Alarm Management: First Things First
Using Reliable Data to Eliminate Unnecessary Alarms
By Tim Vanderveen, PharmD, MS
When Hedy Cohen, RN, BSN, MS, a critical care nurse and former vice president of the Institute for Safe Medication Practices (ISMP), was admitted to a major medical center with severe abdominal pain, she and her husband, Michael Cohen, RPh, MS, ScD (hon.), DPS (hon.), president of ISMP, were dismayed by the number of medical device alerts and alarms she had to endure—especially because many were completely unnecessary. Three intravenous (IV) antibiotic medications had all been hung as primary infusions—one q6h, another q8h, another q12h—generating nine “infusion completed” alarms. If the bags had been hung as secondary infusions, the pump would have restarted the primary infusion automatically once the antibiotic infusion was complete—without sounding an alarm.
Some alarms kept on sounding for such a long time that Hedy (herself a critical care nurse; therefore, not a typical patient), finally reset the pump herself before using the call to summon a nurse to take the empty bag down and restart the primary infusion. In addition, because the emergency department staff had inserted the catheter in the antecubital fossa site, simply moving her arm triggered too many occlusion alarms to count. Still more alarms could be heard coming from other patients’ rooms and were especially frequent at shift change and during the night.
Hedy estimated she experienced more than 50 alarms in one 24-hour period. She observed, “I totally understand the need for alarms and why a nurse can’t always respond immediately. But when you’re sick or worried about a loved one, your tolerance is totally gone. After a while, you just can’t stand it.”
A Pressing Need
In today’s hospital environments, patients can be inundated with alarms from a wide variety of medical devices. The number of alarm signals per patient per day can reach several hundred depending on the unit within the hospital, translating to thousands of alarm signals on every unit and tens of thousands of alarm signals throughout the hospital every day (Joint Commission, 2013a). Nurses compare the sound to that of a casino or a carnival (Welch, 2012). The continuous noise, especially from unnecessary, “nuisance” alarms, can lead to “alarm fatigue”—desensitization, elevated stress, and clinical complacency. If not properly managed, the unintended effects of clinical alarm systems can seriously compromise patient safety (Joint Commission, 2013a).
A fatigued clinician may resort to unsafe practices such as turning down the volume so that an alarm can’t be heard or even removing the device from patient use. The results can be tragic (Joint Commission, 2013a). A patient in Boston died after alarms had been turned off, which ultimately resulted in an $850,000 settlement (Kowalczyk, 2011). ECRI Institute has rated alarm hazard as the number one or two patient safety issue in its “Top 10 Technology Hazards” since 2010 (ECRI, 2010-2013. A recent Association for the Advancement of Medical Instrumentation (AAMI) Healthcare Technology Foundation (2011) survey showed that 20% of respondent hospitals reported an avoidable, alarms-related adverse event in the past two years.
Another reason for hospitals to make alarm management a priority is the impact of patient satisfaction on the Centers for Medicaid and Medicare Services (CMS) reimbursements. CMS uses the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey to measure patients’ perceptions of their hospital experiences (CMS, 2014). Lower patient satisfaction scores can result in lower value-based incentive payments to hospitals. At ECRI Institute’s 2013 alarm safety web conference, 53% of respondents (n = 359) said that alarm issues have impacted their facility’s patient satisfaction scores (ECRI, 2014a).
In recent years, many organizations have begun to focus intently on this issue, including the Food and Drug Administration, ECRI Institute, American College of Clinical Engineering, and AAMI. In June 2013, The Joint Commission announced National Patient Safety Goal (NPSG) 06.01.01, which set the following deadlines for hospitals and critical access hospitals (Joint Commission, 2013b):
- Beginning January 2014, hospitals are required to establish alarms as an organization priority and to identify the most important alarms to manage based on their own internal situations.
- Beginning January 2016, hospitals will be expected to develop and implement specific components of policies, procedures, and staff education about alarm system management.
|New National Coalition Tackles Alarm Management in Healthcare|
Close to 100 clinicians, healthcare technology experts, regulators, patient safety advocates, researchers, and leaders in the medical device industry gathered on April 24-25, 2014, in Arlington, Virginia, to kick off an unprecedented two-year initiative to improve alarm management in healthcare.
