Identifying Pediatric Emergence Delirium by Using the PAED Scale: a Quality Improvement Project

Identifying Pediatric Emergence Delirium by Using the PAED Scale: A Quality Improvement Project

MATTHEW J. STAMPER, DNP, CRNA; SHARON J. HAWKS, DNP, CRNA; BRAD M. TAICHER, DO, MBA; JULIET BONTA, BSN, RN; DEBRA H. BRANDON, PhD, RN, CCNS, FAAN

ABSTRACT

Pediatric emergence delirium is a postoperative phenomenon characterized by aberrant cognitive and psychomotor behavior, which can place the patient and health care personnel at risk for injury. A common tool for identifying emergence delirium is the Level of Consciousness-Richmond Agitation and Sedation Scale (LOC- RASS), although it has not been validated for use in the pediatric population. The Pediatric Anesthesia Emergence Delirium Scale (PAED) is a newly validated tool to measure emergence delirium in children. We chose to implement and evaluate the effectiveness and fidelity of using the PAED Scale to identify pediatric emergence delirium in one eight-bed postanesthesia care unit in comparison with the traditional LOC-RASS. The overall incidence of pediatric emergence delirium found by using the LOC-RASS with a retrospective chart review (3%) was significantly lower than the incidence found by using the LOC-RASS (7.5%) and PAED Scale (11.5%) during the implementation period. Our findings suggest that the PAED Scale may be a more sensitive measure of pediatric emergence delirium, and, in the future, we recommend that health care personnel at our facility use the PAED Scale rather than the LOC-RASS. AORN J 99 (April 2014) 480-494. Ó AORN, Inc, 2014. http:// dx.doi.org/10.1016/j.aorn.2013.08.019

Key words: pediatric emergence delirium, Pediatric Anesthesia Emergence Delirium Scale, PAED Scale, Level of Consciousness Richmond Agitation and Sedation Scale, LOC-RASS, quality improvement.

he term emergence delirium is deﬁned as in the pediatric population.2-9 Children who expe- “a mental disturbance during the recovery rience emergence delirium postoperatively are at T from general anesthesia consisting of hal- risk for unintentional self-injury, including removal lucinations, delusions, and confusion manifested by of catheters and drains as well as wound dehis- moaning, restlessness, involuntary physical activ- cence.10 Additionally, management of pediatric ity, and thrashing about in the bed.”1(p19) Although emergence delirium requires extra attention from emergence delirium occurs in patients of all ages, it nursing personnel, placing undue strain on nurse- appears to occur three to eight times more frequently to-patient ratios and resource allocation in the

http://dx.doi.org/10.1016/j.aorn.2013.08.019 480 j AORN Journal April 2014 Vol 99 No 4 Ó AORN, Inc, 2014 PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org postanesthesia care unit (PACU).11 A recent study frequently. The incidence is highest among chil- showed that 49% of children who developed dren who emergence delirium required extra PACU personnel n are younger than six years of age, to care for them compared with only 15% of chil- n have preoperative anxiety, and 9 dren not experiencing emergence delirium. Pedi- n emerge rapidly from general anesthesia.4,7,16 atric emergence delirium also can place the RN at risk for injury if he or she tries to restrain the child Additionally, the use of sevoflurane and surgical to keep the child from hurting himself or herself.11 procedures involving the head and neck are asso- One of the keys to managing pediatric emer- ciated with an increased risk of developing pedi- 17 gence delirium is prompt recognition of the con- atric emergence delirium. Several commonly dition by anesthesia and PACU personnel. However, used IV medications (eg, propofol, fentanyl, keta- to date, there is not a recognized standard assess- mine, dexmedetomidine) that are administered at ment to identify pediatric emergence delirium. various times during the perioperative period are Consequently, a variety of rating and visual analog associated with a decreased incidence of pediatric 5,6,13,18 scales are used, many of which have not been emergence delirium. validated in the pediatric surgical population.12 Although these findings offer information about This practice variation, combined with differing preemptive medication management for children at anesthetic techniques, likely contributes to the wide risk for developing emergence delirium, the true range in the incidence of pediatric emergence effect of these strategies remains unknown because delirium (ie, 2% to 80%) that is reported in the patients often are not assessed with a validated literature.13 pediatric tool. Furthermore, without appropriate tool use, it is difficult to differentiate between DESCRIPTION OF THE PROBLEM emergence delirium and the signs and symptoms of At our facility, health care providers use the Level of pain because both emergence delirium and pain Consciousness-Richmond Agitation and Sedation may be characterized by restless behavior.12 One of Scale (LOC-RASS) to identify pediatric emergence the greatest attributes of the PAED Scale, however, delirium in the PACU. Although the LOC-RASS is that it was developed by using a theoretical is a valid and reliable measure of sedation and framework of delirium, focusing on changes in agitation in critically ill adults, it has not been consciousness and cognition, which permits dif- validated in the pediatric population.14,15 The ferentiation of emergence delirium behaviors from Pediatric Anesthesia Emergence Delirium (PAED) those behaviors associated with pain.12 Scale is a newly validated measure of emergence The PAED Scale consists of five characteristics delirium for children.12 that are each scored by using a 5-point Likert scale. The scores are summed into a final composite score REVIEW OF THE LITERATURE to determine the presence or absence of pediatric Much of the pediatric emergence delirium research emergence delirium. Two items, “the child makes is focused on determining the etiology of emer- eye contact with the caregiver” and “the child is gence delirium and on using pharmacological in- aware of his [or] her surroundings,” respectively, terventions to reduce its incidence. The cause of are reflections of the child’s state of consciousness pediatric emergence delirium is likely multifacto- and ability to focus attention and to thoughtfully rial; however, the following characteristics, either organize external stimuli. One item, “the child’s individually or in combination, have been identified actions are purposeful,” addresses changes in in children who develop emergence delirium more cognition that cause the child to behave in a manner

