Balancing the Tension BetweenHyperoxia Prevention and AlarmFatigue in the NICUAnastasia K. Ketko, MDa, Craig M. Martinb, Michelle A. Nemshak, RNc, Matthew Niedner, MDd, Rebecca J. Vartanian, MDd

abstractBACKGROUND:After the implementation of narrowed oxygen saturation alarms, alarmfrequency increased in the C.S. Mott Children’s Hospital NICU which could havea negative impact on patient safety. The Joint Commission on the Accreditation ofHealthcare Organizations issued a Sentinel Event Alert for hospitals in 2013 toimprove alarm safety, resulting in a 2014 National Patient Safety Goal requiringinstitutional policies and procedures to be in place to manage alarms.

METHODS: A multidisciplinary improvement team developed an alarmmanagement bundle applying strategies to decrease alarm frequency, whichincluded evaluating existing strategies and developing patient care–based andsystems-based interventions. The total number of delivered and detectedsaturation alarms and high saturation alarms and the total time spent withina targeted saturation range were quantitatively tracked. Nursing morale wasassessed qualitatively.

RESULTS: SpO2 alarms per monitored patient-day increased from 78 to 105 afterthe narrowing of alarm limits. Modification of the high saturation alarmalgorithm substantially decreased the delivery and escalation of high pulseoxygen saturation (SpO2) alarms. During a pilot period, using histogramtechnology to individually customize alarm limits resulted in increased timespent within the targeted saturation range and fewer alarms per day.Qualitatively, nurses reported improved satisfaction when not assigned .1infant with frequent alarms, as identified by an alarm frequency tool.

CONCLUSIONS: Alarm fatigue may detrimentally affect patient care and safety.Alarm management strategies should coincide with oxygen managementwithin a NICU, especially in single-patient-bed units.

The ECRI Institute defines alarmfatigue as a condition of sensoryoverload for staff members who areexposed to an excessive number ofalarms.1 Excessive false alarms canlead to nursing desensitization2 andburnout,3 resulting in complacencytoward alarm response,4 which canultimately result in patient harm. Theimportance of alarm safety andmanagement was publicized by a 2013Sentinel Event Alert from the JointCommission after 80 alarm-relateddeaths were directly reported in the

Joint Commission database between

2009 and 20125 and the Food and

Drug Administration’s Manufacturer

and User Facility Device Experience

database reported 566 alarm-related

deaths between 2005 and 2008.6 The

Sentinel Event Alert called for hospitals

to improve medical device alarm

safety,5 resulting in the 2014 National

Patient Safety Goal (NPSG), which

states that institutional policies and

procedures for alarm safety are to be in

place by 20161 (Fig 1).

aMinnesota Neonatal Physicians P.A., Children’s Hospitalsand Clinics of Minnesota, Minneapolis, Minnesota; bMedicalCenter Information Technology, Departments of cNursing,and dPediatrics and Communicable Diseases, University ofMichigan Health Systems, Ann Arbor, Michigan

Dr Ketko was involved in the quality improvementwork as a neonatal fellow; Ms Nemshak andDr Vartanian were involved in the qualityimprovement work throughout development andimplementation; Mr Martin carried out the alarmdata collection and analysis and developed andimplemented strategies; Dr Niedner providedexpertise in statistical process control and qualityimprovement science; Dr Ketko drafted the initialmanuscript; Ms Nemshak and Dr Vartanian criticallyreviewed the manuscript; Mr Martin and Dr Niednerreviewed and revised the manuscript; and allauthors approved the final manuscript assubmitted.

www.pediatrics.org/cgi/doi/10.1542/peds.2014-1550

DOI: 10.1542/peds.2014-1550

Accepted for publication Feb 26, 2015

Address correspondence to Rebecca Vartanian, MD,1540 E. Medical Center Dr, Ann Arbor, MI 48109.E-mail: rebeccav@med.umich.edu

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online,1098-4275).

Copyright © 2015 by the American Academy ofPediatrics

FINANCIAL DISCLOSURE: The authors have indicatedthey have no financial relationships relevant to thisarticle to disclose.

FUNDING: No funding was secured for this study.

POTENTIAL CONFLICT OF INTEREST: The authors haveindicated they have no potential conflicts of interestto disclose.

