infusing patients safely: priority issues from the aami
TRANSCRIPT
About AAMI
The Association for the Advancement of Medical Instrumentation(AAMI), a nonprofit organization founded in 1967, is a diverse allianceof more than 6,000 members from around the world united by onecritical missionâsupporting the healthcare community in thedevelopment, management and use of safe and effective medicaltechnology.
AAMI serves as a convener of diverse groups of committedprofessionals with one common goalâimproving patient outcomes.AAMI also produces high-quality and objective information onmedical technology and related processes and issues. AAMI is not anadvocacy organization and prides itself on the objectivity of its work.
About FDA
The U.S. Food and Drug Administration (FDA) is an agency within theU.S. Department of Health and Human Services. FDA is responsiblefor protecting the public health by assuring the safety, efficacy, andsecurity of human and veterinary drugs, biological products, medicaldevices, our nationâs food supply, cosmetics, and products that emitradiation, and by regulating the manufacture, marketing, anddistribution of tobacco products.
FDA is also responsible for advancing the public health by helping to speed innovations that make medicines and foods more effective,safer, and more affordable; and helping the public get the accurate, science-based information they need to use medicines and foods,and to reduce tobacco use to improve health.
Š 2010 AAMI
Permission is granted to distribute or reproduce this report in its entirety for noncommercial and educational purposes. For other uses, or to orderreprints of this report, contact AAMI's Joe Bremner at [email protected].
Call to Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 1
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 3
Clarion Themes and Priority Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . page 5
Where We Go From Here . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 37
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 39
Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 40
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 41
Contents
A Call to ActionDear Colleagues,
On Oct. 5â6, 2010, a remarkable group of diverse experts came together to change the world of infusion devices.The AAMI/FDA Infusion Device Summit opened with this quote from Margaret Mead: âNever doubt that a groupof thoughtful, committed citizens can change the world. Indeed, it is the only thing that ever has.â Looking back,that quote captured the essence and mood of the summit.
The spark for hosting this unprecedented, groundbreaking event was the result of listening well. FDA listenedwell to the barrage of reports it received about too many adverse incidents involving infusion devices. Theacademic and industry co-chairs of AAMIâs Infusion Device Standards Committee listened well to their instincts:the international standard up for revision needed more work than simply an up or down vote. AAMI as anorganization listened well to the need for a neutral convener to bring the right people together from across the entire healthcare community. Industry listened well by answering the call to actively engage in the event.The clinician, clinical engineering, and expert consultant community listened well by giving two days of theirtime, expertise, and passion for patient safety for the âgood of the order.â The time was right, and from listeningwell, the event was conceived and took shape.
The spark for what happened during the event also was the result of listening well by the 330 people in theroom. Attendees heard challenges presented by speakers and amplified by audience discussion. Questions,audience comments, suggestions, frustrations, and opinions were brought to the floor. Participants togetherfulfilled the challenging assignment of developing a list of prioritiesâ13 in allâthat all agreed must beaddressed.
What made the event even more remarkable was the overwhelming commitment of attendees that, as acommunity, they would continue to work together on implementing action plans based on the agreed-uponpriorities. Each of these committed professionals, reaching across all parts of healthcare, deserves a round ofapplause for this great work product. Thank you.
The challenge for all of us will be to keep the momentum going, to nurture the spark that gave us all the energy to conceive and shape the eventâand to stay actively engaged in addressing the 13 priorities. TheMedical Device Safety Council/Infusion launched by the AAMI Foundation has been tasked with keeping thepassion alive by organizing working groups, overseeing the progression of solutions to the 13 priorities, andworking with the many people who have already volunteered their time and expertise to infusion initiatives. If the going gets tough, weâll reach out for that common bond we all share: improving patient safety.
Sincerely,
Mary Logan Carol L. HermanPresident DirectorAssociation for the Advancement Standards Management Staffof Medical Instrumentation Office of Science and Engineering Laboratories
Center for Devices and Radiological HealthU.S. Food and Drug Administration
Carol L. HermanMary Logan
Priority Issues From the AAMI/FDA Infusion Device Summit | 3
THE LEGACY OF an unprecedented AAMI/FDAInfusion Device Summit could be a more coordinatedand focused approach to solving a multitude ofchallenges with infusion devicesâsome of the mostwidely used medical technologies in healthcare.Ultimately, this effort could result in improved patientsafety from drug infusions.
That, indeed, is the shared expectation of 330 summitparticipants who gathered at FDA headquartersoutside Washington, DC, in October 2010 to identifyand prioritize the most salient issues with infusiondevices. The summit drew a broad range ofprofessionals, including physicians, nurses,pharmacists, clinical engineers, biomedicaltechnology professionals, human factors engineers,manufacturers, academicians, regulators, andorganizations that represent them.
Framed by expert presentations, summit participantsspent two days building consensus for 13 priorityissues that they believe are critical for improvingpatient safety. The range of issuesâwhichencompass incident reporting, design, engineering,manufacturing, and use of infusion devicesâcallsattention to the increasing complexity of infusiondevices. At the same time, these issues gainedtraction because summit participants believe thehealthcare community is capable of addressing themnow with innovative solutions.
At the summit, AAMI announced the launch of asafety council to spearhead action on the priorityissues. Shortly after the summit, AAMIâs InfusionDevice Standards Committee grouped the 13 priorityissues into five clarion themes, which serve as a callto action for improving infusion devices.
Clarion Themes and Priority Issues
1. Standardize systems and processes forreporting, aggregating, and analyzing infusiondevice incidents.
Priority Issues:
1. There is a poor (incomplete and inadequate)system for reporting aggregate state andnational data about adverse events (e.g.,MAUDE [Manufacturer and User FacilityDevice Experience] and PSOs [Patient SafetyOrganizations]).
a. There is a lack of standardization tosupport data aggregation.
2. The reported incidents do not convey thebigger picture in terms of the volume ofincidents involving infusion devices. Userfacilities are encouraged, but not required, to report âclose callsâ and ânear missesâ and to determine their root causes.
3. There is often an inability by manufacturersto determine root cause of infusion deviceincidents due to difficulty accessing andanalyzing incident data from all sources. This also applies to continuous qualityimprovement (CQI) reporting.
4. There is no process for collaborative failureanalysis.
a. There is no safe space for disclosing oraccessing information about infusiondevice incidents or problems. PatientSafety Organizations (PSOs) should beconsidered.
Executive SummaryâThe most important aspect of the summit is the hugemultidisciplinary turnout. There is no way this issuecan sink back into obscurity.â
âNat Sims, M.D., anesthesiologist and physician advisorin biomedical engineering at Massachusetts
General Hospital in Boston and co-chair of AAMIâs Infusion Device Standards Committee
4 | Infusing Patients Safely
2. Improve the integration of infusion deviceswith information systems and drug libraries.
Priority Issues:
5. There is incompatibility across devices andwith systems (e.g., consistent bar coding,wireless, power supply, and healthinformation technology [HIT] systems). Theunavailability of wireless in a natural disastershould be considered.
6. There is a lack of formulary and standards for drug libraries, including standardizationof drug concentrations and transparency(e.g., for sharing of drug libraries betweenfacilities).
7. Uploading, managing, and maintaining druglibraries can be difficult.
a. There is a lack of coordination betweenpump requirements and hospitalcapabilities.
b. There is a steep learning curve forconfiguring and managing drug libraries.
c. There is difficulty in managing the samedrug used in multiple units in multipleways.
3. Mitigate use errors with infusion devices.
Priority Issues:
8. A high percentage of sentinel/adverse drugevents (ADEs) are due to use errors. It isimperative to figure out how to developdesign safety features that make it easy forthe user to do the right thing. Applicablehuman factors, automatic identification (e.g., bar coding), and the value of all thesteps involved in drug administration shouldbe considered.
9. There is a lack of standardization ofterminology used in infusion systems(upstream and downstream devices)âand a clear need for the same wording, same spelling, etc., across the process,devices, containers, etc.
10. There is a lack of knowledge/familiarity with infusion devices and a lack of effectivetraining in their useâfrom bothmanufacturers and facilities.
4. Improve management of multiple infusions.
Priority Issues:
11. There is difficulty in infusion line manage-mentâincluding containers, manifolds,catheters, and transportâreflecting thecomplexity of multiple infusions, includingsecondaries, disposables, etc.
5. Reconcile challenges and differences in the useenvironments of infusion devices.
Priority Issues:
12. Alarm management is not effective.
a. There are high numbers of false alarms,which also can lead to true alarms beingignored (e.g., air).
b. Alarms are difficult to prioritize.
c. It is unclear how to resolve alarm issues.
13. Injuries are caused by a lack of differentiationbetween the use of infusion devices inhospitals and in other environments (e.g., home use). Products designed for the hospital environment are being used in home environments (and vice versa). There are design and user issues anddifferences among home, hospital, and other environments.
About This Report
By design, this report is not a chronological summaryof summit proceedings. Summit participants madeclear that the most welcome and productive nextstep would be concerted action to address thepriority issues. Thus, this report is organized aroundthese priority issues, with the expectation that it willserve as a touchstone document for moving forward.
The report does highlight the summit presentationsand discussions, along with expert perspectivessolicited after the summit and an update of theinfusion systems safety councilâs progress indeveloping a comprehensive, collaborative, andmultidisciplinary action plan to tackle the priorityissues.
An important disclaimer: This publication reportson the 13 priority issues developed by consensusat the summit, summarizes summit presentations,and provides additional context from experts.The 13 priority issues have not been endorsed by AAMI, FDA, or any of the summit sponsors orsupporting organizations. The views expressedby individuals in summit presentations andexpert perspectives do not necessarily representthese organizationsâ views.
More Summit Information on AAMI Web Site
The summit agenda, PowerPoint presentations of summitspeakers, reference materials, and updates are posted onthe AAMI Web site. Visit the Web site to view or downloadthis information.
www.aami.org/infusionsummit
The Scope of the Problem
The vast majority of patients who spend even a few hours in a hospital encounter one of the most widely used medical technologies in healthcare:infusion pumps. Increasingly, computer-controlledsmart pumps can be programmed to delivercontrolled amounts of painkillers, antibiotics, insulin,chemotherapy drugs, nutrients, or other fluids. Smart pumps also can keep electronic records ofinfusions, which are captured in the pumpsâ software.Infusion devices are used in homes and otherhealthcare settings as well.
When they work as intended, infusion pumpssupport the âfive rightsâ of medication safetyâright medication, right dose, right time, right route,and right patient. In fact, infusion pumps can provide important health benefits and reducemedical errors, according to summit presenterWilliam Maisel, Deputy Center Director for science atthe FDAâs Center for Devices & Radiological Health.
But infusion errors and pump failures can causeserious harm, and even death, to patients. BetweenJan. 1, 2005, and Dec. 31, 2009, more than 56,000adverse events and 710 deaths associated withinfusion devices were reported to FDAâmore thanfor any other medical technology. During this period,
there were 87 pump recalls.