AORN Journal j 481 April 2014 Vol 99 No 4 STAMPER ET AL outside of what is deemed appropriate. Two addi- development within this population, the assessment tional items, “the child is restless” and “the child is criteria appear to be applicable to pediatric patients inconsolable,” are measures of psychomotor and outside of this age group, which has led researchers emotional behavior, respectively, that are associ- to examine the incidence of emergence delirium in ated with delirium. Although the last two items children of all ages.6,7,9,20 Therefore, we chose to also may indicate that the patient is experiencing implement the PAED Scale to provide a more pain, assessing these measures in conjunction clearly defined set of criteria for identification of with other indicators of consciousness and cogni- pediatric emergence delirium at our facility. tion helps the health care provider differentiate between emergence delirium and pain.12 PROJECT GOALS AND METHODS The developers of the PAED provided strong The primary goal of this project was to evaluate the evidence of measurement reliability and validity, identification of pediatric emergence delirium by with an internal consistency of 0.89, with delirium implementing and evaluating the use of the PAED characteristics of the Diagnostic and Statistical Scale in the children’s PACU at our facility. We Manual of Mental Disorders,19 interobserver re- compared the incidence of pediatric emergence liability of 0.84 (95% confidence interval [CI], delirium as determined by using the LOC-RASS 0.76-0.90), and sensi- during a retrospective tivity and specificity chart review to the of 0.64 and 0.86, Although the PAED Scale originally was incidence as deter- respectively, when developed to assess for emergence delirium mined by using the the composite PAED among patients between the ages of 18 months PAED Scale during scores were higher and six years of age because of a higher pro- the implementation than or equal to 10.12 pensity for emergence delirium development period. Additionally, In contrast, the within this population, the assessment criteria given that the LOC- LOC-RASS originally appear to be applicable to pediatric patients RASS remained in use was developed as “a outside of this age group. throughout the project, clinically useful tool we examined differ- to assess the level of ences in the incidence consciousness and agitated behavior in (adult) of pediatric emergence delirium by using the LOC- [intensive care unit (ICU)] patients [to] guide RASS between the retrospective and implementa- sedation therapy and improve communication tion periods. A secondary goal was to assess whether among health care providers.”14(p1341) The LOC- patient characteristics (ie, age, gender, medical RASS can be an effective component of an algo- diagnosis of attention deficit hyperactivity disorder rithm to identify pediatric emergence delirium, but [ADHD], premedication with midazolam, parental it should be used in conjunction with the PAED presence on induction, type of anesthetic mainte- Scale.15 To our knowledge, no published literature nance, type of surgery) predicted the development compares LOC-RASS scores with the construct of pediatric emergence delirium during the imple- of pediatric emergence delirium, nor has a thresh- mentation period. old score for the presence of pediatric emergence Our final goal was to assess the fidelity (ie, a delirium been identified for the LOC-RASS. Although quality measurement of a tool or intervention the PAED Scale originally was developed to assess available to the health care providers that, when for emergence delirium among patients between used correctly, results in an accurate or intended the ages of 18 months and six years of age because outcome) of the PAED Scale by evaluating the of a higher propensity for emergence delirium structure, process, and outcome of implementation

482 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org according to the Quality of Care Framework.21 the children’s PACU were excluded from this Measures included the following factors: project. Retrospective patient data were retrieved n StructuredAre PACU RNs using the PAED from perioperative electronic health records exactly Scale consistently and appropriately? one year before the implementation period. We n ProcessdAre the attending anesthesiologists chose this time period to avoid influence from the responding to episodes of pediatric emergence education that was required for perioperative delirium when notified by a PACU RN? personnel in the implementation period. n OutcomedAre patients experiencing subse- A power analysis, based on an estimated 15% quent or sustained episodes of pediatric emer- incidence of pediatric emergence delirium, indi- gence delirium during the next assessment cated that a sample size of 150 patients in each period after pharmacological treatment? period would provide 80% power to detect group differences in the incidence of pediatric emergence delirium. However, we decided to include 200 SETTING children in each group in case the incidence of We conducted this quality improvement project in pediatric emergence delirium was less than ex- an eight-bed children’s PACU housed within a 928- pected. We estimated the incidence based on a bed academic facility in the southeastern United prospective cohort study that included 521 patients States. Anesthesia professionals (ie, certified reg- between three and seven years of age that showed istered nurse anesthetists [CRNAs], anesthesia that 18% of the patients undergoing general anes- residents, attending anesthesiologists), and RNs thesia for an outpatient surgical procedure experi- provide perioperative services for approximately enced an episode of emergence delirium in the 5,800 pediatric patients who present to our facility postoperative period.4 Although the authors of that annually for surgical and diagnostic procedures study referred to the event as “emergence agitation,” with a daily PACU census, ranging from 15 to 35 the definition they used to classify the event was patients. The anesthesia professionals work col- consistent with the definition of emergence de- laboratively within an Anesthesia Care Team lirium used in our project (ie, agitation with non- Model to manage preoperative anxiety, intra- purposeful movement, restlessness, or thrashing; operative anesthesia and analgesia, postoperative incoherence; inconsolability; unresponsiveness4). 22 pain, and cardiovascular or respiratory events. The More recent studies using the PAED Scale to assess RNs perform preoperative and postoperative as- for the presence or absence of pediatric emergence sessments, manage postoperative pain, and notify delirium have shown incidences ranging from 4.8% the anesthesia professionalsdprimarily attending to 47.2%, depending on the type of maintenance anesthesiologistsdof deteriorations in a patient’s status. anesthetic used or the administration of prophy- lactic medication.5-9 SAMPLE We used a convenience sample of 400 patients, 200 DESIGN AND MEASURES from the retrospective audit and 200 from the We used a pre-post observational design in which implementation period. We included any patient we compared the baseline incidence of pediatric 17 years of age or younger who presented to the emergence delirium determined by LOC-RASS children’s preoperative area and who subsequently data collected from a retrospective chart review returned to the PACU after undergoing general an- with the incidence of emergence delirium deter- esthesia for a surgical or diagnostic procedure in mined by PAED Scale and LOC-RASS data the sample for each period. Only those patients who collected during the implementation period. We were transported to an ICU or recovered outside of took the LOC-RASS data that we collected during