PEDIATRICS Volume 136, number 2, August 2015 QUALITY REPORT by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

Alarm frequency became a significantchallenge in the C.S. Mott Children’sHospital NICU at the University ofMichigan during implementation ofnew oxygen management guidelinesthat focused on oxygen targeting toreduce severe retinopathy ofprematurity (ROP).7–9 The guidelinesincluded narrowing of oxygensaturation alarms measured by pulseoximetry (SpO2) to maximize timespent within a target range. BedsideSpO2 alarms and signals from oursecondary alert notification system(SANS) escalated during initialevaluation. This increase in bothprimary and secondary alarmsdetrimentally affected nursing moraleand created concern for patientsafety, impeding further work on theoriginal quality improvement (QI)

effort to reduce ROP through oxygentargeting.

The goal of the QI effort described herewas to identify and implement alarmmanagement strategies that wouldreduce the burden of tightened alarmsto levels commensurate withpreintervention data. It was anticipatedthat attention to and improvement inalarm frequency would improvepatient care by returning focus ofpatient care to oxygen targeting ratherthan alarm management.

METHODS

Setting

The level IV NICU consists of46 single-patient rooms over19 580 square feet of patient care

space. Bedside cardiorespiratorymonitoring is achieved by use of GESolar 8000i monitors (GE Healthcare,Wauwatosa, Wisconsin) from time ofadmission through discharge. DefaultSpO2 alarm parameters were 80% to97%, and the default averaging timewas 16 seconds based on preexistingunit guidelines. A new institution-wide addition to workflow with thesingle-patient room layout was SANStechnology through ConnexallMiddleware System (GlobestarSystems, Toronto, Ontario, Canada),which escalates alarms detected atthe bedside monitor to a nurse’sphone through a modifiablealgorithm. The initial SANS escalationalgorithm was based onbenchmarking, opinion, andmanufacturer recommendations(Fig 2A) and transmitted highsaturation alarms, low saturationalarms, and low heart rate alarmsthrough the same algorithm to thenurse’s phone after 20 seconds. Analarm was considered detected if theSpO2 value breached the set high orlow parameter limits on the bedside

FIGURE 1Adapted from the Joint Commission, Sentinel Event Alert Issue 50.7

FIGURE 2A, Original SANS escalation algorithm for high and low SpO2 and heart rate. If the bedside alarm is not acknowledged by the primary nurse within40 seconds, it is escalated to a predetermined secondary nurse, and then to a backup team. B, Modified SANS escalation algorithm for high SpO2 only. Thesignal to the primary nurse was delayed by an additional 25 seconds to allow for autocorrection. The primary nurse now has 60 seconds to acknowledgethe bedside alarm before escalation to the secondary nurse.

2 KETKO et al by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

monitor. Institutional review boardapproval was not sought for this QIeffort, as our focus was on activities(alarms) rather than human subjects,thereby not meeting the definitionof human subjects researchrequiring institutional reviewboard approval (per Code ofFederal Regulations 45x46 and21x56).A multidisciplinary task forceconsisting of physicians, nurses, andrespiratory therapists (dubbed the“Sat Pack”) reviewed results fromrandomized controlled trials onoxygen targeting.7–9 Among manychanges, narrowed alarm parametersbased on gestational age wereadopted. SpO2 alarm frequencywas a predetermined balancingmeasure, with prospective datacollection of SpO2 alarms beginning

in May 2012 and retrospectivelyavailable from December 2011.Informal staff education throughe-mail began in February 2013, withformal training starting in April 2013.In May 2013, we conducted a small-scale pilot to study the impact andusability of the guidelines.Narrowed SpO2 alarm limits wereintroduced methodically to includemore labile infants, allowing staff togarner skill in oxygen managementstrategies. Concerns regardingpatient safety and nursing moralewere voiced, along with a parent-voiced “near miss,” which redirectedthe team to shift focus to meetthe urgent needs of alarmmanagement and delay theunitwide implementation of theoxygen management guidelines untilAugust 1, 2013.

Planning the Initiative

The Sat Pack expanded to includebiomedical engineering and aninformation technologist. Patient carepractices and systems/operationalpractices were determined to be keydrivers of alarm frequency (Fig 3).Processes to affect these key driverswere identified, and appropriatemeasures were selected and modifiedto align with those recommended bythe Joint Commission (Fig 1).