The National Coalition for Alarm Management Safety is an effort launched by the AAMI Foundation’s Healthcare Technology Safety Institute (HTSI). Its inaugural meeting brought together experts to hear the experiences of seven hospitals that have worked to improve alarm management and the perspectives of industry leaders. They also began to discuss what is needed to become a “high-reliability organization,” one in which systems are in place to radically reduce failures. The presentations sparked lively discussions, and attendees identified several areas worthy of follow up, including a need for additional research to determine if adverse patient outcomes have resulted from enhanced alarm parameters for clinical monitors.
The tangibles the coalition plans to deliver over the next year include the following:
Over the next two years, the coalition hopes to produce a list of suggestions on how hospitals can become high-reliability organizations as devices and clinical-care processes evolve over time. Members also plan to write preliminary recommendations about the management of ventilator and infusion pump alarms (AAMI, 2014a).
For further information, visit www.aami.org/htsi/alarms/
The AAMI Foundation Healthcare Technology Safety Institute (HTSI) helped form a National Coalition for Alarm Management, which held its inaugural meeting in April 2014 (see sidebar).
Alarm-management efforts to date have focused primarily on physiologic monitors and ventilators. Alarms related to IV infusion pumps presumably will be a next area of focus. Infusion pumps typically have a fraction of the alarms and alerts generated by multi-parameter monitors; however, virtually every patient has one or more pumps, which means that across an entire unit, infusion pumps generate a very large number of alarms.
In this article I will briefly describe current efforts to improve alarms management, provide an early look at infusion pump alarm data, identify some trends and observations, and suggest possible considerations for hospitals as they begin to focus on reducing alarms and making the patient care environment quieter. Clearly with regard to alarms, alerts, and overall noise within the hospital, we are now at a tipping point; dramatic reductions have already been reported from some leading centers that have focused on multi-parameter monitors (Graham, 2013; AAMI, 2014b).
First Things First
Patients and family members often ask a very good question: “Why do the devices in patients’ rooms have to alarm at the bedside?” They go on to point out the obvious: “Make the devices so that they alert my nurse, who can do something about the alarms!” This is a reasonable question and often is identified by caregivers as the “holy grail” of alarms management, especially with infusion pumps. However, its implementation poses numerous challenges. Today very few infusion pumps are assigned to specific patients, so sending an alarm notification directly to a nurse without knowing the patient context is not a viable solution.
Data from a recent analysis of almost 40,000,000 alarms collected over a three-month period from 730 hospitals show that up to 30% of alarms occur while the clinician is interacting with the pumps (Vanderveen, 2014). Examples include placing the IV tubing into a free-flow condition (safety-clamp-open alarm), starting an infusion without first installing the IV tubing (no-set-detected alarm), and forgetting to open a clamp at the start of a new infusion (patient-side-occlusion alarm). In these cases, the last thing a nurse wants is a call to her mobile phone telling her a patient’s pump is alarming—when she is standing next to that pump at the patient’s bedside. ECRI Institute has warned that “…applying [alarm integration systems] without correcting the underlying flaws in the process—such as reducing the number of unnecessary or avoidable alarms—will invariably lead to disappointing results” (ECRI, 2014a).
No doubt, the patient-association gap ultimately will be solved; the current standards for types, levels, and location of alarms will be modernized; and third-party alarms-management and enterprise health information technology (HIT) vendors will provide innovative solutions to this longstanding noise issue. In the meantime, rather than disseminating (and duplicating) the alarms, a more proactive approach with greater “speed to impact” is to begin to reduce the alarm noise through targeted changes in practices, policies, and device settings.
The old adage, “You cannot change what you cannot measure,” fits today’s efforts to improve alarm management. You need data to identify opportunities for improvement and to measure the success of improvements. The introduction of wireless connectivity greatly improved the ease of updating IV infusion pump datasets (drug libraries) and downloading the continuous quality improvement (CQI) data. Now connectivity allows hospitals to collect all infusion-related data, including alarms and alerts.