AORN Journal j 483 April 2014 Vol 99 No 4 STAMPER ET AL the retrospective chart review from the patients’ particular time period. We set the threshold score PACU records. The number of LOC-RASS as- for the presence of pediatric emergence delirium by sessments conducted on any one patient from the using the PAED Scale at higher than or equal to 10 retrospective chart review ranged from one to three, based on sensitivity and specificity reported in the and the time between assessments was not stan- literature.12 dardized. This was in contrast to the way LOC- RASS and PAED Scale scores were collected LOC-RASS during the implementation period. The LOC-RASS was developed to “establish a During the implementation period, the RNs used clinically useful tool to assess the level of con- the LOC-RASS and PAED Scale to assess for pe- sciousness and agitated behavior in ICU patients diatric emergence delirium during four time periods that might guide sedation therapy and improve com- 14(p1341) separated by 10-minute intervals. We chose 10-minute munication among health care providers.” time intervals based on a previous study20 that used One of the salient features is the single-item a similar protocol to compare the incidence of pe- scoring system, which negates the need to add diatric emergence delirium found by using the multiple subscale scores and can be completed PAED Scale to the incidence found by using the quickly through observation and assessment of 14 Watcha behavior scale for emergence delirium, patient responses to auditory and physical stimuli. the Cravero emergence agitation scale, and by an The LOC-RASS was found to have high interrater experienced anesthesia observer. In our project, reliability (0.956) and strong construct validity with we denoted the first assessment time period as respect to a visual analog scale (r ¼ 0.84-0.98, all 14 “Time 0,” which corresponded with arrival in the P < .0001). PACU. The second, third, and fourth assessment The LOC-RASS assesses five levels of sedation: time periods occurred 10 minutes, 20 minutes, and n 5 ¼ unarousable with no response to voice or 30 minutes after arrival in the PACU, respectively. physical stimulation n 4 ¼ deep sedation with no response to voice PAED but movement to physical stimuli ¼ The PAED Scale consists of five characteristics that n 3 moderate sedation with movement to voice are each scored by using a 5-point Likert scale. The but no eye contact ¼ first three items are scored in reverse order: n 2 light sedation and briefly awakens (ie, less than 10 seconds) with eye contact to voice n 4 ¼ not at all n 1 ¼ drowsy (ie, not fully alert) but has sus- n 3 ¼ just a little tained (ie, more than 10 seconds) awakening n 2 ¼ quite a bit with eye contact to voice n 1 ¼ very much n 0 ¼ extremely A score of zero corresponds to a calm state of alertness. Positive scores on the LOC-RASS con- Items four and five are scored in regular order: tinuum denote some level of agitation: n 0 ¼ not at all n þ1 ¼ restless, anxious, or apprehensive but n 1 ¼ just a little movements are not aggressive or vigorous n 2 ¼ quite a bit n þ2 ¼ agitated, frequent, nonpurposeful move- n 3 ¼ very much ments or ventilator dyssynchrony n 4 ¼ extremely n þ3 ¼ very agitated, pulls on or removes tubes Individual scores for each item are then added or catheters, or has aggressive behavior toward together to determine the composite score for a personnel

484 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org n þ4 ¼ combative, overtly combative, or violent, whether the results from this project were consis- posing an immediate danger to personnel tent with current research findings. We set the threshold score for the presence of pediatric emergence delirium by using the LOC- Fidelity RASS at higher than or equal to þ2 because the Members of PQIT used the fidelity data that RNs corresponding characteristics are more closely in the pediatric PACU collected during the imple- aligned with the definition of emergence delirium mentation period to evaluate implementation of the that we used for our project. Additionally, setting PAED Scale. the threshold score at þ1 would have classified all n We assessed the construct of fidelity structure patients as having emergence delirium who were by examining the number of patients receiving not sedated or who were in a calm state. Compar- PAED and LOC-RASS scores as documented atively, the PAED Scale allows the potential for by the RNs during all four assessment time alterations in behavior periods relative to and cognition beyond the total number that of being in a calm The LOC-RASS was developed to “establish of patients who and alert state, thus we a clinically useful tool to assess the level of presented to the chose to include the at- consciousness and agitated behavior in ICU PACU. We felt tributes associated with patients that might guide sedation therapy that successful im- a LOC-RASS of þ1 and improve communication among health plementation of the within the acceptable care providers.” innovation would range of emergence have been achieved characteristics. if 85% or more of all patients were assessed by using the LOC- PAED Data Collection Form RASS and PAED Scale in all four assessment The CRNAs and anesthesia residents used the time periods. PAED Data Collection Form to document infor- n We evaluated the construct of the fidelity pro- mation describing the patient population during cess by examining how anesthesia and nursing the implementation period. Data collected included professionals used the patient assessment data age, gender, medical diagnosis of ADHD, pre- from the LOC-RASS and PAED Scale. Thus, medication with midazolam, midazolam effective- the process examined the number and type of ness, effect of parental presence on induction, type anesthesia professionals’ responses to each pa- of anesthetic maintenance, and type of surgery. tient who met the criteria for pediatric emer- Members of the PAED Quality Improvement Team gence delirium, as communicated by the PACU (PQIT)dconsisting of a CRNA, an attending an- RN, relative to the total number of patients esthesiologist, a pediatric PACU RN, and Duke who met the criteria for pediatric emergence Nursing School facultydused these data as com- delirium. A response was defined as coming parative measures within and between the retro- to the patient’s bedside, writing an order for spective and implementation groups as well as to or administering medication, or a combination determine whether there was an association be- of both. We considered a provider response tween these characteristics and the development of rate of 85% or more to be a positive reflection pediatric emergence delirium during the imple- of this measure. mentation period. The measure of these character- n We measured the construct of fidelity outcome istics within our PACU population served to determine by examining the number of patients who