Patient Care Practices

1. Develop processes for safe alarmmanagement and response. TheSat Pack guidelines included algo-rithms about when and how torespond to SpO2 alarms. Real-timesupport for staff regarding alarmsafety was provided throughoutthe improvement process, and

FIGURE 3Driver diagram of alarm management strategies demonstrating primary and secondary drivers and process measures. No changes were made to signalaveraging time after initial assessment, so no process measure was identified.

PEDIATRICS Volume 136, number 2, August 2015 3 by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

safety reports were retrospectivelyreviewed for patient harm relatedto alarm strategies.

2. Tailor alarm settings for individualpatients. Premature neonates arenotorious for fluctuations in SpO2,and the ability to keep an infant’sSpO2 within target range isvariable10 (from 16% to 64% in1 study11).We theorized that theabsolute number of alarms doesnot provide the quantitative datato determine the time spent aboveor below a certain target range,and that perhaps a better tool fortargeting oxygen therapy existed.Histogram technology was identi-fied as a possible tool, as it quan-tifies the time spent within, above,and below the prespecified targetrange10,12 (Adams KK, GoldsteinM, Ninnis J, Hopper A, Deming D,unpublished observations) andprovides visual display of this data.In September 2013, a pulse oxim-eter with histogram technology(Masimo Radical-7, Masimo Corp.,Irvine, California) was piloted on7 infants who were identified tohave the most alarms. Histogramdata were monitored by the careteam at least every 4 to 8 hours.Individualized (usually broad-ened) saturation alarm limits weregranted to patients who rarely re-quired intervention for theiralarms or who showed improvedtime within their target saturationrange. The number of SpO2 alarmsand time spent within target rangewere collected at baseline andthroughout the pilot. After the pi-lot, equipment resources alloweda maximum of 10 patients to bemonitored with histogram tech-nology at any point.

Technology Interventions

1. Ensure clear guidelines for alarmsettings. Guidelines for alarm set-tings were standard practice in ourNICU, and the Sat Pack modifiedthem to incorporate gestationalage and degree of respiratorysupport. Electronic order sets and

nursing documentation were cus-tomized to ensure clarity duringphysician ordering and patientcharting, respectively. Oxygen tar-gets and alarms were displayedvisually at the bedside and wereexpected to be reviewed dailyduring medical rounds. The dailyreview was informally audited bythe Sat Pack and formalized in May2014.

2. Incorporate alarm data in patientcare assignments. Review of SpO2

data showed that 10% of patientswere responsible for .65% ofalarms, providing a new measureof acuity for nursing assignments.The information technologisttranslated individual patient alarmfrequency data into a “patientalarm barometer.” This barometercolor-codes the number of SpO2

alarms from a 24-hour period ina unitwide summary view. Chargenurses used this tool in con-structing daily nurse-patientassignments, with instructions toavoid pairing patients having fre-quent alarms with the same nurse.The barometer was reviewed withthe on-service medical teams dur-ing the daily multidisciplinarymorning report. Nurse opinionsand all provider opinions weresurveyed to assess the utility ofthis tool in October 2013 and April2014, respectively.

3. Evaluate and modify SANS escala-tion algorithms. Although bothhigh and low SpO2 are quantita-tively associated with harm invulnerable populations overtime,7–9 the immediate need todecrease false alarms requiredcareful consideration of the sec-ondary notification system algo-rithms. Because low SpO2 andheart rate changes can be imme-diately associated with poor pa-tient outcome (ie, death), andbecause the intervals for escala-tion are arbitrarily determined,changes were first made to thehigh SpO2 alarm escalation

algorithm. Working with our in-formation technologist, the algo-rithm for high saturation alarmswas modified (Fig 2B). The totalnumber of alarms detected at thebedside monitor and the numberof alarms delivered through SANSwere measured. Severe ROP wastracked as a balancing measure.

