Friedrich & Custer: Designing for Failure

You are having surgery. It’s relatively minor, hoping to relieve the leg pain you have while walking. After the procedure is done, the surgeon tells you everything went well expects you will soon experience relief from the pain.

But the pain actually gets worse. You see the surgeon in clinic the next week and follow-up testing reveals that the tip of a long catheter used in your leg had sheared off and was unknowingly left inside your body. That 2-cm catheter fragment has been sitting in your artery, causing blood to clot around it, impeding blood flow to your leg. You will need emergency surgery to remove the catheter fragment and restore circulation. You may have residual weakness in that leg.  

Medical devices sometimes fail; a piece may break off inside of a patient. We don’t know how often this happens, as the data relies on recognition (which is unreliable) and reporting (which is voluntary). Regardless of the incidence, any one of these retained catheter fragments could severely harm the patient depending on the size of the fragment and where gets lodged inside their body.

The outcome hinges on early recognition that the entire device wasn’t removed. Early recognition allows the clinician to determine if removal is needed and intervene promptly to prevent complications such as infection, thrombus (clot), and aneurysm. As technology advances, we perform more minimally-invasive procedures and outpatient therapies that use endovascular catheters. As we use more catheters, reliable identification of incomplete removal becomes increasingly important.

Medical devices are designed for when they perform as intended. We need the same thoughtfulness in design for when they fail.

In patient safety, we often get involved when things go wrong. We are generally successful in preventing adverse event recurrence by changing things in control: our environment, internal processes, or the equipment we purchase. However, we see a recurring issue that is also plaguing other healthcare organizations, and we fear we cannot improve it without seeking external change.

When medical devices fail, we report it to the manufacturer, and the response in our experience is often that it was due to “user error.” We are writing this as a call to action for better human factors design across/between medical device manufacturers.

If “user error,” one would simply re-educate the user. In the case of identifying broken catheter tips, there are several reasons why re-education won’t work without design standardization:

  • Multiple manufacturers produce those devices, and the catheters/devices all look different. Sometimes there are markings or a taper closer to the end, other times not.
  • The clinician removing a catheter is often not the same person who placed it. The removing clinician may not know what the catheter looked like before use.
  • Patients move between different hospitals and clinics. The first time you see “brand X” catheter may be the first time you remove it and have to account for tip integrity.
  • Catheters are sometimes stretched or navigated through tortuous blood vessels. Even if there is an unused catheter available for comparison, it may not look the same.

We need to think differently.

For decades, cars from all manufacturers have been made with similar design requirements: gas pedal on the right and brake pedal beside it to the left, all lights on the rear of the vehicle are the same color red, etc. Driving cars is risky; that basic standardization helps the humans at the wheel.

Delivering healthcare is also high risk. Why not require some level of design standardization to help clinicians who are already struggling with an overburdened cognitive load?

Our proposed solution: Require the tip of every catheter made by any manufacturer to have the same visual indicator. A suggestion from our work is to color the 5-mm tip of every catheter lime green.

A colored tip indicator helps in two ways:

  1. Easier identification from anyone removing any catheter. This is key to early identification. If the clinician does not see the lime green indicator (tip), they would be cued to investigate.
  2. Easier education. Any clinician or student (e.g., nurse, physician, pharmacist) needs only to be taught to visually check for the lime green tip after removing a catheter.

This easy identification strategy frees up cognitive space for the tasks that deserve more thought, such as assessing the patient or performing complex procedures.

Safer healthcare is an imperative. We work with incredibly talented healthcare providers who challenge themselves to provide innovative, high-quality care, every day. This sentiment is echoed by the Institute of Medicine: “People working in healthcare are among the most educated and dedicated workforce in any industry. The problem is not bad people; the problem is that the system needs to be made safer.”

We can create that safer system by designing devices in a way that promotes early identification of failure and minimizing cognitive load.

Rebekah Friedrich, MS, RN, CCRN, CPPS, is the senior performance improvement leader at the University of Maryland Medical Center.