AORN Journal j 485 April 2014 Vol 99 No 4 STAMPER ET AL

experienced a second (ie, either subsequent or the proposed plan to implement the PAED Scale sustained) episode of emergence delirium dur- into the PACU RNs’ assessment. The RNs who ing the RN assessment period after pharmaco- were unable to attend the inservice program logical intervention compared with the number received individual education before project of patients receiving no pharmacological inter- implementation. vention. We defined a subsequent or sustained After the inservice program, the RNs were given emergence delirium episode as a PAED score the opportunity to practice using the PAED Scale. of higher than or equal to 10 or LOC-RASS Two members of the PQIT were available during of higher than or equal to þ2 in the ensuing the five-day trial period to answer questions and assessment time period. We set an overall rate address any concerns expressed by the RNs as they of 15% or fewer patients experiencing subse- practiced using the new assessment tool. During quent or sustained emergence delirium after this same time period, a member of the PQIT pharmacological intervention as a positive contacted all anesthesia professionals who had the measure for this construct. potential to care for pediatric patients during the PAED Scale implementation period in person or An answer of “N/A” to the fidelity questions meant via e-mail if they were not readily available to the questions were not applicable to the patient (ie, inform them of the implementation start date and the patient did not experience an episode of emer- of their data collection responsibilities. Addition- gence delirium [ie, LOC-RASS < þ2orPAED ally, a member of the PQIT asked the anesthesia < 10 during each assessment period] and required professionals not to change their current practices no pharmacological intervention). The fidelity of managing pediatric emergence delirium. questions were Preoperative RNs, anesthesia professionals, and n Did you call the attending anesthesiologist to PACU RNs worked cooperatively to collect report a PAED score of higher than or equal prospective perioperative data on 200 pediatric to 10? patients during a three-week period. Specific re- n Did you call the attending anesthesiologist to sponsibilities during the implementation period report a LOC-RASS score of higher than or included the following: equal to þ2? n A preoperative RN placed the PAED Data n Did the attending anesthesiologist come to the Collection Form in the front of each patient’s bedside or order or administer medication? chart in the preoperative area, and this form n If medication was administered, did the patient stayed in the chart until the patient was dis- experience a sustained or subsequent episode of charged from the children’s PACU. emergence delirium during the next assessment n The CRNAs and resident anesthesiologists period (ie, 10 minutes later)? documented patient age and gender, premedication with midazolam, the effectiveness IMPLEMENTATION AND DATA COLLECTION of midazolam administration in reducing After the PQIT received approval from the internal the patient’s anxiety, medical diagnosis of review board for the project through an expedited ADHD, parental presence during induction, review process, and five days before implementa- type of maintenance anesthetic, and type of tion of the PAED Scale, a member of the PQIT procedure. conducted an educational inservice program for the n The PACU RN completed the remainder of the PACU RNs. This inservice program focused on the PAED Data Collection Form during the post- implications of pediatric emergence delirium and operative period, which consisted of LOC-RASS

486 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org

and PAED assessments, along with answering time periods. We compared age by using a Wil- the project fidelity questions. coxon rank sum test as a result of non-normality of A member of the PQIT posted laminated copies the distribution, and we compared nonparametric of the PAED Scale on the tables in each of the data, including gender, premedication with mid- recovery bays to ensure that the RNs had instant azolam, parental presence at induction, and type access to the assessment criteria. Nurses conducted of maintenance anesthetic by using Fisher exact assessments for each patient by using the PAED tests to identify differences between the retro- Scale and LOC-RASS during all four assessment spective and implementation periods. We used z time periods. The RNs documented both scores scores and corresponding ESs to test for differ- on the PAED Data ences in the propor- Collection Form at the tion of patients who time of the assessment The decision to provide a pharmacological experienced pediatric and notified the at- intervention was left to the discretion of the emergence delirium tending anesthesiolo- attending anesthesiologist. The RNs docu- relative to the patients gist if the patient met mented whether the attending anesthesiologist who did not as deter- the threshold score met the criteria of a “response” and whether mined by the LOC- criteria for experi- the patient experienced a subsequent or a RASS and PAED encing pediatric em- sustained episode of pediatric emergence Scale used in their ergence delirium by delirium at the next assessment time period. respective periods. using either assess- Additionally, we ment tool, regardless used logistic regres- of the findings from the other tool. The decision to sion models to determine whether any significant provide a pharmacological intervention was left to differences between group variables or patient the discretion of the attending anesthesiologist. The characteristic data collected during the imple- RNs documented whether the attending anesthesi- mentation period relative to PAED Scale scores ologist met the criteria of a “response” and whether (ie, age younger than or equal to six years; pre- the patient experienced a subsequent or a sustained medication with midazolam; a medical diagnosis episode of pediatric emergence delirium at the next of ADHD; parental presence at induction; sevo- assessment time period. The RNs placed the PAED flurane used for maintenance anesthetic; procedures Data Collection Form in a locked box after the involving the ear, nose, and throat) were associated patient was discharged from the PACU and a with the development of pediatric emergence member of the PQIT collected all of the forms delirium. The level of significance was set at P ¼ before the start of the next day. After the imple- .05 for all tests. mentation period, a member of the information technology team pulled the retrospective informa- RESULTS tion consisting of LOC-RASS and patient charac- Members of the PQIT and a statistician analyzed teristic data from the patients’ charts. data from 200 patients in both the retrospective and implementation periods (N ¼ 400). Perioperative DATA ANALYSIS data collected for 49 patients were not included in A member of the PQIT and a statistician analyzed the analysis because these patients did not receive data by using SAS/STATÒ 9.3 software. We used PAED Scale and LOC-RASS assessments during descriptive statistics to determine frequency distri- all four time periods. There were no statistically butions among the patient characteristic and fidelity significant differences between retrospective and data collected during the respective assessment implementation groups with respect to age, gender,