4. Evaluate signal averaging time ona primary monitor. Pulse oxi-meters average the detectable sig-nal over time, and previous studieshave shown that short averagingtimes result in more alarms butlonger averaging times can resultin falsely lengthened events, sincean average of multiple shortevents may appear as 1 longevent.13 An averaging time of16 seconds was an existing alarmintervention, and small tests weremade with different averagingtimes. We maintained the 16-secondaveraging time to prevent an un-necessary increase in alarms, andthe importance of averaging timefor bedside care providers in clin-ical decision making was includedin educational materials.

Data Collection and Analysis

Alarm data were collected with theConnexall Middleware System andthe data stored in a structured querylanguage (SQL) server database(Microsoft, Redmond, Washington).Inbound alarm data were processedby redundant application componentsand archived to a SQL server databasewith data located on a storage areanetwork. Data were extracted viastandard SQL queries. Datapresentation was achieved througha combination of Excel charts andSQL server reporting services.

Weekly alarm rates were rationallysubgrouped (detection, escalation tosecondary nurse, and escalation toregional team), then plotted onannotated control charts with QICharts version 2.0.22 (ProcessImprovement Products, Austin,Texas). Because large denominators

4 KETKO et al by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

created overdispersion in u charts,u-prime (u9) charts were used whenappropriate. Statistical processcontrol was used, and natural breakpoints were identified throughspecial-cause variation using the4 conventional rules.14 Twenty-fivedata points were used to establishbaseline rates and initial controllimits, during which time special-cause variation rules were notapplied. After interventions began, ifspecial-cause variation was sustained,mean lines were adjusted to reflectthe new steady-state rates. Spurioussignals (eg, isolated data pointsoutside the control limits in theabsence of attributable causes) didnot result in mean line adjustments.Alarm data were not case-mixadjusted to account for periods ofhigher and lower aggregate acuity inthe NICU. For the purposes of thisanalysis, the alarm rates of 4 epochswere defined and analyzed: baselineperiod (August 26, 2012, to May 4,2013: 36 weeks), narrowed SpO2

period (May 5, 2013, to July 20, 2013:11 weeks), revised escalation period(July 21, 2013, to February 22, 2014:31 weeks), and histogram period(February 23, 2014, to August 16,2014: 25 weeks) (Table 1), althoughthe histogram intervention applied toonly 20% to 25% of the unit census.Wilcoxon rank sum was used asappropriate to generate traditionalP values (StatCrunch, IntegratedAnalytics, Great Falls, Virginia).

Although not designed for statisticalpower, the number of alarms andpercentage of time spent within thetarget range from the histogramtechnology pilot were compared byusing a nonparametric repeatedmeasures ANOVA (Friedman test)(GraphPad InStat, La Jolla, California).

Nursing and physician opinionssurrounding interventions wereassessed in October 2013 and againin April 2014 using Qualtricssoftware (Qualtrics, Provo, Utah). TheSat Pack monitored and responded topatient safety concerns prospectively,and deidentified voluntary patientsafety reports were retrospectivelyqueried for reference to “alarm” andindividually reviewed to determinewhether resuscitation was required,and if so, the incident’s relevance toSpO2 alarm management. Thediagnosis of ROP was tracked througha deidentified database from theVermont Oxford Network that isqueried quarterly for qualityimprovement review. ROP stagesbefore and after narrowed limits(January 2012 to April 2013 and May201 to October 2014, respectively)were compared using Wilcoxon ranksum to test for significance.

RESULTS

In a 32-month monitoring period,3 459 637 saturation alarms weredetected, and 9 031 344 total alarmswere detected at the bedside monitor

(Fig 4). Natural fluctuation in data ispresent and is attributed to unitacuity, patient characteristics, andcompliance with alarm settings. Theaverage number of SpO2 alarms permonitored patient-day increased from78 before narrowing of alarms to 105after intervention. Informaldiscussions and education led to aninitial increase in alarm frequency,followed by a sustained increase withthe pilot of narrowed alarms (Fig 4).The modified SANS algorithm for highSpO2 delivery resulted in animmediate and sustained decrease inthe escalation of high SpO2 alarms tonursing phones (Fig 5). Reorientationto histogram technology in early 2014was temporally associated witha decrease in high SpO2 alarms(Fig 5A). The histogram technologypilot for 7 patients resulted in a 7.2%maximum increase in time spentwithin the targeted saturation rangeand a maximum reduction of 120alarms per monitored patient (P = .13and P = .34, respectively). The pilotwas not designed for statisticalsignificance, but the staff reportedthat histogram technology wasa helpful tool in oxygen targeting(Fig 6), even though we could onlymonitor one-quarter of the unit withthis technology at any time point.