âAdverse events are amplified because of the numberand frequency of use of infusion pumps,â Maisel said.âFailures can occur wherever pumps are used, withevery type of pump, with any manufacturer. Manyproblems are due to deficiencies in design andengineering. But itâs not just an issue of devices, itâs about users and user interfaces.â
In response to its safety concerns, FDA in April 2010launched an infusion pump improvement initiative,held a public workshop, published a white paper,issued draft guidance for manufacturers, and createda Web site devoted to infusion safety.
www.fda.gov/InfusionPumps
A Call to Action and 13 Priority Issues to Address
Framed by expert presentations, summit participantsspent two days building consensus on 13 priorityissues that they believe are the most critical forimproving patient safety. The range of issues callsattention to the increasing complexity of infusiondevices and the healthcare environments in whichthey are used.
At the same time, these issues gained tractionbecause summit participants believe the healthcarecommunity is capable of addressing them now withinnovative solutions to improve incident reporting,design, engineering, manufacturing, and use ofinfusion devices. In the words of summit presenterPat Baird, Systems Engineer at Baxter International,Inc., a manufacturer of infusion devices, and co-chairof AAMIâs Infusion Device Standards Committee,going after the âlow-hanging fruitâ is a viable place to start.
At the summit, AAMI announced the formation of asafety council to spearhead action on the priorityissues. Shortly after the summit, AAMIâs InfusionDevice Standards Committee grouped the 13 priorityissues into five clarion themes, which serve as a callto action for improving infusion devices.
The clarion themes and priority issues are presentedbelow, with highlights from summit presenters andexpert perspectives on the issues and potentialsolutions.
Clarion Themes and Priority Issues
Priority Issues From the AAMI/FDA Infusion Device Summit | 5
6 | Infusing Patients Safely
âWhat you donât know, you canât fix.â
âBryanne Patail, biomedical engineer at the Veterans Health Administrationâs
National Center for Patient Safety
CALL TO ACTION: NEXT-GENERATION REPORTING
CLARION THEME 1Standardize systems and processes forreporting, aggregating, and analyzinginfusion device incidents.
PRIORIT Y ISSUES:
1. There is a poor (incomplete and inadequate)system for reporting aggregate state and national data about adverse events (e.g., MAUDE [Manufacturer and User Facility DeviceExperience] and PSO [Patient SafetyOrganizations]).
a. There is a lack of standardization to supportdata aggregation.
2. The reported incidents do not convey the biggerpicture in terms of the volume of incidentsinvolving infusion devices. User facilities areencouraged, but not required, to report âclosecallsâ and ânear missesâ and to determine theirroot causes.
3. There is often an inability by manufacturers todetermine root cause of infusion device incidentsdue to difficulty accessing and analyzing incidentdata from all sources. This also applies tocontinuous quality improvement (CQI) reporting.
4. There is no process for collaborative failureanalysis.
a. There is no safe space for disclosing oraccessing information about infusion deviceincidents or problems. Patient SafetyOrganizations (PSOs) should be considered.
Priority Issues From the AAMI/FDA Infusion Device Summit | 7
Understanding the Issues: Reporting Systems and Processes
A frustrating aspect of infusion device incidents isthe dearth of information about themâanswers tothe essential âwho, what, when, where, why, and howâquestions, according to Baird.
Baird shared findings from his examination of asample of incident data from public sources, focusingon serious incidents that resulted in patient harm. He sought to answer basic questions about infusionpumps, including how the broad patterns in thepublic data would inform further investigation intothe design, manufacture, and use of infusion pumps.Specifically, Baird reviewed:
⢠Three years of data from MAUDE (2007, 2008, and2009), FDAâs database of voluntary reports ofmedical device incidents from users, facilities,distributors, and manufacturers, covering patient-controlled analgesia (PCA), elastomeric,enteral, and general-purpose infusion devices
⢠More than five years of information from the FDAClass I recall list, the most serious medical devicerecalls
⢠10 years of research literature, including casestudies, reports, and conference proceedings oninfusion devices
A key takeaway message from this exhaustiveinvestigation is that, even for the most seriousincidents, answers to basic questions are sometimesunknown. A sampling of Bairdâs synthesis of MAUDE-reported incidents with PCA infusion pumps, shownin Figure 1, highlights this shortcoming. Bairdpresented similar findings about other types ofinfusion devices.
MAUDE is a self-reporting system that might notcontain all the necessary information about medicaldevice incidents, Baird noted. In addition, for theincident reports included in his study, MAUDE coversonly device-specific issues, which means that relatedissuesâsuch as misconnections, pharmacy errors,and adverse drug interactionsâwill not show up.(However, FDA said that related issues such as theseare included in other kinds of incident reports.)Rarely is there any âwhoâ associated with deviceincidents in MAUDE reports, Baird said. The quality ofreports varies by device type and by type of incident,with injury reports providing much more informationthan death reports.
8 | Infusing Patients Safely
Source: Pat Baird, âIncident Data: Filling the Gaps, Identifying the Value, and PrioritizingNeeded Information.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 6, 2010.
Figure 1. A Synthesis of Patient-Controlled Analgesia (PCA) Infusion Pump Incident Reports to MAUDE, 2007â09
CALL TO ACTION: NEXT-GENERATION REPORTING
Priority Issues From the AAMI/FDA Infusion Device Summit | 9
There are shortcomings to other sources of informa-tion as well, Baird said. Published case reports coverspecific incidents, with no aggregation or analysis of data into broad categories of information.Conference proceedings often make best-practicerecommendationsâfor unstated problems.
Even with incomplete and inadequate data, however,there is important information to be gleaned fromMAUDE data. Figure 2 shows the identified causes of harm and average yearly tally of adverse patientoutcomes for PCA, enteral, and general-purposedevices over the three-year period Baird examined.
Summit participants elaborated on the inadequateand incomplete system and processes for reporting,aggregating, and analyzing infusion device incidentswith these observations, which are now priorityissues:
⢠User facilities are encouraged, but not required, to report âclose callsâ and ânear misses.â Thus, thereported numbers of incidents do not convey theâbigger pictureâ of the true volume of infusiondevice incidents.
⢠The shortcomings of the incident report systemmake it difficult for manufacturers to accessinformation from all sources to conduct rootcause analysis of infusion device incidentsâa problem that applies to continuous qualityimprovement information captured by smartpumps.
⢠There is no process or âsafe spaceâ for collabor-ative failure analysisâcareful, after-the-factexamination of infusion pump incidents by multi-disciplinary teams in which the goal is identifyingand resolving issues rather than placing blame.
Source: Pat Baird, âIncident Data: Filling the Gaps, Identifying the Value, and PrioritizingNeeded Information.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 6, 2010.
Figure 2. Identified Causes of Harm and Average Annual Number of Reported Incidents: Patient-Controlled Analgesia (PCA), Enteral, and General-Purpose (FRN) Pumps, 2007â2009
10 | Infusing Patients Safely
Addressing the shortcomings of the systems andprocesses for reporting, aggregating, and analyzinginfusion device incidents must be a sharedresponsibility among all stakeholders, not just FDA.
This is because MAUDE is a reporting system,intended to signal safety issues, but not to determineor convey root causes of incidents. Not all incidentsare required to be reported. And the information inthe MAUDE database is only as good as the reporterswho submit it. FDA receives two kinds of medicaldevice reports:
⢠Mandatory medical device reports, or MDRs,submitted by the medical device industry and by certain healthcare facilities by three mainreportersâmedical device manufacturers,medical device importers and user facilities, suchas hospitals, which use the devices. The term âuser facilityâ actually describes five types offacilities responsible for mandatory reporting:hospitals, ambulatory surgical facilities, nursinghomes, outpatient diagnostic facilities, andoutpatient treatment facilities.
The federal Food, Drug, and Cosmetic Act (FD&CAct) gives FDA the authority to collect MDRs. The reporting requirements, which are found inChapter 21 of the Code of Federal Regulationsbook at Part 803, are implemented based on thelaws in Section 519 of the FD&C Act.
⢠Voluntary reports submitted by reporters whodo not have a mandatory requirement to tell FDAabout device-related deaths, serious injuries, ormalfunctions, but are encouraged to do so.
According to FDA, near misses and events thatcould result in harm can be reported using thismethod. Additionally, the Medical Product SafetyNetwork (MedSun) is an adverse event reportingprogram launched in 2002 by FDA. The primarygoal of MedSun is to work with the clinicalcommunity to identify and solve problems withthe use of medical devices. FDA also collaborateswith MedSun facilities to more fully understandpossible emerging public health signals. MedSunparticipants are educated and encouraged tovoluntarily report problems with devices, such as
CALL TO ACTION: NEXT-GENERATION REPORTING
Priority Issues From the AAMI/FDA Infusion Device Summit | 11
âclose calls,â events that could result in harm, andother safety concerns in addition to the death andserious injury events required.
Both voluntary and mandatory reports containinformation that helps FDA monitor the safety of medical devices. At any time, if a post-marketanalyst requires more detail regarding a report, FDA may obtain additional information to fill in anyobserved gaps or missing pieces in reporting.
FDA encourages manufacturers to engage in activeconversations with the agency to obtain guidanceabout their reporting requirements. In addition, FDA strives to provide outreach through directcommunication with manufacturers and throughtraining materials and guidance documents postedon the FDA Web site.
www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/default.htm
FDA also supports medical device reporting by manufacturers, voluntary reporters, importers, and user facilities with training presentations, alsoavailable on the FDA Web site. Two presentationsprovide guidance and technical information abouthow to use MAUDE, search for adverse eventinformation and submit MDRs to FDA electronically.
www.fda.gov/Training/CDRHLearn/ucm162015.htm
Finally, FDA encourages collaborative action to determine root causes of adverse events.Manufacturers should ensure that staff performingthe root cause analysis are communicating andcollaborating with their customers and the deviceusers at the time of failure to assist in theidentification of the specific cause(s) of the failure.The manufacturerâs team performing the root causeanalysis should contain knowledge from multipleengineering and relevant clinical disciplines. FDA hasposted a number of examples of reported infusionpump problems on its Web site.
www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/GeneralHospitalDevicesandSupplies/InfusionPumps/ucm202496.htm
SEVERAL SUMMIT PARTICIPANTS envision anext-generation system for incident reporting,aggregation, and analysis as robust as the NationalTransportation Safety Boardâs. NTSB is on the sceneof every major transportation incident with a teamof trained investigators, who use a standardprotocol to get to root causes.
Incident records, preliminary and final reports, and conclusions are accessible on a searchabledatabase on a public Web site.
www.ntsb.gov
A Potential Process for Next-Generation Reporting: The National Transportation Safety Board
12 | Infusing Patients Safely
Root Cause Analysis Shapes Safety Initiatives
AT THE VETERANS HEALTH ADMINISTRATION(VHA), root cause analysis is practically an art form.The VHA has conducted more than 12,000 systematicroot cause analysesâor more than 1,000 a yearâsince 1999.
That year, the VHA established the National Centerfor Patient Safety (NCPS) to develop and nurture
a culture of safety, including a non-punitiveenvironment for reporting medical device and otherpatient safety incidents. Since then, âclose callâreports increased 900-fold. âChanging the culture of safety in an organization is like an engine thatcontinues to propel safety,â said summit presenterBryanne Patail, biomedical engineer at NCPS.