AORN Journal j 487 April 2014 Vol 99 No 4 STAMPER ET AL

TABLE 1. Comparison of Sample Characteristics Between Time Periods

Characteristic Retrospective period Implementation period Pa

Age (years) N ¼ 200b N ¼ 198b .511c Mean (SD) 7.9 (5.2) 7.6 (5.2) Minimum, maximum 0.1, 17 0.1, 17 Gender .421d Male n/N (%)b 107/200 (53.5%) 116/200 (58%) Female n/N (%)b 93/200 (46.5%) 84/200 (42%) ADHDe N/Af 12/196 (6.1%) Premedication with midazolam 55/200 (27.5%) 85/200 (42.5%) .002d Premedication with midazolam effective N/Af 79/83 (95.2%) N/Af Parental presence at induction 43/200 (21.5%) 53/197 (26.9%) .241d Maintenance anesthetic N ¼ 181b N ¼ 198b .049d Sevoflurane 166/181 (91.7%) 181/198 (91.4%) Isoflurane 3/181 (1.7%) 11/198 (5.6%) Desflurane 1/181 (0.6%) 0 Total IV anesthesia 11/181 (6.1%) 6/198 (3.0%) Type of surgery N ¼ 196b N ¼ 198b Dental 7 (3.6%) 10 (5.1%) ENT 24 (12.2%) 14 (7.1%) Endoscopy 24 (12.2%) 35 (17.7%) MRI/CT 5 (2.6%) 15 (7.6%) Neurosurgery 9 (4.6%) 11 (5.6%) Orthopedics 32 (16.3%) 22 (11.1%) General surgery 38 (19.4%) 26 (13.1%) Urology 23 (11.7%) 16 (8.1%) Otherg 34 (17.4%) 49 (24.8%)

CT ¼ computed tomography; ENT ¼ ears, nose, and throat; Max ¼ maximum age; Min ¼ minimum age; MRI ¼ magnetic resonance image. a P is significant at .05. b The N for each individual characteristic represents the number of patients with documented data. Total charts reviewed in the retrospective period ¼ 200. Total patient encounters in the implementation period ¼ 200. N < 200 represents data that were not collected for a particular variable. c Medically diagnosed with attention deficit hyperactivity disorder. d Wilcoxon two-sample test. e Fisher exact test. f Information not available. g Minor procedures performed (ie, bronchoscopy, cutaneous arteriovenous malformation treatments, bone marrow biopsy, lumbar punctures with or without administration of intrathecal chemotherapy). or parental presence at induction. Various patient difference (P ¼ .049) in the type of maintenance characteristics that were analyzed to determine anesthetic used between the two periods, with statistically significant differences in patient char- sevoflurane being the most frequently used in each. acteristics are demonstrated in Table 1. We evaluated the potential confounding effects of We found that the proportion of patients re- preoperative midazolam administration and main- ceiving midazolam preoperatively was significantly tenance anesthetic during the retrospective and im- higher during the implementation period compared plementation periods to determine whether these with the retrospective period (P ¼ .002). Seventy- variables affected the incidence. A logistic re- nine (95.2%) of the anesthesia professionals who gression model, including variables “group” (ie, administered the midazolam documented that the retrospective, implementation periods), and “premedication was effective for relieving patients’ medication with midazolam,” demonstrated that preoperative anxiety. We also found a significant midazolam administration was not a significant

488 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org

2 covariate (c 1 ¼ 1.90; P ¼ .1685). Similarly, were evident; pediatric emergence delirium was the logistic regression model with the variables identified in 11 patients with the PAED Scale alone “group” and “maintenance anesthetic” also indi- and in three patients with the LOC-RASS alone. cated that the covariate maintenance anesthetic was The subgroups were pulled out to show that each 2 not statistically significant (c 3 ¼ .002; P ¼ .99). tool identified patients with pediatric emergence Thus, we did not include these two variables in delirium independent of the other tool that suggested the primary analysis of the emergence delirium that these patients did not have emergence delirium. outcomes as described below. Effect sizes (ESs) associated with each tool in A medical diagnosis of ADHD was not commonly its respective time period were calculated to de- recorded in the perioperative electronic health re- termine the clinical significance of the differences cord; therefore, we in pediatric emergence could not ascertain delirium incidence the incidence of Use of the PAED scale in the implementation found by using each ADHD among the period was more likely to identify patients tool. The comparison retrospective group experiencing emergence delirium than use between the incidence for comparison. Data of the LOC-RASS in the retrospective period. of emergence delirium regarding ADHD were found by using the collected during the PAED Scale during implementation period to determine whether ADHD the implementation period and that found by using was associated with the development of pediatric the LOC-RASS during the retrospective period emergence delirium. Only 12 of the patients (6.1%) resulted in an ES ¼ 0.79 (odds ratio [OR] ¼ 4.2 had a medical diagnosis of ADHD documented in [95% CI, 0.996-6.902]). The ESs are an expression their charts. The varied frequency distribution of of the strength of relationship between two vari- the types of surgery did not allow for statistical ables, ranging from 0e1.0, independent of the analysis of this variable. statistical significance (P). The higher the ES, the The overall incidence of pediatric emergence higher the probability that the dependent variable delirium in the implementation period by using (ie, incidence of emergence delirium) was influ- the LOC-RASS and PAED Scale were 7.5% and enced by the independent variable (ie, assess- 11.5%, respectively, whereas the incidence was ment tool), suggesting that use of the PAED only 3% during the retrospective period. The inci- scale in the implementation period was more dence of pediatric emergence delirium found by likely to identify patients experiencing emergence using the LOC-RASS during the retrospective delirium than use of the LOC-RASS in the retro- period was significantly lower than the incidence spective period. The comparison between the in- found by using the LOC-RASS (z ¼ e2.02; P ¼ cidence of emergence delirium found by using .043 ) and the PAED Scale (z ¼ e3.28; P < .001) the LOC-RASS during the implementation period during the implementation period. However, the and that found by using the LOC-RASS during incidence of pediatric emergence delirium was not the retrospective period resulted in an ES ¼ 0.53 significantly different between the PAED Scale and (OR ¼ 2.6 [95% CI, 1.672-10.556]). The com- LOC-RASS assessments during the implementation parison between the incidence of emergence de- period (z ¼ e1.02, P ¼ .31). It should be noted that lirium found by using the PAED Scale and that pediatric emergence delirium was identified in 12 found by using the LOC-RASS during the imple- patients with both the LOC-RASS (þ2) and the mentation period resulted in an ES ¼ 0.26 (OR ¼ PAED score ( 10) during the implementation 1.6 [95% CI, 0.810-3.171]). This suggests that period. Differences in assessment findings also use of the PAED scale may have influenced the

