Medical Devices in the "Wild"
A 75-year-old man with a history of congestive heart failure (CHF), coronary artery disease, diabetes, chronic pain, arthritis, and hyperlipidemia was admitted to the hospital with a CHF exacerbation manifesting as lower extremity edema and weight gain. At baseline, he was able to function independently and perform all activities of daily living. The patient was treated with diuretics, fluid restriction, and was given dietary and medication education. After a short period of treatment, his swelling improved and he was able to ambulate on the hospital ward without difficulty. The medical team was preparing the patient for discharge the following day.
That afternoon, the patient was lying in bed watching television when his nurse came into the room to assess him. The bed was low to the ground and locked in position, so she raised the bed up to perform her assessment. The patient had sequential compression devices (SCDs) in place to prevent deep venous thrombosis (DVT). When the nurse raised the bed, unbeknownst to her, the tubing for the SCDs caught on the bed wheel lock and unlocked the bed. After completing her assessment, the nurse left the room. Having been told he should ambulate several times daily, the patient then sat up on the side of the bed and attempted to stand. In doing so, he pushed down on the bed with his hands. When he did this, the bed rolled out from under him and he fell onto his left side. He immediately complained of hip pain, and on radiographs was found to have a broken left hip.
The next day he went to surgery for planned open reduction and internal fixation of his hip fracture. Unfortunately, he developed respiratory issues and had to be transferred to the medical intensive care unit for closer monitoring postoperatively. He did improve temporarily, but despite receiving appropriate DVT prophylaxis, one week after surgery he suddenly experienced a cardiac arrest and was found to have a massive pulmonary embolism. He was briefly resuscitated but died a short time later.
The unit where the fall occurred called a multidisciplinary "fall huddle" to reenact the circumstances of the incident. The nurse was certain the bed was locked, but a technician noticed that the SCD tubing had fallen around the wheel and brake of the bed. During the reenactment, the staff realized it was possible for the tubing to catch the brake and unlock the bed without being noticed. The hospital's patient safety department immediately informed the company that manufactured the beds and the SCD equipment. The SCDs were replaced with newer equipment that had shorter tubing that could not wrap around the bed brake and unlock it inadvertently.
Medical device–associated errors are common, costly, and often cause preventable harm. A 2008 estimate indicates the annual cost of medical errors to be $17.1 billion to the United States economy, with device-associated errors among the top 10 contributors to this number.(1) A systematic review found that more than 23% of errors in the operating rooms involve device-related issues.(2) A substantial proportion of device-related errors are use errors that occur due to unintended or unanticipated interactions between devices and users in complex work environments. Evidence indicates that medical device-use errors may be more frequent and cause more harm than device failures (e.g., malfunctions).(3,4)
Medical device-use errors can be significantly reduced using human factors engineering (HFE) principles and techniques.(5) The present case demonstrates the importance of two HFE principles. First, secondary defensive mechanisms (e.g., nurse checking the locked status of the bed before leaving the room) may help to a limited extent, but they are much less reliable than safety features integrated into the product.(6) Second, environments in which any medical devices are used, including patient beds, typically are multifaceted complex sociotechnical systems. The safety of care provided in these systems depends not only on the design and complexity of a particular technology (i.e., a typical patient bed has more than 25 points of control), but also the interactions among different technologies (e.g., patient bed and SCD [sequential compression device] tubes) (7) and other components of a work system, including a multitude of users.(8,9)
Although new SCD equipment with shorter tubes may work for this particular problem, what about other lines and tubes? In such care settings, problems may be more complex than first realized, and the expertise of human factors engineers may be required to ensure all system considerations are addressed. An HFE approach for such a problem would be to redesign the system while involving the users throughout the entire design and development process utilizing a user-centered design approach.(10) One solution could be an effective means to control the position of tubing in a manner that it does not affect bed controls. Another solution could be a very short raised edge or cleat affixed to the floor sufficing as a wheel chock to prevent the bed from rolling. A complementary and effective solution could be to develop a status indicator that makes it easy for team members (nurse, patient, family, nursing aid, etc.) to be aware of the bed's locked status even with linens draped about the foot. Bed manufacturers could provide a "bed lock status" indicator displayed more prominently on the bed and capability for remote display at a central nursing monitoring position. Another, safer design could involve consolidating all bed status information, including lock status, on one visual indicator for better awareness. Ideally human factors engineers identify the safety risks in these complex work systems early in the design process and develop solutions in the form of design requirements and user instructions to mitigate these risks. Human factors engineers and others concerned with patient safety continue this process throughout the entire product lifecycle. Ultimately, without scientific evidence concerning solution effectiveness and adequate support by robust nationwide reporting, we cannot be certain of achieving safe designs for complex work environments.
The SCDs used in the current case were replaced with newer equipment that had shorter tubing to prevent inadvertent unlocking of the bed brake. This reactive change may be limited to just one hospital and likely did not result in more systemic changes in SCD equipment through the entire health system. Unlike in aviation where safety-critical changes are quickly communicated and acted upon across the entire industry (11), the medical device and health care industries are not efficient knowledge markets.(12) In other words, the medical device industry has an inadequate proactive surveillance mechanism in place; it communicates needed improvements less effectively than the aviation industry. In medicine there are subscription-based services and publicly available Web-based databases (e.g., MAUDE, the FDA's voluntary reporting system) useful for learning about device-related problems and comparing different products. However, these systems are limited in that they do not address device-interaction effects or other contextual factors, and can also be difficult to navigate. On the other hand, one can use common Web sites to compare many consumer products with respect to price, customer satisfaction, and other criteria within minutes.