The goal of root cause analysis is to find out whathappened, why it happened, and what to do toprevent it from happening again. âWhat you donâtknow, you canât fix,â Patail said.
Some 978 of the root cause analyses conducted over 11 years involve medical devices. More than 13 percent of these, 129 in all, pertain to two types of infusion pumpsâ60 on general-purpose pumpsand 69 on PCA pumps. Based on these root causeanalyses, the VAâs integrated product teamrecommended that PCA pumps with an integratedend tidal CO2 monitor is the pump of choice, becauseuse of this technology could have prevented morethan 60 percent of adverse events related to PCApumps.
NCPS has developed a standard protocol, which isspelled out in detail on its Web site and summarizedon laminated, spiral-bound, and portable NCPS Triage Cardsâ˘, to support patient safety teams inconducting root cause analyses. Triage questionshelp teams gather information from interviews,documents, and devices and delve into possible (and often multiple) causes for incidents. Probingquestions examine relevant issues in key areas:
⢠Human factors (communication, training, fatigue,and scheduling)
⢠The environment and equipment
⢠Rules, policies, and procedures
⢠Barriers and controls
Based on their findings, the teams recommendactions to prevent or minimize future adverse eventsor close calls. NCPS tracks the implementation ofthese follow-up actionsâand has enough outcomedata to be able to categorize them as âstrong,intermediate, or weak fixes.â The strongest fix forboth general purpose and PCA pumps is standardiz-ing processesâprotocols, clinical guidelines, ordersets. New device purchases and standardization ofequipment make the strong-fix list as well.
Root cause analysis, follow-up actions, and outcomestatements indicate that there are opportunities forimprovement by all stakeholders. âOpportunities toprevent problems from happening in the first placeexist at many levels,â Patail said. âI am a strongsupporter of âprevention is better than the cure.â We cannot afford to run to fire calls all the time.â
LEAD-USER PROFILE: The Veterans Health Administration
CALL TO ACTION: NEXT-GENERATION REPORTING
Priority Issues From the AAMI/FDA Infusion Device Summit | 13
Pat BairdBaxter International, Inc.
Q. What is the âbigger pictureâ that is missingfrom reporting requirements?A. âHealthcare providers are focused on anindividualâs careâas they should be. It shouldnât besurprising that they sometimes donât have the timeto fill out a detailed, free-form incident report. Thisleads to under-reporting and incomplete reporting.To improve public health, we all need that extra levelof detail. We need to find a way to both makereporting easier and to change the perception thatreporting is removed from patient safety. This is notjust an infusion issue; it affects all medical devices.â
Q. From your investigation, what root causesdo you believe deserve more attention?A. âI believe the industry already knew thatâmisprogrammingâ and âuse errorâ were significantissues, but those are very broad categories and need-ed more specificity. This investigation enlightenedme to the fact that âright data in the wrong fieldâ wasa significant type of misprogramming. I also wassurprised at the number of incidents where physicaldamage to the device wasnât apparent to the user,but resulted in a death or injury. Other types of useerror caused more harm than âphysical damage,â but Idonât want us to forget that not all âuse errorâ isrelated to device programming.â
Q. What roles do you think manufacturers, FDA,and AAMI should play in improving reporting,aggregation, and analysis of infusion deviceincidents over the next year?A. âThere are two things I'd like to see happen:
1) Build confidence in the conclusions from MAUDE.
2) Know what it takes to build a better MAUDE in the future.
âI reviewed three years of MAUDE data and came toone set of conclusions. A review of MedSun* mightcome to the same or different conclusions. Other
databases might have different results altogether.Although I reviewed over 1,000 MAUDE records to bethorough, correlating the MAUDE results with thosefrom other databases would provide an additionallevel of confidence to the MAUDE data. We need tobe sure that we are solving the right problems.
âAlthough we may be looking specifically at infusiondevices in these databases, what we learn isapplicable to nearly all medical devices. Infusionsystems are a microcosm of the broader medicaldevice industry.â
Q. What could be learned from other reportingsystems in the U.S. and abroad? A. âOther reporting systems have their own specificgoals, strengths, and weaknesses. As we look atpumps, letâs also look at how well these reportingsystems are working. How do best-practice hospitalstrack issues internally, and what is the quality of theirresults? Pennsylvania and Minnesota have statewidereporting systems. How well are those systemsworking? MEDMARX and MedSun are national-levelreporting systems worth looking at. Ontario spendsa significant amount of money on healthcareimprovement initiatives. How is its reporting systemset up? What can we learn from the NHS [NationalHealth Service] in the U.K. and AFSSAPS [FrenchHealth Products Safety Agency] in France? Letâs takethis opportunity to truly understand the capabilitiesand limits of our collective reporting systems to helpdesign the next-generation reporting system.â
In response to Bairdâs suggestions, FDA cautionedthat any conclusions drawn from the MAUDEdatabase might not be sound because, as Bairddiscovered, the information submitted is not alwayscomplete or of high quality. Moreover, incidentsreported to FDA do not encompass all incidents thatoccur. In addition, comparing MAUDE data to datafrom other databases could be problematic as well,because different reporting systems collect differentinformation.
EXPERT PERSPECTIVE: Improving Reporting Systems and Processes
* Note: Data from the FDAâs Medical Product Safety Network (MedSun) is included in the MAUDE database.
14 | Infusing Patients Safely
CALL TO ACTION: FROM SMART PUMPS TO INTELLIGENT SYSTEMS
âIV interoperability is cutting-edge technology that we think is really the next additional step for IV medication safety.â
âAmanda Prusch, pharmacist and medication safety specialist at Lancaster General Hospital
CLARION THEME 2Improve the integration of infusion deviceswith information systems and drug libraries.
PRIORIT Y ISSUES:
5. There is incompatibility across devices and withsystems (e.g., consistent bar coding, wireless,power supply, and health information technology[HIT] systems). The unavailability of wireless in anatural disaster should be considered.
6. There is a lack of formulary and standards for drug libraries, including standardization of drugconcentrations and transparency (e.g., for sharingof drug libraries between facilities).
7. Uploading, managing, and maintaining druglibraries can be difficult.
a. There is a lack of coordination between pumprequirements and hospital capabilities.
b. There is a steep learning curve for configuringand managing drug libraries.
c. There is difficulty in managing the same drugused in multiple units in multiple ways.
Priority Issues From the AAMI/FDA Infusion Device Summit | 15
Understanding the Issues: Information Systems and Drug Libraries
Todayâs infusion pumps are complex systems,connected to wired or wireless networks andinformation systems, and, increasingly, to druglibraries and electronic medical records. When thesesystems are well designed and integrated, infusiondevices can:
⢠Improve patient safety
⢠Reduce medical errors
⢠Improve charting
⢠Promote best practices with standardized drugconcentrations
⢠Increase the efficiency of medication use, giving clinicians more time to focus on patients
⢠Facilitate technical support
⢠Capture valuable information
⢠Support information sharing
⢠Streamline information processing
But the bane of all this connectivity is that, in manyhealthcare settings, the technology has not caught upto clinical needs. The many kinds of infusion devicesand systems are not always interoperable, which cancompromise device availability, clinical practices,information collection and useâand patient safety.Healthcare institutions often have to manage thedevelopment of device and system integration andinteroperability into their own information systems,which can be a major undertaking for large, mid-size,and small institutions alike. Wireless technology addsanother layer to this challenge, because robustwireless infrastructures and information security arecritical. Summit participants noted that wirelesstechnology is not always reliableâand it could beunavailable in the event of a natural disaster or otheremergency.
Incorporating drug libraries into this mix offers anadditional set of challengesâand opportunities.Variability among drug formularies (lists of preferreddrugs), infusion concentrations, and dosingparameters used by healthcare institutions are keyissues. Individual institutions must develop their ownparameters for drug librariesâor borrow them fromother facilities. This latter solution might not be anappropriate one, summit participants noted. Theprofiles and needs of different facilities might notmatch. And itâs important for institutions to
understand how the drug libraries they use weredeveloped: the considerations and decisions behinddrug libraries matter. Institutions need to âownâ theirdrug librariesâby knowing whatâs in them and why.
Right now, many healthcare organizations allow awide variety of different infusion concentrationsâfrom five to 45, depending on the drug, according to a survey of pharmacists by the American Society of Health-System Pharmacists (ASHP). ASHP, whichrepresents 35,000 pharmacists in hospitals andhealthcare systems, has been focused on this issuesince a 2008 summit on achieving fail-safe use of IVmedications. ASHP is working on a recommendationfrom that summit to develop nationally standardizeddrug concentrations and dosing units for commonlyused, high-risk drugs used in infusions. Specifically,ASHP aims to:
⢠Standardize infusion concentrations that
â Are limited to as few concentrations as possible
â Are clinically appropriate for the majority of patients
â Include the most commonly infusedmedications
⢠Standardize terminology for dose and rate
⢠Provide evidence-based recommendations forupper and lower dosing and rate limits
⢠Recommend standard procedures for ordering,preparing, and administering drugs
Standard drug concentrations and dosing wouldreduce errors, facilitate technical support, and increaseefficiency, said summit presenter Bona Benjamin,director of medication-use quality improvement atASHP. Use of standardized infusion concentrationscould facilitate faster approval for new drugconcentrations and result in an increased number of products available in ready-to-administer form.
There are barriers to this work, however. For starters,âpatients are not standardized,â Benjamin said, âand we know that itâs neither realistic not clinicallyappropriate to use standard concentrations in allpatients.â Physicians and nurses, not surprisingly,tend to resist changing established practicesâandpractices vary in different care environments, facilities,and geographic regions. For this reason, determiningâbest practiceâ concentrations on a national level is achallenge.
16 | Infusing Patients Safely
Some regional groups have succeeded in this effort,Benjamin said, including the San Diego Patient Safety Consortium, which developed countywidestandardization for 14 hospitals, and the IndianapolisCoalition for Patient Safety, a similar effort for 13hospitals in seven healthcare systems.
Still, Benjamin cautioned, even a consensus-developed, evidence-based, open-source referencefor drug standardization will not be a magic bullet.âWe know that a list of standardized concentrationsis not the same as a drug library,â she said. âHowever,itâs a necessary first step to give practitioners acommon foundation and reference on which to builda smart library.â
A drug library should be âa framework for saferpractice,â according to summit presenter ErinSparnon, senior project engineer at ECRI InstituteâsHealth Devices Group. Well-designed and integratedtechnology features can support the âfive rightsâ ofmedication safety (right medication, dose, time, route,and patient). A drug library can eliminate grossmisprogramming or titration errors, allow for efficient
and correct autoprogramming, and enable bar codereadings of drugs.