AORN Journal j 489 April 2014 Vol 99 No 4 STAMPER ET AL

TABLE 2. Logistic Regression Analysis of Patient Characteristics and Association With Pediatric Emergence Delirium

2 2 Characteristic c c 3 P Odds ratio 95% CI

Age 6 y 0.124 1 .724 1.173 0.483-0.847 ADHDa 2.021 1 .155 0.366 0.091-1.464 Premedication with midazolam 3.445 1 .063 0.431 0.177-1.048 Parental presence at induction 0.165 1 .685 0.821 0.318-2.214 Sevoﬂuraneb 0.645 1 .422 1.725 0.456-6.527 ENT surgery 0.286 1 .592 1.765 0.220-14.161

ADHD ¼ attention deficit hyperactivity disorder; CI ¼ confidence interval; ENT ¼ ear, nose, and throat. a Medically diagnosed with ADHD. b Anesthesia maintained with sevoflurane.

LOC-RASS scores in the implementation period more than or equal to 85%. Only the fidelity pro- because each tool was used simultaneously to cess measure (ie, provider response rate 85%) identify patients who experience emergence de- met the established criteria set for the fidelity lirium, resulting in a smaller ES between the tools measures. Attending anesthesiologists were notified compared during the implementation period. of 13 of the 15 pediatric emergence delirium cases In examination of the 23 patients with pediatric (86.7%) assessed by using the LOC-RASS and 16 emergence delirium identified by using the PAED of the 23 pediatric emergence delirium cases (69.6%) Scale, neither age younger than or equal to 6 years; assessed by using the PAED Scale. Attending an- premedication with midazolam; a medical diag- esthesiologists “responded” to all cases of which nosis of ADHD; parental presence at induction; they were notified, either by coming to the bedside, anesthesia maintained with sevoflurane; nor ear, ordering or administering a medication, or a com- nose, and throat procedures predicted the develop- bination of both, for an overall response rate ment of pediatric emergence delirium (Table 2). of 100%. Only one variable, premedication with midazolam, Thirteen of the 26 patients identified with an trended positively toward being associated with episode of pediatric emergence delirium by using developing pediatric emergence delirium within either assessment tool received medication; six 2 this population, (c 1 ¼ 3.445; P ¼ .063). (46.2%) were given fentanyl and seven (53.8%) We assessed the effectiveness of program im- were given dexmedetomidine. The decision to plementation through compliance with the assess- administer fentanyl versus dexmedetomidine in ments (ie, structure), provider response rate to response to emergence delirium was based on notification of pediatric emergence delirium per the provider preference because emergence delirium assessments (ie, process), and overall rate of pa- is self-limiting and will resolve with time. The tients experiencing subsequent or sustained emergence attending anesthesiologists were not asked to delirium after pharmacological intervention (ie, document their emergence delirium assessment; outcome). Two hundred of the 249 patients who therefore, we could not determine whether they presented to the children’s PACU during the im- chose to administer medication based on their plementation period received assessments during determination that the patients were experiencing all four time periods by using both the PAED Scale pain or emergence delirium. Five of these 13 pa- and LOC-RASS for an overall compliance rate tients experienced a subsequent or sustained ep- of 80.3%, which was less than the set standard of isode of pediatric emergence delirium in the next

490 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org assessment period. Therefore, the rate of subse- can be used to detect disturbances in several psy- quent or sustained pediatric emergence delirium chomotor characteristics simultaneously, which after pharmacological intervention (38.5%) was allows providers to assess both cognition and be- higher than fidelity outcome measure set standard havior at the same time. As a result, patients who of less than or equal to 15%. experience pediatric emergence delirium may be found to have a decreased level of consciousness DISCUSSION and exhibit agitated behavior by using the PAED The primary goal of this quality improvement Scale. As a result, it may be easier to recognize project was to evaluate the identification of pedi- pediatric emergence delirium because the tool atric emergence delirium by implementing and allows practitioners to better identify patients who evaluating the use of are demonstrating a the PAED Scale in the decreased level of children’s PACU at The PAED Scale can be used to detect consciousness, ex- our facility. The re- disturbances in several psychomotor char- hibiting agitated trospective chart re- acteristics simultaneously, which allows pro- behavior, or both. view revealed that viders to assess both cognition and behavior This is in contrast only 3% of patients at the same time. to the LOC-RASS experienced pediatric assessment charac- emergence delirium teristics, whereby during this time period when they were assessed by a patient cannot be identified as having both using only the LOC-RASS. Although this incidence a decreased level of consciousness and being rate was within the range reported in the literature agitated. Therefore, use of the LOC-RASS to (ie, 2% to 80%13), it was significantly lower than identify pediatric emergence delirium could lead the incidence found by using either the LOC- to false-negative assessments if a patient does not RASS or PAED Scale during the implementation make eye contact or interact with the provider and period (ie, 7.5% and 11.5%, respectively). receives a score between e5 and e1. The potential It could be argued that the incidence of pediatric for false-positive assessments does exist when using emergence delirium that we found between the two the PAED Scale; however, determination of the time periods was a result of heightened awareness clinical effect could not be ascertained from data of emergence delirium among anesthesia and collected during this project; therefore, additional nursing personnel during the implementation investigation is required. It should be noted that we period. However, analysis of our findings suggests based the threshold PAED score of more than or that the PAED Scale may be a more sensitive equal to 10 on data from the developers’ receiver assessment of pediatric emergence delirium after operating characteristic curve that maximized the emergence from general anesthesia. In addition to number of true positives and minimized the number the z scores and corresponding P values, the asso- of false positives under the area of the curve.12 ciated ESs suggested that the use of the PAED had We attributed the increased incidence found by a clinically significant effect on the identification using the LOC-RASS in the implementation period of pediatric emergence delirium during the imple- to two factors. The first was the education the PACU mentation period. We attributed this significant RNs received during the inservice program, which difference to the fact that, unlike the PAED Scale, may have led to a heightened awareness and re- the LOC-RASS was not designed to identify pedi- newed commitment to vigilantly assess for pediat- atric emergence delirium, nor has it been validated ric emergence delirium in patients at the children’s in the pediatric population.12,14,15 The PAED Scale PACU. The second factor was the concomitant use