Learning from device-related errors one organization at a time is not an acceptable approach for assuring safety. There is a great need for universal reporting and learning systems that leverage the power of information technology and public health informatics. FDA's recent regulatory efforts on development of a unique device identification system and the publication of the AAMI [Advancing Safety in Medical Technology] technical report providing guidance on post-market surveillance of use error management are two laudable efforts in this direction (13); however, there seem to be many hurdles to overcome before the medical device industry becomes an "efficiently learning industry."
Higher levels of collaboration between manufacturers, clinicians, human factors engineers, patients, families, and other stakeholders are needed to achieve safer designs. Such collaborations can produce the minimum product-specific design requirements necessary to avoid any known safety hazards. It is important to view this as an enabler for innovation rather than a barrier. Instead of each manufacturer determining the minimum acceptable safety design features for each of their products, safety requirements could be identified in an organized, systematic way based on rigorous usability evaluation studies and learnings from surveillance systems. Subsequently, this product-specific information would be made publicly available. An example for this kind of approach is the recent collaboration between the Johns Hopkins Armstrong Institute for Patient Safety and Quality and the Johns Hopkins University Applied Physics Laboratory, along with input of the 41 representatives from industry, regulators, clinicians, pharmacists, engineers, and human factors specialists to develop a prototype infusion pump and simulation test environment that investigates the design features for a safer infusion pump.(14) In summary, medical device and health care industries can produce safer products with increased collaboration and joint efforts to eliminate preventable harm and save lives.
- Medical device use errors are responsible for a large portion of medical errors and preventable harm.
- Health care is provided in complex, sociotechnical work systems. Introducing any new technology into these systems, including medical devices, can result in both anticipated and unanticipated effects, some of which may cause medical errors and preventable harm.
- By designing safer medical devices based on human factors engineering principles and methods, patient safety can be improved and device-associated preventable harm can be reduced.
- A more robust and efficient knowledge market is needed in health care, which should include a systematic, organized mechanism to provide evidence on and standards for safe design characteristics for each major product type.
- Stakeholders (including clinicians, patients, engineers, device manufacturers) should collaborate for safer design of medical devices.
- Health care organizations should work with human factors engineers and conduct basic usability evaluation themselves before purchasing a new device to reduce risks. However, it is neither reasonable to expect nor feasible for each individual entity to conduct robust and comprehensive usability evaluations prior to purchase, which makes the importance of a more centralized body of knowledge globally available more apparent.
Ayse P. Gurses, PhD Human Factors Engineer Associate Professor Armstrong Institute for Patient Safety and Quality Schools of Medicine, Bloomberg Public Health, Whiting Engineering Johns Hopkins University
Peter A. Doyle, PhD Human Factors Engineer Johns Hopkins Hospital Clinical Engineering Department
1. Van Den Bos J, Rustagi K, Gray T, Halford M, Ziemkiewicz E, Shreve J. The $17.1 billion problem: the annual cost of measurable medical errors. Health Aff (Millwood). 2011;30:596-603. [go to PubMed]
2. Weerakkody RA, Cheshire NJ, Riga C, et al. Surgical technology and operating-room safety failures: a systematic review of quantitative studies. BMJ Qual Saf. 2013;22:710-718. [go to PubMed]
3. Cooper JB, Newbower RS, Long CD, McPeek B. Preventable anesthesia mishaps: a study of human factors. Anesthesiology. 1978;49:399-406. [go to PubMed]
4. Leape LL. "Smart" pumps: a cautionary tale of human factors engineering. Crit Care Med. 2005;33:679-680. [go to PubMed]
5. Weinger MB, Wiklund ME, Gardner-Bonneau DJ, eds. Handbook of Human Factors in Medical Device Design. Boca Raton, FL: CRC Press; 2010. ISBN: 9780805856279.
6. Association for the Advancement of Medical Instrumentation. ANSI/AAMI HE75: 2009: Human factors engineering: Design of medical devices. Arlington, VA: AAMI; 2010.
7. Pennathur PR, Thompson D, Abernathy JH III, et al. Technologies in the wild (TiW): human factors implications for patient safety in the cardiovascular operating room. Ergonomics. 2013;56:205-219. [go to PubMed]
8. Carayon P, Schoofs Hundt A, Karsh BT, et al. Work system design for patient safety: the SEIPS model. Qual Saf Health Care. 2006;15(suppl 1):i50-i58. [go to PubMed]
9. Gurses AP, Ozok AA, Pronovost PJ. Time to accelerate integration of human factors and ergonomics in patient safety. BMJ Qual Saf. 2012;21:347-351. [go to PubMed]
10. Abras C, Maloney-Krichmar D, Preece J. User-centered design. In: Bainbridge WS, ed. Encyclopedia of Human–Computer Interaction. Berkshire Publishing Group; 2004;37:445-456. ISBN: 9780974309125.
11. Donaldson L. When will health care pass the orange-wire test? Lancet. 2004;364:1567-1568. [go to PubMed]
12. Nichols LM, Ginsburg PB, Berenson RA, Christianson J, Hurley RE. Are market forces strong enough to deliver efficient health care systems? Confidence is waning. Health Aff (Millwood). 2004;23:8-21. [go to PubMed]
13. Rumsfeld JS, Peterson ED. Achieving meaningful device surveillance: from reaction to proaction. JAMA. 2010;304:2065-2066. [go to PubMed]
14. Ravitz AD, Sapirstein A, Pham JC, Doyle PA. Systems approach and systems engineering applied to health care: improving patient safety and health care delivery. Johns Hopkins APL Tech Dig. 2013;31:354-365.