Jason Custer, MD, is the director of patient safety and medical director of the pediatric intensive care unit at the University of Maryland Medical Center and an associate professor of pediatrics and the University of Maryland School of Medicine.

10 thoughts on “Friedrich & Custer: Designing for Failure

  1. I like the idea of the green tip color code. Would greatly help everyone identify the tip at the end of the procedure.

  2. And another sad truth. Engineers are not taught about device failures post design. They are taught to design, period. Part of the reason I became an engineer after 10 years as a tech was frustration over designs that didn’t account for failure. As I earned my degree, I realized that limitation; and it would take a wholesale revamping of engineering education to expand beyond ‘get it out the door’ design philosophy.

    And that doesn’t account for the points others have brought up.

  3. While I agree with all points made in the article and comments and hope you are successful at your particular endeavor, it is more than a bit frustrating to read yet another column on the general topic of safety (I’ve written many, as has Bill Hyman, as have many others). Blaming device failures on how they are used is nothing new.

    Regarding your point on how cars are designed, over a decade ago I made the point that I could go into any showroom and test drive a car in minutes, but every time a new medical device shows up in a hospital, a carefully planned in-service is needed. Not for a new type of device, mind you, but for devices that have been around for decades, e.g., infusion pumps.

    Now consider this: I entered the field in 1979. Back then, devices didn’t change much if at all during their lifetimes, often exceeding a decade. Clinical users and technical support staff could be trained on acquisition and over time gain a sense for the operational and technical nuances of the device, both on its own and in concert with other devices at a point of delivery of care. Now? Software upgrades are common. Consider the implications not just in terms of the device itself but also in terms of the devices with which it interacts.

    Speaking of interactions, the integration of information presented by devices that used to occur in the mind of the clinician now increasingly occurs behind the scenes. Consider the implications.

    Before writing me off as a Luddite, please understand that I have been active throughout my career in promoting the development of technology-based tools to implement features unavailable when I entered the field, including medical device communications interoperability. But my grounding in all of that work was the promotion of doing so systematically and safely. At one point, I was sponsored by AAMI to serve as the US user representative on an international standards committee focused on risk management of medical devices on general purpose IT networks.

    I have two requests:

    1. Study the history of medical device safety issues since the adoption of the Medical Device Amendments of 1976.

    2. Stick with this going forward. Don’t back down. You will encounter resistance and, worse, indifference.

    Thanks for reading this screed. Good luck.

  4. Thanks for taking the time to read our editorial. If this resonates with you, please leave a comment! Do you agree that sensible design standardization could help improve patient safety?

    • Yes, in particular when it comes to human factors. Your analogy to a car is spot on.

      A number of years ago, I brought up UI standardization at a meeting focused on communications interoperability standards. My argument was both were necessary to better assure safety. A marketing executive argued that often the only differentiator between their and other products was the UI. I didn’t pursue that further; perhaps I should have, (you might consider investigating safety cases and applying them to your work, if you haven’t already:

  5. In some cases reporting s not voluntary under the FDA’s MDRs.
    Color coding and other standardization is a long recognized design approach has seen remarkably little adoption in healthcare. We should remember here that deliberate unstandardization has its place as well.
    I agree that “user error” is often a lame excuse when the design did not address the issue in the first place.In the present case did the instruction adequately address tip inspection? But even if it did, instructions are not a cure for bad design

    • William,

      Great points! We wouldn’t want to interfere with properties of the device that allow it to function.

      In my experience speaking with clinicians, they are aware of the expectation to check the tip for integrity. But the design makes it difficult to do quickly and reliably.

      There is an additional layer of complexity; occasionally, clinicians purposefully cut (shorten) catheters before using them on patients. Recently, a nurse removed a vascular catheter and wasn’t certain of the tip integrity. 20 minutes later, after asking the opinions of several colleagues, they determined that the tip was missing. Someone at the previous hospital cut the catheter to shorten it before placing the line. A standard colored indicator could make that assessment fast and reliable.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s