The October 2013 survey showedthat 94% of responding chargenurses reported the patient alarmbarometer to be at least a somewhathelpful tool in directing nursing

TABLE 1 High SpO2 Alarms and Escalations During Intervention Periods

Monitor and PrimaryRegistered Nurse

Escalation to SecondaryRegistered Nurse

Escalation to Team

Aggregate high SpO2 alarms 635 338 22 768 4479Aggregate monitored patient-days 28 003 28 003 28 003Aggregate alarm rate per monitored patient-day 22.69 0.81 0.16Baseline alarm rate per monitored patient-day 16.76 0.95 0.24Narrowed SpO2 period alarm rate per monitored patient-day 38.64 2.19 0.49Alarm rate change compared with baseline, % 131 132 108P compared with baseline rate ,.001 ,.001 ,.019

Revised escalation period alarm rate per monitored patient-day 25.45 0.51 0.05Alarm rate change compared with baseline, % 52 246 279P compared with baseline rate ,.001 ,.001 ,.001

Histogram period alarm rate per monitored patient-day 20.35 0.40 0.04Alarm rate change compared with baseline, % 21 258 281P compared with baseline rate ,.033 ,.001 ,.001

PEDIATRICS Volume 136, number 2, August 2015 5 by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

assignments, and 100% of fellowsand attending physicians found it tobe somewhat helpful in medicaldecision making on subsequentsurvey. Figure 6 displays the resultsof the multidisciplinary survey inApril 2014 (n = 69) regardingattitudes and perceptions in alarmfrequency, demonstrating that mostrespondents felt that alarm frequencyhad improved and alarm fatigue wasbeing addressed.

Patient safety reporting identified noserious events requiring resuscitationsecondary to alarm management.Although the study was not poweredto detect the impact of narrowedSpO2 alarms on ROP at a singleinstitution in this time period, thesedata were nevertheless reviewed.Reductions in incidence and severitywere noted, although they did notmeet statistical significance.

DISCUSSION

This report emphasizes theimportance of selecting appropriatebalancing measures for QI efforts.Alarm frequency wasa predetermined balancing measurefor the Sat Pack oxygen managementguidelines, ultimately allowing the

team to respond to increasing alarmfrequency before full-scaleimplementation of the newguidelines. Measuring alarms alsoquantified the unfortunate burden ofusing pulse oximetry alarms as a toolfor oxygen targeting. The number ofpulse oximetry alarms is staggering,but pulse oximetry alarms are only 1source of alarms in an intensive caresetting, with additional signalscoming from cardiac monitoring,respiratory support devices, andmedication delivery devices. Becausenot all alarms require intervention,the sheer number of false alarms is ofconcern. False alarms contribute toalarm fatigue; it has also beendemonstrated that as alarm reliabilitydecreases, response to the alarmsalso decreases.15 Bliss and Dunn16

demonstrated that the reliability ofthe alarm system becomes moreimportant in situations withincreasing workload, which certainlyhas implications when caring forcritically ill patients. Other reportshave identified the number of false-positive alarms and inappropriatelyset alarms to be key factors in alarmdesensitization.17

Acknowledging that alarmmanagement must be a collaborative

effort among disciplines was a pivotalfirst step in this initiative. A culturalchange was necessary throughout ourunit, transitioning from excessivealarm frequency being solelya nursing concern to everyone takingownership of the problem. It was onlyby engaging multiple diverse skill setsthat we were able to developstrategies to address unnecessaryalarm frequency.

We demonstrate the importance ofdetermining key drivers of alarmfrequency to develop improvementstrategies. The development of clearstrategies and guidelines for alarmsettings and training of personnel canminimize alarms18 and was includedin our Sat Pack guidelines. Gross et alemphasized the increasing need tocustomize alarm limits in health carefacilities,19 as individualizing alarmthresholds has been shown todecrease alarms, resulting in lessnursing distraction and alarmfatigue.4 By collaborating with ourinformation technologists, we wereable to maximize data collection toaid us in customizing the SpO2 alarmlimits for some of our mostchallenging patients while stilltargeting specific SpO2 goals.