The early CQI data provided a treasure trove of previously unavailable information on infusion practices, effectiveness of drug dose limits, and typical clinician response to safety warnings. Today’s alarms and alerts data provide a similar “window to practice” that we’ve never had before. Data analysts can evaluate pump alarms for hundreds of hospitals for a macro look, and individual hospitals for their specific alarm profiles. Just as the early CQI data opened our eyes to previously unrecognized variability in infusion practice, the new data are revealing similar variability for infusion pump alarms and alerts.
Alarms and Alerts Data – An Early Look
Once healthcare organizations began to analyze the data on infusion alarms and alerts, it quickly became apparent that interpreting the data would require more than merely creating charts and graphs; understanding and investigating current practice is also required. In some cases, the cause of an alarm is readily apparent. For example, frequent “downstream” (between the pump and the patient) occlusions in adult medical/surgical units most likely result from catheter placement. However, the causes of alarms such as the following are less easily identified:
- Why are 10% of alarms for occlusions above the pump? Most of the infusion tubing above the pump does not have clamps, so is the alarm caused by a failure to open air vents with bottles? By wetted air vents that are no longer functional? Is crimped tubing a possible cause?
- Why would 28% of air-in-line (AIL) alarms in the intensive care units in one IDN be associated with the drug amiodarone, but only 3% in a similar IDN? Presumably, same drug, same infusion pumps, and same disposables—but a very different result. In the IDN with the high incidence of AIL alarms, why would switching from premixed amiodarone sourced from an outside pharmacy-compounding center to amiodarone prepared in-house with a pharmacy robot drop the AIL alarms to 1/10th their previous frequency? Do refrigeration and storage impact infusion delivery? Are cold IV solutions more problematic? This was not an isolated example. Similar variability was seen when the same two IDNs were compared with regard to other drugs associated with AIL alarms; only one drug was in the top 6 in both IDNs.
These and many other related examples make it clear that multiple factors, including preparation, storage, methods of administration, configuration of the alarm settings, and nurse/device interaction, will all require investigation, as hospitals begin to address the pump alarms.
It may turn out that some pump alarms simply have no value, but (as with monitors) since they are available, they are configured. One children’s hospital eliminated 24,000 near-end-of-infusion (NEOI) alarms in just two weeks after realizing that NEOI alarms serve no real value and that the pump configuration could be easily changed via the wireless network. In another case, the use of an anti-siphon valve to slightly increase the pressure in the IV line during a lengthy infusion of etoposide reduced the number of AIL alarms from 26 to 0. The slight increase in pressure in the infusion tubing had a dramatic impact on eliminating the typical “frothy” bubbles of this problematic medication.
Where to Begin
“Throwing technology” at the alarm problem is not sufficient; you need to fix the process first,” ECRI Institute points out in its Alarm Safety Handbook (ECRI, 2014a). Hospitals need to drive reductions in preventable alarms before sending alarms to wireless devices and nurse call stations. Reducing the number of alarms overall makes the remaining alarms more meaningful. Staff members are less stressed and can spend more time providing better patient care. Patients and family members will certainly appreciate the reduction in alarms. And if and when a distributed alarm system is put in place to alert nurses directly, its implementation and use will be simpler and more effective.
The Joint Commission has emphasized, however, that “These issues vary greatly among hospitals and even within different units in a single hospital” (Joint Commission, 2013b). Because of this variability, efforts to improve alarms management must begin with the collection and analysis of reliable data to identify specific issues and opportunities for a particular setting. Hospitals need to benchmark the number and types of alarms, compare findings within and among organizations, and then identify and investigate problem areas.
What Can We Do Now?
With advanced connectivity and data analysis, the stage is now set for addressing infusion pump alarms. Unfortunately many hospitals “have grown frustrated with the slow progress of the work and the challenge of exporting usable data from the devices. In their experience, it takes many person-hours in order to utilize and apply the data within their organization” (AAMI, 2014a).
Fortunately, the strides already made in integrating IV pumps and health information technology now make it possible to address alarm fatigue in a comprehensive way. Applications that combine wireless connectivity, a hosted site for data storage and retrieval, and advanced analytics and reports allow hospitals and health systems to store, access, and analyze huge amounts of data without requiring additional software or staff.