AORN Journal j 491 April 2014 Vol 99 No 4 STAMPER ET AL of the PAED Scale, which may have influenced the sionals (ie, attending anesthesiologists, residents, RNs’ LOC-RASS assessments. However, 11 pa- CRNAs) to the pediatric perioperative setting tients who met criteria for experiencing pediatric may have contributed to increased use of mid- emergence delirium when using the PAED Scale azolam; however, the effect of this variable is did not meet the criteria when using the LOC- unknown because provider-specific practices were RASS. These findings suggested that, even when not tracked. Similarly, the type of maintenance a validated pediatric emergence delirium assess- anesthetic used was left to the discretion of the ment tool is used, the LOC-RASS led to under- anesthesia professional unless the use of volatile identification of emergence delirium among our agents was contraindicated during certain neuro- pediatric patients. logical procedures. Therefore, we could not de- A secondary goal of the project was to compare termine the exact source of the difference in patient characteristics among and within the two anesthetic maintenance between the two time pe- time periods, and to determine whether any of the riods but was likely because of provider preference. characteristics of the patients in the implementation None of the patient characteristics entered in- period were associated with the development of to the regression model were predictive of the emergence delirium. The only significant differ- development of pediatric emergence delirium ences that we noted during the imple- between the groups mentation period. during each time When the episodes of pediatric emergence This could be attrib- period pertained to delirium were reported, the overall response uted in part to the low administration of rate among the attending anesthesiologists incidence of pediatric midazolam in the was 100%, suggesting that the attending emergence delirium. anesthesiologists valued the PAED Scale preoperative area and Additionally, neither findings. thetypeofmainte- parental presence at nance anesthetic used induction nor a medi- during the surgical procedure, and neither variable cal diagnosis of ADHD have been shown to be was found to be a significant contributing factor to specifically associated with the development of the development of pediatric emergence delirium. pediatric emergence delirium; however, we included Midazolam trended toward a positive correlation, these measures in the analysis because they are but it was not associated with the development associated with separation anxiety and impulsive of emergence delirium. The use of midazolam is behavior, respectively, which have been identified provider-specific and not a standard of care in our as potential factors that increase the risk of emer- institution. There was no identifiable reason for the gence delirium development.16,23 The lack of var- significant increase in midazolam administration iability in maintenance anesthetic practices and the in the preoperative area during the implementa- low number of ear, nose, and throat procedures tion period. performed during the implementation period may Preoperative anxiety and fear are commonly have influenced the insignificant findings related experienced among children who present for to these patient characteristics. surgery.16 The decision to administer midazolam Our final goal was to assess the fidelity of the always has been left to the discretion of the an- PAED Scale during the implementation period esthesia professional and is based on a subjective as it related to the Quality of Care Framework assessment of the child’s level of preoperative measures (ie, structure, process, outcome).21 The anxiety. Introduction of new anesthesia profes- RNs completed LOC-RASS and PAED Scale

492 j AORN Journal PEDIATRIC EMERGENCE DELIRIUM www.aornjournal.org assessments of the majority of the patients (80.3%), CONCLUSION indicating adoption of the practice change. We Analysis of our findings suggests that the PAED attributed the lack of higher compliance to two Scale may be a more sensitive measure for pedi- factors. First, RNs documented their assessment atric emergence delirium within our children’s findings on the PAED Data Collection Form rather PACU than the LOC-RASS, and that anesthesia than as part of their standard documentation prac- and nursing professionals were capable of using the tices. The second factor was the high acuity and scale effectively to identify and respond to episodes patient workload that was associated with our pa- of emergence delirium. Future goals should be to tient population. Our children’s PACU provided institute the PAED Scale as standard of care when postoperative recovery services for patients with assessing for the presence of pediatric emergence a wide range of comorbidities, and the patient delirium, build on the current knowledge base, turnover rate was high at times, both of which discover ways to improve communication between may have contributed to some patients not being health care personnel, and help prevent patient and assessed according to protocol. personnel injury when episodes of pediatric emer- The attending anesthesiologists were very respon- gence delirium do occur. sive to episodes of pediatric emergence delirium Editor’s note: d when they were notified by the RNs. The reason SAS/STAT Statistical Analysis some RNs did not report episodes of emergence System is a registered trademark of Statistical delirium is unknown. It is possible that the RNs did Analysis System Institute, Inc, Cary, NC; 1982. not feel confident with their assessment findings, or, if medication orders were already in place, they References 1. Wilson TA, Graves SA. Pediatric considerations in a may have felt that they did not need to contact the general postanesthesia care unit. J Post Anesth Nurs. attending anesthesiologist despite being asked to 1990;5(1):16-24. do so as part of the project protocol. When the 2. LepouseC,LautnerCA,LiuL,GomisP,LeonA.Emer- gence delirium in adults in the post-anaesthesia care unit. episodes of pediatric emergence delirium were Br J Anaesth. 2006;96(6):747-753. reported, the overall response rate among the 3. Radtke FM, Franck M, Hagemann L, Seeling M, Wernecke KD, Spies CD. Risk factors for inadequate attending anesthesiologists was 100%, suggesting emergence after anesthesia: emergence delirium and that the attending anesthesiologists valued the hypoactive delirium. Minerva Anestesiol. 2010;76(6): 394-403. PAED Scale findings. 4. Voepel-Lewis T, Malviya S, Tait AR. A prospective Analysis of fidelity outcomes suggests that cohort study of emergence agitation in the pediatric medication administration was not a definitive postanesthesia care unit. Anesth Analg. 2003;96(6): 1625-1630. treatment for all cases of pediatric emergence 5. Aouda MT, Yazbeck-Karam VG, Nasr VG, El-Khatib MF, delirium. Medication administration effectively Kanazi GE, Bleik JH. A single dose of propofol at the end of surgery for the prevention of emergence agitation abated emergence delirium in eight of the 13 pain children undergoing strabismus surgery during sevo- tients. However, analysis of our findings suggests flurane anesthesia. Anesthesiology. 2007;107(5):733-738. 6. Abu-Shahwan I. Effect of propofol on emergence be- that it is important to weigh the risks and benefits of havior in children after sevoflurane general anesthesia. medication administration, including its effects on Paediatr Anaesth. 2008;18(1):55-59. respiratory, cardiovascular, and neurological func- 7. Bong CL, Ng AS. Evaluation of emergence delirium in Asian children using the Pediatric Anesthesia Emergence tion, which may prolong the length of stay in the Delirium Scale. Paediatr Anaesth. 2009;19(6):593-600. PACU. In addition, interventions to keep the pa- 8. Hasani A, Ozgen S, Baftiu N. Emergence agitation in children after propofol versus halothane anesthesia. Med tient safe until the episode of emergence delirium Sci Monit. 2009;15(6):CR302-CR306. subsides should be used, such as securing IV 9. Faulk DJ, Twite MD, Zuk J, Pan Z, Wallen B, Friesen RH. Hypnotic depth and the incidence of emergence agitation catheters to prevent removal; padding bed siderails; and negative postoperative behavioral changes. Paediatr and maintaining a calm, quiet environment. Anaesth. 2010;20(1):72-81.