âNo single safety technology can be perfect, but ifyou add enough layers of âSwiss cheese,â you mightbe able to prevent more errors,â she said, referring toBritish psychologist James Reasonâs model of systemsafety. In this model, the holes in a slice of Swisscheese represent areas where errors or failures couldoccur. Minimizing the holes and creating multipleoverlapping layers of Swiss cheese decreases thelikelihood that an error will slip through. To preventerrors, ECRI recommends:
⢠A comprehensive dose error reduction system(DERS) with both drug library and log-analysissoftware
⢠A two-way wireless communication and serversystem, capable of interfacing to other systems,including pharmacy and documentation
⢠A pump design that supports future softwareupgrades
CALL TO ACTION: FROM SMART PUMPS TO INTELLIGENT SYSTEMS
WHEN DUKE UNIVERSITY MEDICAL CENTERdecided to install smart pumps in its three-hospitalsystem, it began not with the technology but withthe people who would use it.
The planning started 18 months before theimplementation, with focus and work groups thatinvolved stakeholders throughout the system. âThis required a collaborative effort that spanned the entire culture of the medical center,â said summitpresenter Glenn Scales, certified biomedicalequipment technician and patient safety specialist at Duke. âIt required buy-in. We engaged peoplewho had typically not been involved.â
This extensive planning resulted in a modular smartpump system, with smart pumps that went live in2005, wireless connectivity added in 2007, andpatient-controlled analgesia (PCA) pumps and awireless expansion in 2008. Since then, Duke hasintegrated the drug library into a computerizedphysician order entry (CPOE) systemâanotherbenefit that took months of planningâand added
hard limits on heparin and insulin infusions, madelow-limit adjustments, and minor tweaks andexpansions for new therapies.
âSmart pump technology enabled us to get thingsdone that we had not been able to do before,âincluding consolidating the drug library withstandardized dosagesâmost notably in pediatrics,Scales said. âSmart pumps have had a dramaticimpact on the severity and number of events weâvehad over time.â
Smart pumps have brought new challenges, however.âThe pumps arenât smart enough,â Scales said.People can be âextraordinarily inventiveâ in overridingthe checks that are built into the devices. Crosschecking of infusion settings remains limited. Andthere is a sense of information overload from thecontinuous quality improvement (CQI) data collectedby the infusion devices. Data analysis is highly labor-intensive, he said, and staff time for this process isconstrained.
LEAD-USER PROFILE: Duke University Medical Center, NC
Priority Issues From the AAMI/FDA Infusion Device Summit | 17
Developing a Smart System of Pump Alerts
IT TAKES MORE THAN SMART PUMPS to produce the greatest benefits to patient safety. Ittakes a smart system. At Intermountain Healthcare,which serves the medical needs of Utah andsoutheastern Idaho in 22 hospitals, a smart systemdeveloped in-house is producing results and helpingpeople work smarter as well.
The smart system features wireless connectivity ofinfusion pumps, bedside computers, and nursingstation computers into a central database, which isconnected to electronic medical records, as shown in Figure 3.
The system sets pump alerts for high-riskmedications, including initial start limitsâtypicallyset limitsâand rate change limits that take intoaccount unit protocols, patient work flows, andpatient variables, such as nutrition and insulintherapy. Simple or complex pump alerts can bedeveloped, with information about the alertsdelivered via e-mail, pages, or text messages tocomputers or mobile devices, according to summitpresenter Rick Carlson, pharmacist and boardcertified pharmacotherapy specialist atIntermountain Medical Center. Rules for the alertingdatabase can be modified easily with the help ofmedical informatics and computer programmers.
Analysis of the smart system over almost two yearsshowed that 137 alerts saved patients from harm.
LEAD-USER PROFILE: Intermountain Healthcare, UT
Source: Rick Carlson. âEnhanced Notification of Infusion Pump Programming Errors.âPresentation at the AAMI/FDA Infusion Device Summit, Oct. 5, 2010.
Figure 3. Intermountain Healthcareâs Smart System
Interoperability: The Next Step in Medication Safety
LANCASTER GENERAL HOSPITAL harnessed thepower of emerging technologiesâand its vendorsâto create a seamless, interoperable system thatsupports medication safety. The Ăźbersystem linksinfusion pumps and the people who use them with a host of data systems, including medication orders, a drug library, electronic medical administration(eMar) records, bar code medication administration(BCMA), and reporting, as shown in Figure 4.
âItâs all wirelessâand hospitals, if youâre not alreadythere, you should be,â said Amanda Prusch, pharmacistand medication safety specialist at Lancaster General.The two-way system builds in checks and counter-checks in the clinical workflow to satisfy the fiverights of patient safety (right medication, dose, time,route, and patient), as shown in Figure 5.
âDrug library compliance pretty much doubled afterintroducing IV interoperability,â Prusch said. âWethink IV interoperability is really the next additional
step for IV medication safety. It is cutting-edgetechnology and itâs still evolving. This was a hugecollaborative effort with our pump vendors, barcoding vendors, pharmacists, information services,and biomedical engineering.â
The system also streamlined the workflow andpotential for errors by moving from manual toautomated programming, reducing a 17-step processto seven steps and reducing nursing time for pumpsetup by 24.8 percent, said summit presenter Tina Suess, a nurse and system administrator at thehospital. The hospital conducted a validation studycomparing manual and automated programming bynursesâand found that manual programming wasreplete with errors. âWe wanted to make it safer forour nurses to administer IV medications,â she said.âAnd we wanted to prove that this was worth ourtime and money.â
Figure 6 shows the challenges of manual pumpprogramming and the advantagesâno longer inquestionâof the interoperable system.
18 | Infusing Patients Safely
CALL TO ACTION: FROM SMART PUMPS TO INTELLIGENT SYSTEMS
LEAD-USER PROFILE: Lancaster General Hospital, PA
Source: Amanda Prusch and Tina Suess. âIV Interoperability: Smart Pump and BCMAIntegration.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 6, 2010.
Figure 4. The Data Flow in an Interoperable System
Priority Issues From the AAMI/FDA Infusion Device Summit | 19
Source: Amanda Prusch and Tina Suess. âIV Interoperability: Smart Pump and BCMAIntegration.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 6, 2010.
Figure 6. A Comparison of Manual Programming and IV Interoperability
Source: Amanda Prusch and Tina Suess. âIV Interoperability: Smart Pump and BCMAIntegration.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 6, 2010.
Figure 5. Clinical Workflow Builds in Medication Safety Checks
20 | Infusing Patients Safely
CALL TO ACTION: FROM SMART PUMPS TO INTELLIGENT SYSTEMS
Alan LipschultzChristiana Care Health System
WHAT COULD THE HEALTHCARE COMMUNITYDO to improve infusion device safety? AlanLipschultz, director of clinical engineering atChristiana Care Health System in Delaware, pointedto the common themes of mitigating use error andmaking it easy for people to do that right thing inoffering these three recommendations:
⢠Automate processes. Every step in the infusionprocess should be examined, from the moment a physician writes a drug order until the drug isdelivered to the patient, with an eye towardmaking each step less dependent on humansalone. âThe more you can streamline andautomate processes, the more effective thesystem will be in meeting the five rights ofmedication safety,â Lipschultz said, referring to theright medication, dose, time, route, and patient.
Ideally, bar coding technology should be used ondrugs, drug orders, drug bags, pumps, healthcareproviders, and patients to ensure correct matches.Interoperability of systems, âthe Holy Grail thatpeople have been chasing,â should be the goldstandard.
⢠Standardize devices and protocols. Infusiondevices should be intuitive to use. âAs an example,when you rent a car, you can be sure that theaccelerator is to the right of the brake,â Lipschultzsaid. Standard layouts for the most importantdevice features would make it easier for people touse different devices as well.
Likewise, he said, âphysicians and clinicians havebeen learning that there are benefits to standardprotocols. When a patient has chest pain, thereare standard things that you doâand if you dothem, outcomes go up.â Clinicians andpharmacists could apply this same mindset toevery step of the drug ordering and deliveryprocess. âStandards are doable if people arewilling. Clinical engineering can act as afacilitator, making sure that the right people aretalking, but it has to come from clinicians andpharmacists.â
⢠Expedite fixes. A software glitch in ChristianaCareâs infusion pumps illustrates Lipschultzâsconcern that problems are solved too slowly. The healthcare system, which has more than 1,100 patient beds in two hospitals, rolled outnew infusion pumps between February and Mayof 2009. By July of 2009, the healthcare systemhad identified a problem. The pump softwaresometimes crashed, resulting in what Lipschultztermed âthe blue screen of death,â which requiresrestarting the pump to get it back into service, or,worse, âthe white screen of death.â For the latterfailure, âWeâre hosed. We canât reboot,â he said.
Christiana Care reported these problems in July2009. By October 2009, the manufacturer told the healthcare system that it had resolved theproblem with a software fix. However, because ofFDAâs stepped-up static analysis of softwarechanges in infusion devices, the updated softwaremight not be released until mid-2011. In themeantime, Christiana Care has documented morethan 100 pump crashes, Lipschultz said.
The pump does sound a loud alarm when itcrashes, which so far has helped avert any adversepatient incidents beyond delay of medicationâwhich for some patients could be risky. Still, thisdoesnât make it easier for Lipschultz to sleep atnight. He advocates releasing the software thatfixes the known bug before the thoroughsoftware analysis is completed.
In response to this recommendation, FDA said: âIt would be a premature move to approve anunfinished product without a thorough softwareanalysis. Our role is not to jump to quick fixes, butto actually make sure a fix fixes something. Weâvehad unfortunate experiences where we have triedto be speedy and the fixes were not beneficial.â
EXPERT PERSPECTIVE: Automate, Standardize, Expedite
Priority Issues From the AAMI/FDA Infusion Device Summit | 21
Glenn ScalesDuke University Medical Center
SMART PUMPS COLLECT A TROVE OF DATA. But they donât yet provide adequate tools foranalyzing this data, according to Glenn Scales ofDuke University Medical Center.
Healthcare systems can review the continuousquality improvement (CQI) data accumulated bysmart pumps on a regular basis. Making sense of this data is a challenge, however.
âWhen we pull the data out of the pumps, we get a spreadsheet thatâs about 30 columns wide with85,000 rows of data,â Scales said. âItâs hard to find time with the right kind of people to pull the datatogether and analyze this information. We spendhours trying to find patterns in this data.â
While there are some software tools that supportdata analysis, theyâre not yet powerful enough, he said. It takes human beingsâphysicians, nurses,pharmacists, and other healthcare professionals withan understanding of clinical practices in specificsettingsâto spot institutional issues. Even then, âthe data is interesting, but it never really answers the question, why? The software hasnât kept up.Thereâs not enough artificial intelligence to do whatthe human brain can do.â
EXPERT PERSPECTIVE: Develop Artificial Intelligence to Analyze Data
22 | Infusing Patients Safely
CALL TO ACTION: A SYSTEMS APPROACH TO USERS AND USER INTERFACES
âWe need to design safety features that make it easyfor the user to do the right thing.â
âMatthew B. Weinger, M.D., Norman Ty Smith chair in patient safety and medical simulation and professorof anesthesiology, biomedical informatics, and medicaleducation at Vanderbilt University School of Medicine
CLARION THEME 3Mitigate use errors with infusion devices.