The goal for reduced hyperoxia/ROPcan cause unintended secondaryquality metrics to deteriorate(alarms), but this process can bemitigated and even improved withthe modification of technology. Wehave shown that the modification oftechnology applications (ie,a secondary alert notification systemdelivery algorithm) can affect alarmfrequency and that evaluation ofequipment is essential to alarmsystem design. We havedemonstrated the potentialapplication of an alarm frequencymonitoring tool to heightenawareness and engage multiple teammembers in the problem-solvingeffort of alarm management. Finally,our data support thatindividualization of alarms combinedwith histogram technology can

FIGURE 4All alarms per monitored patient-day were captured from the bedside patient monitor starting inJanuary 2012. SpO2 alarms account for nearly half of all bedside monitor alarms. The increase inSpO2 and total alarms with education and narrowing of alarm limits in 2013 is shown.

6 KETKO et al by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

FIGURE 5Three u9 control charts demonstrate high SpO2 alarm rates, with steady-state means represented as solid lines and control limits as dashed lines.A (dots), alarms detected by the monitor, including those automatically sent to the primary nurse’s phone after a short delay. B (diamonds), subset ofalarms from panel A that were delivered to a secondary nurse’s phone, either from automatic time delay or purposeful escalation by the primary nurse.C (triangles), subset of alarms from panel B that were escalated to the phones of the regional care team, either from automatic time delay or purposefulescalation by the primary or secondary nurse. Narrowed SpO2 alarm criteria significantly increased detected alarms as well as alarms escalated tosecondary nurses and broader teams. Redesign work on the alarm delivery and escalation algorithms maintained above-baseline high SpO2 detection butreduced alarm escalations to below baseline rates.

PEDIATRICS Volume 136, number 2, August 2015 7 by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

contribute to decreasing alarmfrequency and improvedpatient outcomes and staffsatisfaction.

Strengths

This report provides numerical datato support the concerns raised by theJoint Commission. The ability toquantify the volume of alarms anduse this data to test strategies foralarm system design improvement isunique. This initiative outlinespotential alarm managementstrategies focused on bothunderstanding and integratingtechnology and patientindividualization that could aid otherICUs in complying with the 2014NPSG. Our strategies heightenawareness of the potential risksassociated with clinical alarms (NPSGphase 1) while also identifyingdifferent means to mitigate thoserisks (NPSG phase 2).20 The JointCommission aims for health carefacilities to identify which alarmsneed to be targeted for bettermanagement based on input frommedical staff, the potential risks for

patient harm if not appropriatelyattended, and which alarmscontribute most significantly tounnecessary alarm frequency. In theNICU, saturation alarms were readilyidentified as the most appropriatealarms to be targeted owing to theirfrequency. Our alarm managementstrategies address how to determineclinically appropriate values foralarm limits, when to individualizesettings, and how to monitor alarmfrequency along with false-alarmfrequency.

Limitations

The Sat Pack oxygen managementguidelines and subsequent alarmmanagement strategies arepredicated on current best evidencefor oxygen targeting that itself hascontroversies.21 After itsdevelopment, pulse oximetry quicklybecame a standard of care inmedicine,22 yet the application tomitigate morbidity and mortality hasremained elusive.23 Results fromrecent oxygen targeting trials inextremely premature neonatesfurther emphasize this point.24

Acknowledging the effect of alarmsettings on the time spent within thetarget range,11 we elected to use theSpO2 alarm limits of 85% to 95% thatwere used in the Surfactant, PositivePressure, and OxygenationRandomized Trial (SUPPORT) as aninitial step in narrowing our limitsfrom 80% to 97%. A secondlimitation is that our auditing processfor alarm and target compliance wasinformal and insufficient. We havesince developed processes to betteraudit alarm compliance. Last, despitethe strategies implemented tomanage alarms, we have not achieveda full return to baseline (Fig 5). Webelieve that increased emphasis onoxygen targeting, application ofhistogram technology to a minority ofthe unit, and the process for alarmindividualization contribute to this.Individualization of patient alarms isnot an automated system, and it maytake several days of evaluation andadjustments to set new limits,accumulating alarms in the meantime.Most importantly, providers believethat alarm management has improvedin the NICU.