AORN Journal j 493 April 2014 Vol 99 No 4 STAMPER ET AL

10. Vlajkovic GP, Sindjelic RP. Emergence delirium in children: many questions, few answers. Anesth Analg. Matthew J. Stamper, DNP, CRNA, is certified 2007;104(1):84-91. RN anesthetist at Duke University Medical 11. Hudek K. Emergence delirium: a nursing perspective. AORN J. 2009;89(3):509-516. Center, Durham, NC. Dr Stamper has no de- 12. Sikich N, Lerman J. Development and psychometric clared affiliation that could be perceived as evaluation of the pediatric anesthesia emergence delirium scale. Anesthesiology. 2004;100(5):1138-1145. posing a potential conflict of interest in the 13. Dahmani S, Stany I, Brasher C, et al. Pharmacological publication of this article. prevention of sevoflurane- and desflurane-related emergence agitation in children: a meta-analysis of published Sharon J. Hawks, DNP, CRNA, is an assistant studies. Br J Anaesth. 2010;104(2):216-223. professor and the director of the Duke University 14. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult School of Nursing, Nurse Anesthesia Program, intensive care units. Am J Resp Crit Care Med. 2002; Durham, NC. Dr Hawks has no declared affili- 166(10):1338-1344. 15. Schieveld JN, van der Valk JA, Smeets I, et al. Diag- ation that could be perceived as posing a po- nostic considerations regarding pediatric delirium: a re- tential conflict of interest in the publication of view and a proposal for an algorithm for pediatric this article. intensive care units. Intensive Care Med. 2009;35(11): 1843-1849. Brad M. Taicher, DO,MBA,isanassistant 16. Kain ZN, Caldwell-Andrews AA, Maranets I, et al. Pre- operative anxiety and emergence delirium and post- professor of anesthesiology and an assistant operative maladaptive behaviors. Anesth Analg. 2004; professor in pediatrics at Duke University 99(6):1648-1654. 17. Kuratani N, Oi Y. Greater incidence of emergence Medical Center, Durham, NC. Dr Taicher has no agitation in children after sevoflurane anesthesia as declared affiliation that could be perceived as compared with halothane: a meta-analysis of randomized posing a potential conflict of interest in the controlled trials. Anesthesiology. 2009;109(2):225-232. 18. Cravero JP, Beach M, Thyr B, Whalen K. The effect of publication of this article. small dose fentanyl on the emergence characteristics of pediatric patients after sevoflurane anesthesia without Juliet Bonta, BSN, RN, is a staff nurse in Duke surgery. Anesth Analg. 2003;97(2):364-367. Pediatric Perioperative Services at Duke Uni- 19. American Psychiatric Association. Diagnostic and Sta- tistical Manual of Mental Disorders. 4th ed. Text Revi- versity Medical Center, Wilson, NC. Ms Bonta sion. (DSM-IV-TR) Arlington, TX: American Psychiatric has no declared affiliation that could be per- Association; 2000. ceived as posing a potential conflict of interest in 20. Bajwa SA, Costi D, Cyna AM. A comparison of emergence delirium scales following general anesthesia in the publication of this article. children. Paediatr Anaesth. 2010;20(8):704-711. 21. Donabedian A. The quality of care. How can it be as- Debra H. Brandon, PhD, RN, CCNS, FAAN, sessed? JAMA. 1988;260(12):1743-1748. is an associate professor and the director of the 22. Anesthesia care team. American Society of Anesthesi- ologists. https://www.asahq.org/Lifeline/Who%20Is% Duke University School of Nursing, PhD, and 20An%20Anesthesiologist/Anesthesia%20Care%20Team Postdoctoral Programs, Durham, NC. Dr Bran- .aspx. Accessed January 23, 2014. don has no declared affiliation that could be 23. Mountain BW, Smithson L, Cramolini M, Wyatt TH, Newman M. Dexmedetomidine as a pediatric anesthetic perceived as posing a potential conflict of in- premedication to reduce anxiety and to deter emergence terest in the publication of this article. delirium. AANA J. 2011;79(3):219-224.

494 j AORN Journal Copyright of AORN Journal is the property of Elsevier Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.