The applications provide an information dashboard that displays actionable retrospective data aggregated from across a smart pump system, hospital- or enterprise-wide. The information includes alerts such as “infusion completed,” alarms such as “AIL” or “occlusion,” the care area, drug or fluid infused, and frequency. If a smart pump system is integrated with an electronic health record (EHR), a particular pump can now be associated with a particular patient, providing further information to help determine the cause of the alarm.
With these new applications, authorized users can work with the data by computer anywhere, anytime. Nurses, pharmacists, data analysts, risk/quality managers, and senior executives can also use the applications to compare performance within facilities, across their IDN, and to incorporate performance goals, without having to incur the burden of daily data management themselves. With everything right in front of them, they can easily review data, create graphs and slides, generate reports, and share information with nursing managers, educators, senior management, bedside clinicians, and others. Examining trends in the data can help identify key areas of focus and needed changes within smart pump datasets, drug and fluid libraries, pump configurations, and nursing education and practice. System-generated reports can also be used to demonstrate compliance to Joint Commission standards.
Analyzing the data on millions of alarms from thousands of pumps over several months’ use provides some insight into causes and possible mitigations or, at least, areas for focus. Many unanswered questions remain; the following is one approach for beginning to answer such questions and to improve infusion pump alarm management.
- First, establish a multidisciplinary Infusion Pump Alarms Committee—ideally a separate committee, not concerned with monitors or ventilators, with a clear focus on the pumps. Members should be well versed in the pump system itself and how the pumps are configured for different care areas and profiles.
- Second, survey nurses with regard to how they view alarms. Nursing perceptions are very important, and valuable insights can be captured that might otherwise be overlooked. Survey results can be used as a baseline to help identify areas for improvement and to measure the effectiveness of any interventions.
- Third, if your infusion system provides alarms and alerts data, by all means create graphs and charts to show where, when, and what types of alarms are occurring in your hospital or system. Your infusion pump vendor(s) can help you determine if these types of data are available.
- Fourth, conduct direct observations in various patient care areas, using trained observers who are familiar with the pump technology, to measure the frequency and types of alarms.
- Fifth, consider using whatever consulting services your vendor may provide. Given the high degree of variability in the types and causes of alarms within and between different care units and hospitals, a vendor’s clinicians with multi-hospital experience can help identify causes and possible mitigations, develop an educational program, and monitor the effectiveness of interventions.
With the availability of extensive, reliable, relevant data, hospitals can begin to investigate variation in patterns, identify problematic practices, and share best practices to address the pump alarms and alerts. Alarm management is not just about managing the noise, it’s about reducing it. For nurses, fewer nuisance alarms means fewer unnecessary trips to the bedside. Remaining alarms become more meaningful and more likely to prompt a timely response. For patients, a more peaceful environment can help patients rest, improve patient and family experience, and possibly result in higher patient satisfaction scores and CMS reimbursement. Optimizing the current process is also essential to prepare for the introduction of new technologies, such as alarm integration systems.
Improving alarms management is not a one-time effort. NPSG 06.01.01 sets the process in motion, but the use of advanced data reporting tools to identify opportunities for improvement needs to be an ongoing part of multidisciplinary robust process improvement efforts, as hospitals work to continuously improve alarm management.
Tim Vanderveen is vice president of CareFusion’s Center for Safety and Clinical Excellence. He is also a member of the AAMI Board of Directors. He may be contacted at Tim.Vanderveen@CareFusion.com.
This article is based on my more than 30 years’ experience in intravenous (IV) infusion safety. I am an employee of CareFusion, which holds many patents, including those associated with clinical device operating parameters. I also was employed by the predecessor companies that were involved in developing modular smart IV infusion safety systems and the associated wireless connectivity. As vice president of the CareFusion Center for Safety and Clinical Excellence, I have the privilege of interacting with thought leaders and senior personnel at healthcare organizations on the leading edge of efforts to improve patient and medication safety. This article presents my perspective on what hospitals need to consider as they attempt to the pervasive problem of “alarm fatigue” and the need for improved alarms management. Other companies are also working to address the need for improved alarms management, but I have little knowledge of these efforts and cannot report on what is being developed or put into practice.
