for creating designs less prone to user error in conjunction with their efforts to increase customer satisfaction.

The means for assigning root causes to medical errors is changing as a result of legal precedents, evolving industry practices, and changing regulatory standards. Although the distinction between user error and design error is not well established, it is clear that the demand for more-robust and error-resistant designs is increasing.

This is not to say that all user errors are the fault of the manufacturer; clearly, the manufacturer must make certain assumptions about the expertise of the user, the user's observance of the labeling (and advertising) claims for the intended use of the product, and his or her adherence to the instructions. Table II presents examples of design changes made by manufacturers to reduce the likelihood of user error. Nevertheless, society is increasingly expecting the manufacturer to produce designs that minimize the risk of operator-related problems even further.

SPECIFIC ISSUES AND SOLUTIONS

In designing a device to minimize user error, it is helpful for manufacturers to refer to specific principles that make the device easier to use and understand. One example is creating a simple user interface that includes a well-organized layout of controls for easy system operation. Others include incorporating easy-to-understand reference documentation and considering such environmental factors as lighting, ambient noise, and space availability.

Data Overload. A common mistake made by many manufacturers is to present too much data on displays. This information overload can lead the operator to overlook important data, or to make mistakes due to the complexity of operating the device. (Many people have felt a similar feeling of information overload when trying to program their VCRs).

The greater the capability of the display, the greater the manufacturer's tendency to fill it up with more data, graphics, menus, buttons, and so on. The benefits of white space so often discussed with respect to printed materials should also be applied to displays. Too much text or too many graphics on a display decrease the reader's comprehension, just as they do for printed reports.

Designers of devices and displays must exercise restraint to include only essential information. Simpler is better. For example, consider an indicator for an electrical stimulator that displays "Stimulation ON" or "Stimulation OFF." This design could be simplified by having an indicator turn red when the device is on and disappearing altogether when the device is off.

Chunking by Sevens. George Miller wrote a paper titled "The Magic Number Seven, Plus or Minus Two."⁴ His idea is that the human mind is limited in its ability to process information. Miller's studies concluded that most humans are unable to retain more than seven "chunks" of information at one time. This philosophy can help manufacturers understand the limits of the amount of information users can readily process and understand.

The rule of sevens has implications for the amount of information presented on displays as well as within software program menus. In designing information displays and menus, engineers should first determine what information is essential, then decide how to best structure secondary information that may be of interest to the user.

A Word about Warnings. Although warnings and cautions printed on labels can be used to prevent errors, these are the least effective methods. A design should take into account that errors will occur, and then permit the user to recover from a mistake. Error-tolerant designs include such techniques as error trapping—which identifies data entries or event sequences that could cause problems—and confirming commands that initiate critical event sequences.

Warnings and cautions sometimes appear on a device even when they are clearly ineffective. For example, a card provided for passengers seated in the exit row of a commercial aircraft reads, "If you are seated in an exit row and you cannot understand this card or cannot see well enough to follow these instructions, please tell a crew member."

In certain cases, however, warning or caution labels might be the only feasible way to reduce risk. The need for these types of labels is recognized by users, and studies have shown that a product with a warning label is perceived by users as safer than the same product without the label.⁵ In other words, warning and caution labels can't hurt, but they are not particularly effective, either.

Designing Home-Care Products. Robust designs become more critical in inverse proportion to the expertise of the user. Devices targeted to home use must be especially easy to operate, as this population includes users who are potentially ill, medicated, handicapped, elderly, and generally less knowledgeable about the specific operation of the device. The wide variety of home users requires device designers to target the lowest level of education and skill across this population. Manufacturers should avoid relying on training and product labeling alone as preventive measures against user error.

BUILDING HUMAN FACTORS INTO EXISTING DESIGN PROCESSES

Human factors engineering is a user-centered design process. When devising a product, designers should research how different individuals use or might use the device in both a clinical setting and its various associated environments. Their research should determine the following characteristics:

• The functions that the device can perform.
• The types of potential user and their respective levels of expertise.
• The context of use, such as within a home or hospital environment.
• The workload of the user.
• The potential device abuse and misuse conditions.⁶

Ease of use is difficult to incorporate into a device after the initial design has been rendered. Human factors engineering activities must be integrated into a prospective design process; this minimizes additional overhead and helps realize significant benefits. Successful companies have learned that new product development activities must include assessment and evaluation of the product's ease of use and other human factors issues as essential ingredients for success.

New-product success is also based on the integration of design principles and the allocation of resources to ensure that human factors issues are addressed in existing design activities. Table III provides a list of specific practices associated with requirements definition, including design, implementation, testing, and user labeling, all of which can be applied to design and development activities.

The identification of potential errors is best accomplished by conducting an analysis of events and sequences in the user environment that can contribute to errors. This must be performed early in the requirements definition phase.

In the effort to improve human factors processes, it is also useful to learn the reasons why bad designs are sometimes created. Bad designs are never created intentionally; problems arise when the designers are unfamiliar with the actual use environment, are unaware of the expertise of the user population (or lack thereof), do not consider the unique situations that can face the user, or fail to identify and account for the physical limitations of the user population.

Why are designers not aware of these issues as prerequisites for any design project? One explanation may be the intense pressure to get new products to market has made knowledge of the user population a low priority. Yet one way to ensure good design is for the engineers to talk directly with the users and visit them in their operational environments. This simple assignment often sheds light on innovative ways to better serve the real needs of the users.