FIGURE 6Perception and attitudes of NICU nurses, respiratory therapists, nurse practitioners, and physicians in April 2014 regarding alarm frequency andinterventions to improve alarm fatigue.

8 KETKO et al by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

Future Directions

We are securing histogram technologyfor all patients. We anticipate that bystandardizing the use of this technology,we will be able to better identify timewithin the target range andindividualize alarm settings whenneeded, resulting in fewer overallalarms while maintaining targetsaturation goals.12 Our multidisciplinaryimprovement task force will continue tomonitor alarm data as we return focusto the original QI efforts of reducingsevere ROP. We are merging our oxygenmanagement and alarm managementguidelines to have a unified algorithm,emphasizing how the two areinterdependent. We have identified andimplemented several strategies in aninitial phase to improve alarmfrequency, but alarm managementcontinues to be a struggle in ourunit and will require futuremultidisciplinary collaborative efforts inaddition to cultural practice changesto achieve significant qualityimprovement.

ACKNOWLEDGMENTS

A special thanks to RobertSchumacher, MD, and Jake Seagull,PhD, for their editing contributions,mentorship, and overall expertise inthe field of quality improvement andhuman factors, respectively. We thankSusan Hieber, who assisted withfigure editing, and the staff at AkronChildren’s Hospital NICU for theircollaboration and others in theVermont Oxford Network forguidance and feedback. The membersof our unit’s Oxygen and AlarmManagement Task Force—the SatPack—were invaluable to this work.

ABBREVIATIONS

QI: quality improvementROP: retinopathy of prematuritySANS: secondary alert notification

systemSpO2: pulse oxygen saturationSQL: structured query language

REFERENCES

1. Blake N. The effect of alarm fatigue onthe work environment. AACN Adv CritCare. 2014;25(1):18–19

2. Harris RM, Manavizadeh J, McPhersonDJ, Smith L. Do you hear bells? Theincreasing problem of alarm fatigue.Pa Nurse. 2011;66(1):10–13

3. Topf M, Dillon E. Noise-induced stress asa predictor of burnout in critical carenurses. Heart Lung. 1988;17(5):567–574

4. Cvach M. Monitor alarm fatigue: anintegrative review. Biomed InstrumTechnol. 2012;46(4):268–277

5. Citing reports of alarm-related deaths,the Joint Commission issues a sentinelevent alert for hospitals to improvemedical device alarm safety. ED Manag.2013;26(6):suppl 1–suppl 3

6. Weil KM. Alarming monitor problems.Nursing. 2009;39(9):58

7. Carlo WA, Finer NN, Walsh MC, et al;SUPPORT Study Group of the EuniceKennedy Shriver NICHD NeonatalResearch Network. Target ranges ofoxygen saturation in extremely preterminfants. N Engl J Med. 2010;362(21):1959–1969

8. Schmidt B, Whyte RK, Asztalos EV, et al;Canadian Oxygen Trial (COT) Group.Effects of targeting higher vs lowerarterial oxygen saturations on death ordisability in extremely preterm infants:a randomized clinical trial. JAMA. 2013;309(20):2111–2120

9. Khadawardi E, Al Hazzani F. Oxygensaturation and outcomes in preterminfants the BOOST II United Kingdom,Australia, and New Zealand CollaborativeGroups. J Clin Neonatol. 2013;2(2):73–75

10. Laptook AR, Salhab W, Allen J, Saha S,Walsh M. Pulse oximetry in very lowbirth weight infants: can oxygensaturation be maintained in the desiredrange? J Perinatol. 2006;26(6):337–341

11. Hagadorn JI, Furey AM, Nghiem TH, et al;AVIOx Study Group. Achieved versusintended pulse oximeter saturation ininfants born less than 28 weeks’gestation: the AVIOx study. Pediatrics.2006;118(4):1574–1582