PRIORIT Y ISSUES:
8. A high percentage of sentinel/adverse drugevents (ADEs) are due to use errors. It isimperative to figure out how to develop designsafety features that make it easy for the user to dothe right thing. Applicable human factors,automatic identification (e.g., bar coding), and thevalue of all the steps involved in drugadministration should be considered.
9. There is a lack of standardization of terminologyused in infusion systems (upstream anddownstream devices)âand a clear need for thesame wording, same spelling, etc., across theprocess, devices, containers, etc.
10. There is a lack of knowledge/familiarity withinfusion devices and a lack of effective training intheir useâfrom both manufacturers and facilities.
Priority Issues From the AAMI/FDA Infusion Device Summit | 23
Understanding the Issues: Use Errors
Smart pumps are neither as ingenious as humans nor as simple for intuitive use as they should be indemanding healthcare environments, summitpresenters and participants emphasized.
Consequently, âuse errorsâ in relation to infusiondevices is a contentious topic in the healthcarecommunity, especially for clinicians. Their point: the majority of use errors are made unwittingly, notwith malice or willful disregard of potential patientsafety risks.
âI donât think there are many people who come towork saying, âIâm going to try to disable the safetyfeatures,ââ said summit presenter Matthew B. Weinger,co-chair of AAMIâs Human Factors EngineeringCommittee; professor of anesthesiology, biomedicalinformatics, and medical education at VanderbiltUniversity Medical Center; and staff physician at theVA Tennessee Valley Healthcare System.
Thus, summit presenters and participants argued for more user-friendly infusion devices that make iteasier for people to do the right thing. Summitpresenter Peter Doyle, human factors engineer at The Johns Hopkins Hospital, shared a sampling offindings from a user interface study, culled from atask analysis of nurses using large-volume infusionpumps, interviews and discussions with experts, fieldobservations, and a literature review. Inclusion of asystems engineer from The Johns Hopkins AppliedPhysics Lab, Alan Ravitz, and clinical personnel in thisstudy helped to assure a broad range of issues wasaddressed. The findings are presented below:
⢠Medication labeling varies in format and contentfrom pumps to IV bag labels to medical recorddocumentation. Such variation increases risk ofprogramming errors.
⢠Hardware design enables incorrect tubinginstallation and removal of cassettes in a mannerthat invites risk.
⢠Controls are not standardized. Keypad controls,including number layout and decimal pointplacement, vary from device to device. âSoftâcontrols such as channel selections are notproperly associated with labels in some instances.Functional grouping of controls and displays via
color, for instance, is not used to best effect. Pump controls are difficult to read in darkenvironments.
⢠Information displays are hard to read and themarquee style display takes too much time topresent needed information. While doseinformation is more important, rate is displayedrather than dose. Pressure history is not providedfor fluid delivered. Indications of occlusion areoverlooked because they are not intuitive. Cuesto distinguish between similar drug names areinsufficient.
⢠Battery status is only indicated when there is alow-battery alarm.
⢠Drug lists showing multiple concentrations mayrequire more than one screen to display alloptions. Without a strong cue that there are moreoptions on subsequent screens, selection ofwrong concentrations has become a commonerror.
⢠Programming prompts can result in wrongdoses. Prompts to enter rate or volume to beinfused (VTBI) come before prompts on dose,which have resulted in incorrect rate calculations.Sufficient confirmation information is notdisplayed.
⢠Failure to power the pump down or select anew patient profile could result in incorrect drugdelivery.
A Definition of âUse Errorâ
⢠Any device-related event in which the userâdevice interaction did not conform to eitherthe userâs expectations or the designerâsintentions.
⢠The only use errors that are outside the scopeof discussion are criminal intention orintentional negligent misuse.
Consistent with ANSI/AAMI/IEC 62366:2007, Medicaldevices: Application of usability engineering to medicaldevices, an FDA-approved document
24 | Infusing Patients Safely
CALL TO ACTION: A SYSTEMS APPROACH TO USERS AND USER INTERFACES
⢠Drug calculations required of cliniciansintroduce the potential for errors in rates.Programmable rates do not allow those neededfor bolus dosing of specific drugs.
⢠Timing issues can impede drug delivery. Thepower-up cycle and input of user recognition datacan delay pump startup and drug administration.Pump menus do not include options to selectsmaller bag sizes so that the time of alerts isaccurate for those bag volumes. There is notimely notification for preparation of subsequentmedication bags. There is no delayed start optionto support nursesâ workflow. Sequences to correctprogramming errors may delay drug delivery.
⢠Error checking may be inadequate in someinstances. A dose error reduction system (DERS) is not a fail-safe means of controlling dose. While it prevents extreme programming errors,
it does not take patient factors into account. The pump display can be deceptive in indicatingwhich medication is being delivered; the drugprogrammed might not match the bag. Barcoding the bags or networking the physicianorder entry (POE) system to the pump couldprovide this check. Pumps do not supportprogramming checks by requiring a secondclinician to repeat (match) the inputs. Proceduralcontrols to assure the five rights of medicationsafety are not always sufficient. Errors in weight-based calculations could be reduced if patientweights from smart patient beds werecommunicated to the electronic health record.
⢠Drug library options may differ among hospitalunits, according to the level of care and patientneeds, but not all options are appropriate for allunits. The range of available patient profilesshould be limited, depending on the unit.
Source: Peter Doyle and Alan Ravitz. âStandardization of the User Interface.â Presentation at the AAMI/FDA Infusion Device Summit, Oct. 5, 2010.
Figure 7. Good User Interface Design Addresses Factors That Shape Performance
Priority Issues From the AAMI/FDA Infusion Device Summit | 25
⢠Alarm volume changes affect signal tones(pitch). No notification is provided when an audiosignal is locked out. Only one alarm signal isprovided for all alarms, regardless of differences incriticality. Timing for âcallbackâ notification is after,not before, infusion completion. Screen test andalarm test functions are not provided.
Doyle recommended a systems engineeringapproach to addressing these and other safety issues.The user interface, device design, human behavior,and the patient care environment are not mutuallyexclusive, as shown in Figure 7, indicating a need toaddress issues via a comprehensive systemsengineering approach. âIt would be much preferred,âhe said, ânot to have to rely on training tocompensate for interface issues that could beaccomplished with design.â
Weinger elaborated on user interface considerationsfrom a clinicianâs perspective. He noted that a userinterface is any aspect of a device with which ahuman can interact, including software, hardware,documentation, labeling, and packaging. Usersinclude clinicians, lay care providers, cleaning andmaintenance staff, and patients.
User interfaces should be informed by humanfactorsâthe formal study of human interactions with tools, devices, and systems to enhance safety,efficiency, and satisfaction. Human factors research is commonly applied in a wide range of domains,such as the airline, energy, and consumer productsindustries. Use errors, Weinger said, are often due toviolations of human factors design principles, asshown below.
Some Human Factors Principles
⢠Meet usersâ needs
⢠Be visible
⢠Design for error
⢠Keep it simple
⢠Be consistent
⢠Communicate
⢠Fail gracefully
⢠Minimize workload
⢠Automate wisely
⢠Focus on core tasks
In practice, these principles should be used to satisfycliniciansâ needs and requirements with infusiondevicesâdoing actual work, accommodating thecomplexity of care, and accounting for the context of use, such as interactions with other technology.Weinger offered these recommendations tomanufacturers and other stakeholders for improvinginfusion pump design:
⢠Develop a meticulous understanding of theuser profile and user workflow. âWe spend toomuch time adapting our workflow to the pump,âhe said.
⢠Make the pump configurable. Add options tothe software design to meet local cultural andworkflow needs. âUsers are different,â he said. âWe need to be able to tweak the options.â
⢠Use bar code technology. Bar code standards are available todayâand bar coding connected tothe pump and the rest of the system provides asafety check and a backup when wireless systemsfail.
⢠Use design standards. Manufacturers canât make a different pump for every hospital, but they could simplify and standardize the design.âPredictability is more important thanconfigurability,â he said. âNurses work in morethan one unit. The pumps might look the same on the outside, but they donât work the same onthe inside. Thatâs a big problem.â
Still, a one-size-fits-all pump would invariably be acompromise suitable to no one, he said. Usabilitytasks and criteria, reporting of usability test results,and core and high-risk user interface features areaspects of infusion devices for which standardizationmakes sense.
The ANSI/AAMI standard, Human factorsengineeringâDesign of medical devices (ANSI/AAMIHE75:2009), is an FDA-approved tool with humanfactors engineering design guidance, case studies,and checklists. Another FDA-approved tool,Medical devicesâApplication of usability engineering
to medical devices (ANSI/AAMI/IEC 62366:2007), is intended primarily for manufacturers and technical committees preparing medical device standards.
Peter DoyleThe Johns Hopkins Hospital
Q. What prompted you to undertake your userinterface study of large-volume pumps?A. âDr. Peter Pronovost of The Johns HopkinsUniversity envisions a Public-Private Partnership to Promote Patient Safety (P5S). This partnership ismodeled after an aviation safety initiative thatincluded all system stakeholders in the pursuit ofsafety. Because the Johns Hopkins University Applied Physics Lab (JHU APL) has expertise insystems engineering, we invited them to participatewith us in a study as a means to grow and developthe P5S methodology. Infusion pumps wereidentified as a good place to start because werecognized the risks of use and they constitute arelatively discrete device with many external âarmsand legs,â if you will, deserving attention for the sakeof safety. Those arms and legs include programminginterface design, drug libraries, labels, user skill andtraining, and environmental considerations. Werecognized pumps are a good subject for partneringwith stakeholders in the pursuit of safety. It was notlong after that we learned of the FDAâs initiative withAAMIâs collaborative efforts.
âSystems engineering is a professional discipline that addresses the increasing difficulty of designingall aspects of complex new products. The goal is to produce products that meet all performancerequirementsâto include safety, operability,reliability, and supportabilityâover the life cycle of the system. A very broad, interdisciplinary teamuses a variety of engineering processes andmethodologies to control and coordinate diversedevelopment tasks, manage risks, and ensureperformance.â
Q. Your study identified a number of design and user issues. Are you collaborating withmanufacturers or any other stakeholders toaddress the issues you raised? How?A. âFor more than two years a diverse team at The Johns Hopkins Hospital has met monthly with one pump manufacturer to share clinicalexperiences and obtain and review safety data.
We also conducted a hazard analysis (a failure modesand effects analysis, or FMEA) and communicated the results to the manufacturer. Realizing the issuesidentified were not unique to one manufacturerprompted us to look more broadly at the challengeto affect design industry-wide. Then, as the FDAinitiative developed, we rolled our efforts into theFDA/AAMI undertaking.â
Q. How widely used is the systems engineeringapproach in healthcare? Is it becoming moreimportant as medical devices and IT convergeand become more complex?A. âWhile I am not sure how widely it is used, I suspect that it is used in varying degrees bydifferent organizations. It is our opinion that the very detailed and all-encompassing systemsengineering processes practiced at JHU APL wouldadd value to the methods used by commercialhealthcare developers. We hope to progressivelydemonstrate the value of employing thesetechniques as we develop the P5S approach.