Vanderveen, T. (2014). Alarm management: First things first. Patient Safety & Quality Healthcare, 11(6), 38–45.
AAMI. (2011). Healthcare Technology Foundation 2011 National Clinical Alarms Survey. Retrieved October 20, 2014, from http://www.aami.org/htsi/alarms/pdfs/2011_HTFAlarmsSurveyFinal.pdf
AAMI. (2014a). New national coalition tackles alarm management in healthcare [Press release]. Retrieved October 20, 2014, from http://www.aami.org/news/2014/043014_Press_New_National_Coalition_Tackles_Alarm_Management.html
AAMI (2014b). Using data to drive alarm system improvement efforts: The Johns Hopkins Hospital experience. AAMI Foundation HTSI, Safety Innovations. With thanks to Cvach M, Currie A, Sapirstein A, et al. Retrieved October 20, 2014, from http://www.aami.org/htsi/SI_Series/Johns_Hopkins_White_Paper.pdf
CMS (2014). HCAHPS: Patients’ perspectives of care survey. Retrieved October 20, 2014, from http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/HospitalHCAHPS.html
ECRI Institute (2009). Top ten healthcare technology hazards for 2010. Health Devices, 38(11), 1-10. Retrieved October 20, 2014, from www.ecri.org/hazards2010.
ECRI Institute. (2010). Top ten healthcare technology hazards for 2011. Health Devices, 39(11), 386-398. Retrieved October 20, 2014, from www.ecri.org/hazards2011
ECRI Institute. (2011). Top ten health technology hazards for 2012. Health Devices, 40(11), 1-16. Retrieved October 20, 2014, from www.ecri.org/hazards2012
ECRI Institute (2012). Top ten health technology hazards for 2013. Health Devices, 41(11), 1-23. Retrieved October 20, 2014, from www.ecri.org/hazards2013
ECRI Institute (2013). Top ten health technology hazards for 2014.
Health Devices, 42(11), 1-13. Retrieved October 20, 2014, from from: www.ecri.org/hazards2014
ECRI Institute (2014a). Alarm load. In: The alarm safety handbook: Strategies, tools and guidance. Plymouth Meeting, PA: ECRI Institute.
ECRI Institute. (2014b). ECRI Institute’s Alarm Safety Handbook helps hospitals minimize a top threat to patient safety [News release]. Retrieved October 20, 2014, from https://www.ecri.org/Press/Pages/ECRI-Institutes-The-Alarm-Safety-Handbook-Helps-Hospitals-Minimize-a-Top-Threat-to-Patient-Safety.aspx. Accessed October 20, 2014.
Graham, K. C., & Cvach, M. (2010). Monitor alarm fatigue: Standardizing use of physiological monitors and decreasing nuisance alarms. American Journal of Critical Care, 19(1), 28-34. doi:10.4037/ajcc2010651.
The Joint Commission. (2013a, April 8). Medical Alarm Safety in Hospitals. Sentinel Event Alert, 50, 1-3 Retrieved October 20, 2014, from http://www.jointcommission.org/assets/1/18/sea_50_alarms_4_5_13_final1.pdf
The Joint Commission. (2013b, July). The Joint Commission announces 2014 National Patient Safety Goal. Joint Commission Perspectives, 33(7), 1-4. Retrieved October 20, 2014, from http://www.jointcommission.org/assets/1/18/jcp0713_announce_new_nspg.pdf
Kowalczyk, N. (2011, November 28). Suit over monitor settled. The Boston Globe. Retrieved October 20, 2014 from http://articles.boston.com/2011-11-28/news/30451184_1_alarm-fatigue-false-alarms-alert-nurses
Vanderveen T. (2014) An early look at infusion pump alarms and alerts data. Presented at the AAMI National Coalition for Alarms Management, Arlington, VA. April 25, 2014.
Welch J. (2012). Alarm fatigue hazards: The sirens are calling. Patient Safety & Quality Healthcare, 9(3), 26–33. Retrieved October 20, 2014, from http://psqh.com/alarm-fatigue-hazards-the-sirens-are-calling