12. Bizzarro MJ, Li FY, Katz K, Shabanova V,Ehrenkranz RA, Bhandari V. Temporalquantification of oxygen saturationranges: an effort to reduce hyperoxia in

the neonatal intensive care unit.J Perinatol. 2014;34(1):33–38

13. Ahmed SJ, Rich W, Finer NN. The effect ofaveraging time on oximetry values in thepremature infant. Pediatrics. 2010;125(1). Available at: www.pediatrics.org/cgi/content/full/125/1/e115

14. Western Electric Company. Statisticalquality control handbook. 1st ed.Indianapolis, IN: Western Electric Co;1956

15. Guardia-Labar LM, Scruth EA, Edworthy J,Foss-Durant AM, Burgoon DH. Alarmfatigue: the human-system interface. ClinNurse Spec. 2014;28(3):135–137

16. Bliss JP, Dunn MC. Behaviouralimplications of alarm mistrust asa function of task workload. Ergonomics.2000;43(9):1283–1300

17. Christensen M, Dodds A, Sauer J, WattsN. Alarm setting for the critically illpatient: a descriptive pilot survey ofnurses’ perceptions of current practicein an Australian Regional Critical CareUnit. Intensive Crit Care Nurs. 2014;30(4):204–210

18. Welch J. An evidence-based approach toreduce nuisance alarms and alarmfatigue. Biomed Instrum Technol. 2011;(Suppl):46–52

19. Gross B, Dahl D, Nielsen L. Physiologicmonitoring alarm load on medical/surgicalfloors of a community hospital. BiomedInstrum Technol. 2011;(Suppl):29–36

20. Alarm system safety. R3 Report, issue 5,2013. Oakbrook Terrace, IL: The JointCommission. Available at: http://www.jointcommission.org/r3_report_issue5/.Accessed March 9, 2015

21. Polin RA, Bateman D. Oxygen-saturationtargets in preterm infants. N Engl J Med.2013;368(22):2141–2142

22. Hay WW. History of pulse oximetry inneonatal medicine. NeoReviews. 2005;6(12):e533–e538

23. Silverman WA. A cautionary tale aboutsupplemental oxygen: the albatross ofneonatal medicine. Pediatrics. 2004;113(2):394–396

24. Saugstad OD, Aune D. Optimaloxygenation of extremely low birthweight infants: a meta-analysis andsystematic review of the oxygensaturation target studies. Neonatology.2014;105(1):55–63

PEDIATRICS Volume 136, number 2, August 2015 9 by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

originally published online July 6, 2015; Pediatrics Rebecca J. Vartanian

Anastasia K. Ketko, Craig M. Martin, Michelle A. Nemshak, Matthew Niedner andNICU

Balancing the Tension Between Hyperoxia Prevention and Alarm Fatigue in the

ServicesUpdated Information &

014-1550http://pediatrics.aappublications.org/content/early/2015/06/30/peds.2including high resolution figures, can be found at:

Subspecialty Collections

y_subhttp://classic.pediatrics.aappublications.org/cgi/collection/neonatologNeonatologyorn_infant_subhttp://classic.pediatrics.aappublications.org/cgi/collection/fetus:newbFetus/Newborn Infantprovement_subhttp://classic.pediatrics.aappublications.org/cgi/collection/quality_imQuality Improvementtion:practice_management_subhttp://classic.pediatrics.aappublications.org/cgi/collection/administraAdministration/Practice Managementfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

https://shop.aap.org/licensing-permissions/in its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://classic.pediatrics.aappublications.org/content/reprintsInformation about ordering reprints can be found online:

ISSN: . 60007. Copyright © 2015 by the American Academy of Pediatrics. All rights reserved. Print American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,has been published continuously since . Pediatrics is owned, published, and trademarked by the Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from

originally published online July 6, 2015; Pediatrics Rebecca J. Vartanian

Anastasia K. Ketko, Craig M. Martin, Michelle A. Nemshak, Matthew Niedner andNICU

Balancing the Tension Between Hyperoxia Prevention and Alarm Fatigue in the

http://pediatrics.aappublications.org/content/early/2015/06/30/peds.2014-1550located on the World Wide Web at:

The online version of this article, along with updated information and services, is

ISSN: . 60007. Copyright © 2015 by the American Academy of Pediatrics. All rights reserved. Print American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,has been published continuously since . Pediatrics is owned, published, and trademarked by the Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on April 24, 2018http://pediatrics.aappublications.org/Downloaded from