âSystems engineering is absolutely more importantgiven the increased electromechanical, IT, operationand maintenance complexity of devices, the supportsystems required, and the training and skill demandsof the users, who conduct their tasks in a fast-pacedenvironments where failures have severeconsequences.â
Q. You noted in your presentation that taskanalysis could be used to analyze additionaltasks, such as anesthesia use and maintenance,and other pump devices. Do you plan toundertake such studies? What additional issues do you anticipate you could discover?A. âWe have a plan to do a comprehensive userrequirements analysis and develop simulation toolsto evaluate many aspects of pump use with someprototype designsâweâre waiting to hear onfunding. Anesthesia use differs in that the pump is not monitored so remotely. Also the users havedifferent training, skill sets, and objectives.Maintainers can identify and prevent failure modes.Each user/maintainer may provide beneficial inputfor new design requirements.â
EXPERT PERSPECTIVE: Studying User Interfaces
26 | Infusing Patients Safely
CALL TO ACTION: A SYSTEMS APPROACH TO USERS AND USER INTERFACES
Priority Issues From the AAMI/FDA Infusion Device Summit | 27
Q. Could the systems engineering approachand usability studies that you employ beundertaken at smaller facilities, or those thatare not focused on research?A. âThe depth or breadth of studies would not denythe possibility of learning valuable information. I would encourage anyone to engage in identifyingfailure modes and improvements. Sometimes thebest findings come from unexpected sources.â
Q. If you were on a design team for infusionpumps, what would you ask designers andengineers to do to create safer pumps? Whatspecific experiences with pumps would youshare?A. âSpecifics are many, but I would encourage theiterative use of usability testing of many operationprotocols, from device concept on paperâyes, youcan test thatâto final and post-production models.â
Q. What are the education and trainingchallenges that you, or the nurses on your team,face in working with infusion devices? A. âA very significant challenge is the time availablefor training. Production pressures and staffinglimitations put the hurt on training opportunities. I believe that training technologies, such ascomputer-based multimedia training and embeddedtraining (pumps with built-in means to providetraining and feedback), are not used to their bestadvantage. These means can be very effective in thetime-stressed environment we face.â
28 | Infusing Patients Safely
CALL TO ACTION: A SYSTEMS APPROACH TO USERS AND USER INTERFACES
Matthew B. WeingerVanderbilt University School of Medicineand VA Tennessee Valley Healthcare System
Q. As a clinician, what prompted you to becomeactive in promoting greater use of humanfactors principles in medical device design?A. âI have been interested in humanâtechnologyinteractions in healthcare since college (30 years ago)when I was a dual electrical engineering and biologymajor at Stanford with an interest in the relativelynew field of biomedical engineering. When I was afirst-year medical student at UC San Diego, I workedwith John Brimm on a decision support system in thecardiothoracic ICU. I was struck by the sheer amountof technology around the fresh post-op heart andthe demands that technology made on the nurses.â
Q. In your presentation, you arguedpassionately for greater consideration ofcliniciansâ needs in the design of userinterfaces. Is the medical device industrybehind other industries you mentioned, such as the airline, energy, or consumer productsindustries, in using human factors designprinciples? If so, why? A. âYes, in general, healthcare is behind otherindustries in adopting state-of-the-art quality andsafety interventions. It is also behind the consumerproducts industries in its ability to be customer-centered and customer-drivenâmeaningpatient-driven. With regard to human factors, Iâd estimate that healthcare is still on average 20years behind other complex, high-risk, technology-oriented industries like aviation and energy.â
Q. Are there special barriers to studying humaninteractions with infusion devices? If so, howcould these barriers be addressed?A. âThe devices are actually quite complex as are thediverse environments in which they are used. Iâdspeculate that infusion pumps are the most widelyused, complex, computer-based technology inhealthcare today. They are much more ubiquitousthan equivalently complex devices like ventilators,have a wider range of uses than most devices, andare used by a wider range of users (includingpatients). Obviously, the use of an infusion pump athome by an elderly, far-sighted, tremulous arthriticdiabetic patient will be worlds apart from use of thesame pump by an anesthesiologist in an operatingroom. So it is perhaps not surprising that it ischallenging to design effective, safe, usable, efficient,and satisfying infusion pump user interfaces.
âHigh-fidelity simulations of infusion pump use aremore challenging and expensive to create. However,high-fidelity simulation has become much moreavailable in the last 10 years.â
Q. You are an editor of a new book, Handbookof Human Factors in Medical Device Design.What topics in the book are pertinent toinfusion device design and mitigation of useerrors? A. âMuch of the book is relevant to infusion pumpuser interface design. The book is intended tocomplement the new ANSI/AAMI HE75 nationalstandard on the design of medical device userinterfaces. For example, chapters of direct relevanceinclude Controls, Displays, and Environment of Use.â
EXPERT PERSPECTIVE: User Interfaces and Use Environments
Priority Issues From the AAMI/FDA Infusion Device Summit | 29
CALL TO ACTION: MANAGE THE COMPLEXIT Y OF MULTIPLE INFUSIONS
âMultiple IV infusion errors are both prevalent andharmful.â
âAndrea Cassano-PichĂŠ, human factors engineer at the University Health Network Centre for
Global eHealth Innovation, Toronto
CLARION THEME 4Improve management of multiple infusions.
PRIORIT Y ISSUE:
11. There is difficulty in infusion line manage-mentâincluding containers, manifolds, catheters,and transportâreflecting the complexity ofmultiple infusions, including secondaries,disposables, etc.
30 | Infusing Patients Safely
CALL TO ACTION: MANAGE THE COMPLEXIT Y OF MULTIPLE INFUSIONS
Understanding the Issue: Multiple Infusions
All of the human factors and user interface issuesthat summit presenters and participants identifiedwith infusion pumps are compounded with multipleinfusions. This priority issue makes such an impacton patient safetyâand raises an even different set ofconcernsâthat both summit participants and AAMIexperts believe it needs its own call to action.
Working with multiple pumps simultaneously, eachof which could be very different, escalates thepotential for errors. The ancillary medical equipmentassociated with many different devices is difficult tomanage as well. The expert perspectives belowelaborate on the challenges.
Matthew B. WeingerVanderbilt School of Medicine and VA Tennessee Valley Healthcare System
Q. What are the clinical challenges of multipleinfusions?A. âManaging multiple lines and pumps addsappreciably to the complexity for clinical users. But it also makes it challenging for manufacturers toassure patient safety because they currently have nocontrol over what happens after a drug leaves theirpump. For example, a pump currently cannot tell
where the pumped drug is going or whether anyincompatible drugs or fluids are being intermixed.
âTo address these important safety issues, we aregoing to need some innovative technologies. I believe that we will need technologies at the IV siteto measure what and how much is actually goinginto the patientâs vein (as opposed to infiltrating). We will need better ways to manage all of theplumbing, with clear correct linkage to medication,pump, and patient. I like to say that what weâll have in the future is way to âteleportâ the drugs directlyinto the patient from the pharmacy Ă la Star Trek.â
EXPERT PERSPECTIVE: Managing Multiple Infusions
Priority Issues From the AAMI/FDA Infusion Device Summit | 31
Andrea Cassano-PichĂŠHuman Factors Engineer, University Health Network Centre for Global eHealthInnovation, Toronto
Q. Can you describe the study of multipleinfusions that you are conducting? Why did youundertake it? A. âWe are conducting a study on the nature andfrequency of errors related to the administration of multiple IV infusions. Our objective is to betterunderstand the types of risks contributing to errorsand then identify and validate potential risk-mitigation strategies.
âThe study came about in response to a heightenedlevel of awareness of multiple IV infusion risks afterperforming a human factors study on smart IVpumps. When comparing user performance ontraditional large-volume IV pumps versus smart IVpumps, it became apparent that errors associatedwith administering multiple IV infusions to a singlepatient were not mitigated by current safetytechnology, and had potentially severe outcomes. A review of two independent incident databasesconfirmed that multiple IV infusion errors are bothprevalent and harmful, and a review of the literatureshowed that this problem has not beensystematically studied and that mitigation strategiescurrently in use have not been validated.â
Q. When will you have results to share?A. âWe expect to have results to share in early 2012.â
Q. What are the confounding issues in hospitalsand other healthcare settings for managingmultiple infusion lines?A. âThe most prevalent issue is the overallcomplexity of the medication administration system.A large number of componentsâincluding an order,a prepared and labeled drug container, tubing, apump, connectors, and patient accessâall need to bearranged in exactly the right way for an infusion tobe administered safely. There are a lot of decisionsand actions that have to be made correctly by severaldifferent clinicians. As the number of IV infusions on
a patient increases, so does the complexity. Morecomplexity leads to more errors.â
Q. How could human factors studies and designprinciples be used to help clinicians avertincidents with infusion devices?A. âHuman factors studies aim to understand thelevel of complexity in the system and identifyopportunities to reduce it to a level that isappropriate for humans to manage safely. That is,they identify the required complexity needed by the users to support their role and the extraneouscomplexity that can be reduced using a variety ofapproaches, such as standardization, reducing thetotal number of steps in a task, automation, and theapplication of usability principles to the design ofuser-technology interfaces.
Q. Youâve contributed to the development andpresentation of Human Factors 101, a courseoffered monthly to healthcare workers at theUniversity Health Network. What kinds ofhealthcare workers take this course? What dothey learn? Should all healthcare workers takea course like this?A. âHuman Factors 101 is a course aimed at cliniciansand healthcare administrators to get them thinkingfrom a broader systems perspective about systemsafety and safety culture. We use case studies tohighlight foundational human factors principles, such as limitations on attention and memory andcognitive biases. We also discuss more controversialtopics such as culpability and team dynamics. Weâvehad just about every type of hospital employee aswell as representatives from industry and other typesof healthcare organizations participate in our course.
âIt is surprising that human factors concepts thathave been talked about by quality and safety practi-tioners in healthcare for over a decade are still newfor most attendees. This is probably because humanfactors education is not an integral part of medicaltraining. I think all clinicians need to be exposed tohuman factors education both in their academictraining and throughout their professionaldevelopment.â
EXPERT PERSPECTIVE: A Human Factors Approach to Studying Multiple Infusions
32 | Infusing Patients Safely
CALL TO ACTION: RECONCILE ALARMS, DIFFERENT USE ENVIRONMENTS
âWhat could go wrong? Let me count the ways.â
âChristine Kessler, R.N., M.N., nurse practitioner at Walter Reed Army Medical Centerâs
Diabetes Institute
CLARION THEME 5Reconcile challenges and differences in the use environments of infusion devices.
PRIORIT Y ISSUES:
12. Alarm management is not effective.
a. There are high numbers of false alarms, whichalso can lead to true alarms being ignored(e.g., air).
b. Alarms are difficult to prioritize.
c. It is unclear how to resolve alarm issues.
13. Injuries are caused by a lack of differentiationbetween the use of infusion devices in hospitalsand in other environments (e.g., home use).Products designed for the hospital environmentare being used in home environments (and vice versa). There are design and user issues anddifferences among home, hospital, and otherenvironments.