TESTING FOR HUMAN FACTORS DESIGN FLAWS

It is important to ensure not only that new products are properly designed, but that they are adequately tested. It is essential that optimal test practices be enforced (i.e., test practices that are effective in identifying potential errors).

Traditional test practices focus on testing derived from established requirements and conducted by internal test resource personnel. One problem with this approach to testing, however, is that these individuals might not be totally familiar with the operational use environment and might not know the common problems that can occur during device use. Identifying human factors problems requires exposure to the scenarios that occur during actual use. Questioning individuals through user focus groups, asking users to conduct beta tests, issuing surveys of user perceptions of device performance, and monitoring actual use scenarios are all effective ways to pinpoint such problems.

To ensure good design, the testing program must also be comprehensive. The testing program should address the following potential operator error conditions: entry of out-of-range or unexpected values or nonstandard command sequences, the use of improper configurations, operation with failed hardware components, a loss of power, operation under maximum loading conditions, and so on. The tester should introduce test sequences that are likely to present realistic error situations. This is best accomplished through testing in an actual use environment and using testers who are experienced in that environment.

EFFICIENCY VERSUS EFFORT

The extent and scope of any human factors effort can be scaled according to specific project needs, such that safety, effectiveness, and usability are optimized at a reasonable cost. The AAMI HE74 standard recommends exerting more-intensive efforts in the following circumstances:

• When developing an entirely new device (rather than making minor changes to an existing design).
• When developing a device that involves extensive or complex user interactions.
• When developing a device that performs a critical, life-sustaining function.
• When introducing an entirely new technology or method that is unfamiliar to users.⁶

These recommendations complement FDA's guidelines for increasing the rigor of formal design controls and documentation for devices that present higher levels of concern for users.⁷ The greater the potential risk associated with a user error, the more effort is warranted for its prevention.

To quantify this, designers can perform a task-and-hazard analysis, with an emphasis on human factors issues. An example, shown in Table IV, assigns a severity ranking to each item. These rankings can be used collectively to evaluate the overall risk.

CONTINUAL IMPROVEMENT OF HUMAN FACTORS DESIGN

Because no design is ever perfect, it is essential to have a program in place for continual improvement. Manufacturers receive information from customers concerning product ease of use and complaints of product errors. This feedback provides opportunities to confirm the robustness of a design and to identify potential enhancements.

Some manufacturers report a reluctance among medical personnel to report user errors; this results in minimal feedback to the manufacturer and therefore an underestimation of the significance of certain design elements. To counteract this tendency, manufacturers should make problem-reporting easy for medical staff, and provide training to service personnel, field sales agents, and customer service operators on effective techniques for soliciting this type of information. Such technologies as downloadable event logs and Web-enabled reporting can also increase the volume of useful feedback.

Once data are collected, techniques such as Pareto analysis can be applied to identify the most common errors—which would provide the greatest benefits if corrected. (Pareto analysis involves identifying the few crucial factors that contribute the most to an overall effect.) It is appropriate to add a caution here: there is a tendency on the manufacturer's part to blame all problems on the user. A more detailed analysis of the problems can uncover methods at the design level that can reduce the occurrence of the errors.

Cause-and-effect diagrams or matrices can be helpful in identifying alternative design techniques to reduce the occurrence of problems experienced by users. Added investment in these types of analyses can prove valuable. It not only can reduce the number of complaints received from customers, but also may result in the creation of a product that is easier to use, which contributes to higher customer satisfaction. Manufacturers must also consider the benefits of reducing the liability risk that might be attributed to the so-called user errors.

CONCLUSION

Poor human factors design is being increasingly identified as a significant cause of medical errors. Those errors result from both failures in the design of medical products and changes in society's expectations for product design. As a result, it is becoming increasingly important for medical product manufacturers to emphasize human factors engineering.

Designers and manufacturers can apply specific design and testing techniques to reduce the initial risk of human error, and they can implement continual improvement programs to reduce risk after new-product introduction. The advantages of participating in a human factors engineering program have traditionally been stressed from a customer-satisfaction perspective; however, the reduced liability risk is potentially a much greater benefit to manufacturers.

REFERENCES

1. "FY 2003 Justification of Estimates for Appropriations Committees," in FY 2003 Budget Summary, Food and Drug Administration, Department of Health and Human Services, (Rockville, MD: FDA, 2002), 2.

2. Do It By Design FDA, (Rockville, MD: 1996), 9.

3. M Harry and R Schroeder, Six Sigma: The Breakthrough Management Strategy Revolutionizing the World's Top Corporations (New York: Doubleday, 2000), 225.

4. GA Miller, "The Magic Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information," The Psychological Review 63 (1956): 81–97.

5. M Sanders and E McCormick, Human Factors in Engineering and Design (New York: McGraw-Hill 1993), 680.

6. "Human Factors Design Process for Medical Devices," ANSI/AAMI HE74:2001 (AAMI, 2001), 13.

7. "Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices" (Rockville, MD: FDA, May 29, 1998), 12.

Christine Engelke is a software quality engineer at Certified Software Solutions Inc. (San Diego). Daniel Olivier is president of Certified Software Solutions Inc.

Copyright ©2002 Medical Device & Diagnostic Industry