Priority Issues From the AAMI/FDA Infusion Device Summit | 33
Understanding the Issue: Use Environments
The healthcare settings in which infusion devices areused are enormously diverse, ranging from hospitalsto alternative care and mobile facilities to homes. In each of these environments are more facets ofdiversityâin terms of such variables as size, location,infrastructure, services, staffing, professional capacity,and patient profiles.
Infusion devices do not always mesh well with theirenvironments. This plays out in ineffective alarmmanagement, inappropriate uses of infusion devicesin some environments, and design and use issuesthat differ, depending on the use environment,summit presenters and participants said.
Ineffective alarm management encompasses anumber of serious problems. False alarms abound.Alarms are too soft or too loud. Alarms from differentdevices cannot be distinguished. Alarms do notalways indicate the severity of the problem, making itdifficult to prioritize alarm responses. Not all infusioncomplications prompt an alarm. Alarms can beturned down or turned off inappropriately. Alarmsfail. And these uncertainties can lead to alarmfatigueâand inadvertent failure to notice andrespond to true alarms.
For this priority issue, summit presenters andparticipants suggested these starting points foraction to reconcile challenges and differences in useenvironments:
⢠Standardize alarms.
⢠Clarify roles and responsibilities of manufacturersand users.
⢠Build protocols for alarm responses into alarms.
Several summit presenters addressed issues withinsulin pumps, which are widely used in home andmobile environments and exemplify challenges anddifferences in different use environments.
FDA is investigating infusion device safety throughthe lens of software, a key device component,according to summit presenter Paul L. Jones, senior systems/software engineer at FDAâs Center for Devices and Radiological Health. FDA ispiggybacking on its Generic Infusion Pump Project,
which began with PCA pumps, to develop model-based software that incorporates safety principles,properties, and methods for insulin pump use.
âWe are in the process of putting together a safetyanalysis, guidelines, and user computer science toolsto reason about the safety of devices,â Jones said.Manufacturers will be able to use or adapt thesesafety models and reference specifications to verifythe safety of different classes of infusion pumps.Academic and clinical researchers are collaboratingwith FDA on this project.
âOur interest is to gain knowledge from technologyto regulate more effectively and efficiently,â Jonessaid. âWe hope to have a regulatory and publichealth impact,â including an infusion pump guidancedocument, an insulin pump standard, and insulinpump safety.
Improving insulin pump safety could, indeed, have a significant public health impact. âThirty-eightpercent of all medication errors are made withinsulin,â said summit presenter Donna Jornsay, B.S.N.,R.N., a diabetes educator and clinical specialist atLong Island Jewish Medical Center in New York.âThese numbers are particularly relevant to theincreasing number of children with diabetes.â
Clinicians who treat diabetic patients with insulinpump therapy make an impact on public health onepatient at a time. The challenges and differenceswith insulin pump use environments vary with everypatient. âWhat could go wrong? Let me count theways,â said summit presenter Christine Kessler, R.N.,M.N., a nurse practitioner at Walter Reed ArmyMedical Centerâs Diabetes Institute. She catalogued a litany of errors associated with pumps, clinicians,and patients, including:
⢠Selecting the wrong pumps for particular patients,based on such patient factors as health, literacy,vision, ability, mobility, activity, and responsibilitylevels, as well as the availability of caregiversupport
⢠Errors in setting up basal and bolus doses andrates
⢠Errors in calculating the duration of insulin actionand âbolus on board,â which differ with differentpumps
34 | Infusing Patients Safely
CALL TO ACTION: RECONCILE ALARMS, DIFFERENT USE ENVIRONMENTS
Matthew B. WeingerVanderbilt School of Medicine and VA Tennessee Valley Healthcare System
Q. How can alarm management be improved?A. âAlarm management will only improve when wehave fully integrated alarm systems that are trulyâintelligent.â This will require a uniform plug-and-playstandard for all medical devices and centralizedalarm management so the information about patientstatus can be shared and utilized.â
EXPERT PERSPECTIVE: Alarm Management
⢠Errors in reminders and alerts
⢠Errors in guestimating and manually inputtingcarbohydrate intake
Smart pumps are making it easier to manage some of these issues. However, Kessler said, âThe pump isonly as good as the patient education. True riskmanagement is follow-up, follow-up, follow-up.â
Priority Issues From the AAMI/FDA Infusion Device Summit | 35
Alarm Fatigue: âAn Issue Across the Countryâ
WHEN AN 89-YEAR-OLD MAN died in his bed at Massachusetts General Hospital, the alarms thatshould have gone offâbut didnâtâset off a differentkind of alarm throughout the hospital, the city ofBoston, the state, and regulatory agencies.
An investigation revealed that the volume on thepatientâs bedside monitor alarm and the alarmdefault setting for lethal arrhythmia had been turnedoff, according to summit presenter Lela Holden, Ph.D.,R.N., and patient safety officer at MassachusettsGeneral. Plus, 10 nurses on duty that day had notdetected a lower-level, two-beep audio alarm at thecentral nursesâ station or on scrolling tickertapemessages on three hallway signs. This alarm signaledthat the patientâs oxygen and heart rate werecritically low.
How was this possible? No one knew when or whythe bedside alarm had been disabled. Very fewpeople knew it was even possible to turn off thealarm default setting for lethal arrhythmia. Furtherinvestigation revealed staff confusion regardingmonitoring and alarm system settings in a complexenvironmentâone with two central systems,multiple models of devices and software versions, a large number of patients on monitors, and a lack of standards for monitoring. Alarm fatigue anddesensitization were uncovered as well.
The hospital drew attention from all corners for thisfatality, including coverage in The Boston Globe, andscrutiny from state and federal regulators and TheJoint Commission. Their collective questioningactually helped the hospital staff learn from theevent and come together in cross-functional teamsto make improvement, Holden said. Now, clinicalstaff and monitor manufacturers do a double-checkwhen devices are rolled out. There are standarddefaults on monitors and a comprehensive processto review defaults. And the biomedical engineeringdepartment reports greater trust from clinicians forits independent validations of equipment, Holdensaid.
âEveryone has a role to playâclinicians, clinicalengineers, vendors, journalists, and regulators,â shesaid.
The tragedy underscored the fact that infusion pumpproblems can occur at even the best and mostinnovative hospitals. U.S. News & World Report rankedMassachusetts General, which pioneered the use ofdrug libraries in infusion pumps in 1996, fifth on its2010 list of best hospitals.
LEAD-USER PROFILE: Massachusetts General Hospital
36 | Infusing Patients Safely
Robert Bernstein, M.D. Southwest Clinical Research Center andRegional Endocrinology Associates
Suellen Minturn, R.N.Regional Endocrinology Associates
DEPENDING ON PATIENTS and their useenvironments, portable insulin pumps can be averitable game changer with a dramatically positiveor negative impact on life. âVirtually all of thepatients whom we have helped get started on pump therapy think itâs a big improvement for them,âsaid summit presenter Robert Bernstein, principalinvestigator at the Southwest Clinical ResearchCenter and an endocrinologist at RegionalEndocrinology Associates in Santa Fe, New Mexico.âWhat theyâve told us is that they are lesshypoglycemic and have less variability in their bloodsugars with the pump. A lot say they just feel better.â
At the same time, he said, âThe great majority ofproblems we have encountered are due to patienterror.â At the summit, he offered several case studiesthat indicate the range of challengesâpatientsdropping pumps in water, exposing them to heat in hot tubs or saunas, reusing disposable equipment,and failing to securely reconnect tubing afterdisconnecting it. Such incidents can damage pumpsor degrade insulin.
âTo use an insulin pump requires a lot of attention,âBernstein said. When patients run into trouble, theytell him, ââI thought the pump would take care ofeverything.ââ
But insulin pumps, like other infusion devices, arenâtyet tough enough to stand up to patient neglect ormisuseâor smart enough to rectify all usershortcomings or use errors.
Thus, clinicians share the responsibility with patientsfor successful insulin pump use, beginning withselecting patients for pump therapy, Bernstein said.Not all patients are candidates for insulin pumps,including patients who are not committed,interested, or capable of managing therapy andpatients with literacy, vision, or hearing problems.
Clinicians who provide training and ongoing supportto patients on insulin pump therapy have a similarâbut magnifiedâchallenge as hospital clinicians:working with a variety of pumps that do not havestandard features. âThey all run in a slightly differentway,â said Suellen Minturn, a nurse and certifieddiabetic educator at Regional EndocrinologyAssociates. If a patient calls her for help when she is not in her office, she has to recall the layout andfunctionality of a particular pump from memory. âI can picture the one or two preferred pumps. Morethan that, I donât know that I could.â
From her experiences with patients, Minturn offeredthese suggestions for pump manufacturers:
⢠Offer a magnifying lens that fits over the screen,larger print, and brighter screens for patients with vision problems
⢠Make icons more intuitiveâand offer a printoption for patients who find it difficult tomemorize ârandomâ icons
⢠Translate the text into more languages
⢠Design a safety feature that alerts patients whenthe cannulaâa small, soft tube that goes underthe skin and connects to the pumpâis partiallyoccluded (not just completely occluded)
⢠Include sensors that can tell when there isresistance or hardened tissue at the infusion site
⢠Test alarm tones on a variety of age groups
⢠Offer a strong vibrating alarm so patients who are hard of hearing or asleep will notice it
⢠Build educational features into the pump, sopatients can learn how their nutritional andexercise habits affect their blood sugars
EXPERT PERSPECTIVE: Insulin Pump Use Environmentsâand Patients
CALL TO ACTION: RECONCILE ALARMS, DIFFERENT USE ENVIRONMENTS
Priority Issues From the AAMI/FDA Infusion Device Summit | 37
NO ONE INDIVIDUAL, working group, ororganization can address the 13 summit prioritieswell in isolation. Prior to the summit, there was nocoordinated, collaborative, multiple-stakeholdermechanism for addressing the challenges withinfusion device systems. Many of the complex issues do not lend themselves to standards; FDA regulations cannot and should not addresseverything; and no single stakeholder has enoughcontext or perspective to see the issues from theperspective of other stakeholders.
At the conclusion of the summit, Pat Baird, co-chair ofthe AAMI Infusion Device Standards Committee, andAAMI President Mary Logan announced the conceptof a medical device safety council to serve as thebody that coordinates implementation of the actionplans that result from the summit. The vision for thesafety councilâs first initiative, infusion device systems,is to convene, cajole, coordinate, collaborate, andcelebrate specific clarion theme projects, with thesuccess measure being safer infusion.
Nearly 100 summit participants signed up toparticipate on specific priority issues, and they will be invited to populate the working groups for eachof the clarion theme projects.
The Medical Device SafetyCouncil/Infusion
The Medical Device Safety Council/Infusion is beingestablished within The AAMI Foundation, which is a501(c)(3) charitable and educational organization,thus supporting the overall goal of improving patientoutcomes by making infusions safer.
While the initial, purposeful focus of the safetycouncil is infusion device systems, there will be a next medical device challenge, flowing from the next summit(s) and other sources. For this reason, the structure is intended to be fluid, nimbleâandready to jump into the next issue if and when thetime is right, as shown in Figure 8.
The AAMI Foundation expresses its gratitude to thefollowing organizing committee members for theirtime, expertise, and passion for launching this work:
⢠Pat Baird, Baxter International, Inc.
⢠Bona Benjamin, American Society of Health-System Pharmacists (ASHP)
⢠Andrea Cassano-PichÊ, University Health Network,Centre for Global eHealth Innovation
⢠Glenn Scales, Duke University Health System
⢠Dennis Schneider, Baxa Corporation
⢠Nat Sims, Massachusetts General Hospital
⢠Erin Sparnon, ECRI
⢠Tony Watson, FDA
Where We Go From Here
âDo or do not. There is no try.â
âYoda
38 | Infusing Patients Safely
Funding
The AAMI Board of Directors in November 2010approved the transfer of the approximately $15,000in income left after paying all direct expenses fromthe summit to help get the safety council off theground. AAMI absorbed all indirect expenses fromthe summit. As advised to all summit participants,AAMI will not be providing funding for the safetycouncil going forward. The safety council wasformed by the will of the community and thus must besupported financially by the will of the community.Contributions by individuals, hospitals, and otherhealthcare organizations, patient safety groups,consultants, industry, and others will be essential ifthe safety council is to be successful.
To make a donation, mail your check or money orderto:
AAMI FoundationMedical Device Safety Council/Infusion4301 N. Fairfax Drive, Suite 301Arlington, VA 22201-1633
For funding questions, contact Lauren Healy at AAMI:703-253-8290.
Next Steps
The organizing committee met on Dec. 8, 2010. Their primary work product is working groupassignments based on the five clarion themes at the summit. The assignment âgridâ will be posted at www.aami.org/foundation by the first week ofJanuary 2011. Updates will continue to be posted
there and at www.aami.org/meetings/infusionsummit/
Let the projects beginâletâs do!
Source: Association for the Advancement of Medical Instrumentation.
Figure 8. Organizational Chart for the Medical Device Safety Council/Infusion
Priority Issues From the AAMI/FDA Infusion Device Summit | 39
THE AAMI/FDA INFUSION DEVICE SUMMITissued a call to action for improving infusion devicesafety: five clarion themes and 13 priority issues thatresonate with urgency.
AAMI, FDA, and summit participants from across the spectrum of the healthcare community issued a renewed pledge to address the issues withinnovative solutions. Many of these solutions couldbe informed by systems approaches, such as:
⢠Thinking systemically about incident reporting,root cause analysis, and continuous improvement
⢠Designing with a systems engineering perspective
⢠Transitioning from smart pumps to smart infusionsystems
⢠Connecting medical technology and informationsystems
⢠Systematizing drug libraries and protocols
⢠Responding to the diverse use environments anduser needs within healthcare systems
Lela Holden, patient safety officer at MassachusettsGeneral Hospital, opened the summit with a keynote about healthcare being a team sport. Sheemphasized that putting patient safety first requiresa willingness by all of us to be completely openabout glitches, near misses, and adverse incidents.Some summit participants expressed curiosity abouthow these two messages about teamwork andopenness relate to the task before us.
By the end of the event, it was clear that Holdenâsmessage had hit home. Not one of the five clarionthemes has a single âownerâ among industry,
hospitals, clinicians, clinical engineers, biomedicalequipment technicians, independent experts, ITprofessionals, or FDA. These are systems issues,requiring the whole healthcare community tocontinue to work together as a team. For example,Clarion Theme 1 will not be solved with a newâMAUDE.â Users and user facilities must makeaccurate, clear, and detailed reporting of glitches,near misses, and adverse incidents to manufacturersa priority; manufacturers must make root causeanalyses a priority; FDA must listen to the needs ofhospitals, industry, and outside subject experts aboutMAUDE. Listening well must continue to be anongoing theme by everyone in the healthcarecommunity if real progress is to be made.
Summit presenters and participants shared insightsand examples of progress on infusion safety, whichcould inform more widespread and comprehensiveimprovements to infusion systems. Other expertsgenerously contributed their perspectivesâandrecommendations for moving forward.
The AAMI Foundationâs new Medical Device SafetyCouncil has been charged with sustaining themomentum from the summit with an action plan foraddressing the priority issues. Like the summit itself,the action agenda will require multidisciplinary,collaborative efforts. No single group can do it alone.With collective action, smarter infusion systems andimproved patient safety are within reach.
Already, nearly 100 people from more than 60organizations have volunteered to help. To volunteerfor this work and to monitor summit follow-upactivity, visit AAMIâs Web site dedicated to infusiondevice safety.
www.aami.org/infusionsummit/index.html
ConclusionâThe AAMI/FDA Infusion Device Summit was a productivefirst step in fostering an open environment wherecollaboration and open discussion is encouraged.â
âAnthony Watson, director of the Division of Anesthesiology, GeneralHospital, Infection Control and Dental Devices in FDAâs Office of
Device Evaluation at the Center for Devices and Radiological Health
40 | Infusing Patients Safely
AAMI Infusion Device Summit
www.aami.org/infusionsummit
ASHP IV Safety Summit
www.ashp.org/iv-summit
FDA Infusion Pump Improvement Initiative
www.fda.gov/InfusionPumps
AAMI Resources
Visit www.aami.org for these and other infusion-related publications.
ANSI/AAMI/IEC 62366:2007, Medical devices:Application of usability engineering to medical devices
ANSI/AAMI HE75:2009, Human factors engineeringâDesign of medical devices
Brady, Jody L. âFirst, Do No Harm: Making InfusionPumps Safer.â (September/October 2010).Biomedical Instrumentation & Technology. Associationfor the Advancement of Medical Instrumentation.
Home Healthcare Horizons, a supplement toBiomedical Instrumentation & Technology. (2010).Association for the Advancement of MedicalInstrumentation.
Other Resources
Evans, R.S., Carlson, R., Johnson, K.V., Palmer, B.K., &Lloyd, J.F. (2010). âEnhanced Notification of InfusionPump Programming Errors.â Studies in HealthTechnology and Informatics, 160: 734â8.
Getting Started Kit: Safe Administration of High-Risk IV Medications
www.nichq.org/pdf/AppendixDSanDiegoPatientSafetyConsortium.pdf
HPM2:2 Medical Instrumentation: Infusion PumpConnectivity to Clinical Systems. (February 2010).Partners Healthcare.
Indianapolis Coalition for Patient Safety
www.indypatientsafety.org/focus_areas.aspx
âIntermountain Healthcare, Baxter Integrate SIGMASpectrum Infusion System with IntermountainâsElectronic Medical Record System.â (Dec. 1, 2010).Business Wire.
www.businesswire.com/smp/intermountain-and-baxter-integrate-sigma-spectrum-infusion-system-with-electronic-medical-record-system/
Weinger, Matthew B.; Wiklund, Michael R.; & Gardner-Bonneau, Daryle J. (Eds.) (2010). Handbook of HumanFactors in Medical Device Design. CRC Press.
Resources
Priority Issues From the AAMI/FDA Infusion Device Summit | 41
Supporting OrganizationsAmerican College of Clinical Engineering (ACCE)Association of periOperative Registered Nurses (AORN)American Society of Health-System Pharmacists (ASHP)Association of Surgical TechnologistsCenter for Integration of Medicine & Innovative
Technology (CIMIT)Diabetes Technology SocietyECRI InstituteIntegrating the Healthcare Enterprise (IHE)IHE USAInstitute for Safe Medication Practices (ISMP)National Patient Safety Foundation (NPSF)Parenteral Drug Association
Summit Presenters and Expert CommentatorsPat BairdBaxter International, [email protected]
Bona BenjaminAmerican Society of Health-System Pharmacists [email protected]
Robert BernsteinSouthwest Clinical Research Center and
Regional Endocrinology Associates [email protected]
Rick CarlsonIntermountain Medical [email protected]
Andrea Cassano-PichĂŠUniversity Health Network
Centre for Global eHealth [email protected]
Peter DoyleThe Johns Hopkins [email protected]
Haley E. Goodwin The Johns Hopkins Hospital [email protected]
Carol Herman FDA Center for Devices and Radiological [email protected]
Lela HoldenMassachusetts General [email protected]
Acknowledgments
Paul L. JonesFDA Center for Devices and Radiological [email protected]
Donna JornsayLong Island Jewish Medical [email protected]
Christine KesslerWalter Reed Army Medical [email protected]
Dorian Liepmann University of California, Berkeleyliepmann@ berkeley.edu
Alan LipschultzChristiana Care Health [email protected]
William MaiselFDA Center for Devices and Radiological [email protected]
Suellen MinturnRegional Endocrinology Associates [email protected]
Bryanne PatailVA National Center for Patient Safety [email protected]
Amanda PruschLancaster General [email protected]
Glenn ScalesDuke University Medical [email protected]
Erin SparnonECRI [email protected]
Tina SuessLancaster General [email protected]
Anthony WatsonFDA Center for Devices and Radiological [email protected]
Matthew WeingerVanderbilt University School of [email protected]
Writing By Martha Vockley, Vockleyâ˘Lang
42 | Infusing Patients Safely
AcelRxwww.acelrx.comCONTACT Nigel Ray
Animaswww.animas.comCONTACT Ramakrishna Venugopalan
Baxawww.baxa.comCONTACT Marian Robinson
Baxterwww.baxter.com/healthcare_professionals/products/sigma_spectrum.htmlCONTACT Eric Shawger
B. Braun Medical Inc.www.bbraunusa.comCONTACT Susan Denby
[email protected] 610-997-4856
CareFusionwww.carefusion.comCONTACT Suzanne Hatcher
FluidNetwww.fluidnet.netCONTACT Brent Nibarger
Fresenius www.fresenius-kabi.com CONTACT Brent Eck
[email protected] 847-969-2700
Thank You, Summit SponsorsThis publication was made possible by the financial support provided by these companies.
Intertech Engineering Assocwww.inea.comCONTACT David A. Vogel
Kimberly-Clark Healthcarewww.kchealthcare.com/CONTACT David W. Johnson
Medical Specialties Distributorswww.msdonline.com CONTACT Jim Milostan
[email protected] ext. 6642
Moogwww.moog.com/medical/ CONTACT Shaunalee Wall
[email protected] ext. 297
Nestlewww.nestle-nutrition.comCONTACT Brian Johnson
[email protected] 952-848-6191
Smiths Medicalwww.smiths-medical.com CONTACT Martha Sewall
Zynxwww.primeapump.comCONTACT Pam Quinlan
The sponsors listed on these pages helped to pay for theproduction, printing and mailing costs for this publication. FDA was not involved in AAMIâs securing of the sponsors. The sponsorships do not constitute any type of endorsementof the sponsors by AAMI or FDA. AAMI expresses its gratitudeto these companies for sponsoring this publication.
Association for the Advancement of Medical Instrumentation
4301 N. Fairfax Drive, Suite 301Arlington, VA 22203-1633
703-525-4890
www.aami.org