Does Teaching Crisis Resource Management Skills Improve Resuscitation Performance in Pediatric Residents?* Jaime Blackwood, MD1; Jonathan P. Duff, MD2; Alberto Nettel-Aguirre, PhD3; Dennis Djogovic, MD4; Chloe Joynt, MD5
Objective: The effect of teaching crisis resource management skills on the resuscitation performance of pediatric residents is unknown. The primary objective of this pilot study was to determine if teaching crisis resource management to residents leads to improved clinical and crisis resource management performance in simulated pediatric resuscitation scenarios. Design: A prospective, randomized control pilot study. Setting: Simulation facility at tertiary pediatric hospital. Subjects: Junior pediatric residents. Interventions: Junior pediatric residents were randomized to 1 hour of crisis resource management instruction or no additional training. Measurements and Main Results: Time to predetermined resuscitation tasks was noted in simulated resuscitation scenarios
*See also p. 382. 1 Division of Critical Care, Department of Pediatrics, Alberta Children’s Hospital, University of Calgary, Calgary, AB, Canada. 2 Division of Critical Care, Department of Pediatrics, Stollery Children’s Hospital, University of Alberta, Edmonton, AB, Canada. 3 Departments of Pediatrics and Community Health Sciences, University of Calgary, Alberta Children’s Hospital Research Institute for Maternal and Child Health, Calgary, AB, Canada. 4 Division of Critical Care, University of Alberta Hospital, University of Alberta, Edmonton, AB, Canada. 5 Division of Neonatology, Department of Pediatrics, Stollery Children’s Hospital, University of Alberta, Edmonton, AB, Canada. This study was performed at Stollery Children’s Hospital, University of Alberta, Edmonton, Alberta, Canada. Supported, in part, by the Women and Children’s Health Research Institute Grant (Edmonton, AB, Canada) and the Royal College of Physicians and Surgeons Medical Education Grant. Dr. Duff received a research grant (as co-applicant, site investigator of Stollery Children’s Hospital) from the Heart and Stroke Foundation of Canada, grant number HSFC PG-10-0482. Dr. Nettel-Aguirre served as a board member for CREBA. Dr. Joynt received grant support from the Women and Children's Health Research Institute. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2014 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0000000000000100
e168
www.pccmjournal.org
immediately after intervention and again 3 months post intervention. Crisis resource management skills were evaluated using the Ottawa Global Rating Scale. Fifteen junior residents participated in the study, of which seven in the intervention group. The intervention crisis resource management group placed monitor leads 24.6 seconds earlier (p = 0.02), placed an IV 47.1 seconds sooner (p = 0.04), called for help 50.4 seconds faster (p = 0.03), and checked for a pulse after noticing a rhythm change 84.9 seconds quicker (p = 0.01). There was no statistically significant difference in time to initiation of cardiopulmonary resuscitation (p = 0.264). The intervention group had overall crisis resource management performance scores 1.15 points higher (Ottawa Global Rating Scale [out of 7]) (p = 0.02). Three months later, these differences between the groups persisted. Conclusions: A 1-hour crisis resource management teaching session improved time to critical initial steps of pediatric resuscitation and crisis resource management performance as measured by the Ottawa Global Rating Scale. The control group did not develop these crisis resource management skills over 3 months of standard training indicating that obtaining these skills requires specific education. Larger studies of crisis resource education are required. (Pediatr Crit Care Med 2014; 15:e168–e174) Key Words: cardiopulmonary resuscitation; crisis resource management; pediatrics; Pediatric Advanced Life Support
E
ducating pediatricians-in-training about resuscitation of acutely ill children presents many barriers. Guidelines and algorithms exist for the resuscitation of children during cardiopulmonary resuscitation (CPR) (1). As resuscitation events in children are rare, it is a challenge to educate pediatric residents in the skills required to manage these events, yet appropriately managing resuscitation is a core competency required by all pediatric trainees (2). It is becoming increasingly recognized that truly successful management of critically ill patients requires strong technical and cognitive skills, as well as complementary strong nontechnical skills (3, 4). May 2014 • Volume 15 • Number 4
Online Clinical Investigations
Crisis resource management (CRM) refers to the nontechnical skills, including leadership, problem solving, situational awareness, communication skills, and resource management (5–7). Recent studies have demonstrated that even brief team building or leadership instruction can improve a resuscitation team’s performance (8, 9). In a pediatric study, CRM teaching led to an increase in perceived team collaboration, satisfaction with care, and observed teamwork skills (10); however, there is a paucity of literature measuring CRM’s effect on medical outcome or changes in ability to perform medical resuscitation maneuvers. These nontechnical skills, although important in the management of acutely ill children, are not acquired passively during residency and hence may require specific education (7, 11). Simulation has been demonstrated to be a useful approach to teach CRM skills (12, 13). To better prepare pediatric residents for caring for acutely ill children, there may be benefit in providing instruction in CRM. We designed this pilot study to determine if a brief small group CRM training session would improve resuscitation performance of junior residents. Our primary objective was to determine if teaching CRM skills leads to improvement in the delivery of pediatric advanced life support (PALS) by junior pediatric residents, measured by a faster initiation of important resuscitation maneuvers. We also aimed to determine if a brief CRM teaching session would improve residents’ CRM skills and finally if teaching CRM had a long-term effect on resuscitation and teamwork behaviors.
MATERIAL AND METHODS Study Population First- and second-year pediatric residents at the University of Alberta were invited to voluntarily participate in this study. Exclusion criteria consisted of extensive acute care experience as defined as more than 1 month in an ICU setting, previous CRM training, or lack of PALS certification. No first- or second-year pediatric residents met exclusion criteria. Written informed consent was obtained from all participants. A prestudy questionnaire determined residents’ level of training, gender, and previous intensive care experience. This information was then used to stratify residents for level of training and critical care experience. Strata were placed in an opaque envelope, and names were randomly drawn to alternately be assigned to CRM and control group from each strata by an assistant not involved in the study. The residents remained in these groups for the duration of the study. The Health Research Ethics Board at the University of Alberta approved this study. Intervention On the day of the research scenario, all participants took part in a 1-hour didactic refresher session, reviewing the knowledge, technical skills, and algorithms important in PALS. Directly after the PALS refresher session, the intervention group had a 1-hour small group CRM training session. This Pediatric Critical Care Medicine
session included didactic teaching, videos of CRM skills, and hands-on interactive practice utilizing CRM skills in an unrelated resuscitation scenario. The format was based on the recent PALS curriculum and the Acute Critical Event Simulation Course (14). Following this session, the group was asked to keep the CRM teaching and intervention content confidential. The control group had no further instruction beyond the PALS refresher session. A standardized orientation to the pediatric high-fidelity patient simulators (BabySim ECS and PediSim ECS; formally METI, now CAE, Saint-Laurent, Quebec, Canada) and crash carts took place for all participants. The assessment scenarios were conducted immediately following the PALS refresher and CRM teaching. Scenarios Each resident assumed the role of “team leader” for two standardized simulated pediatric medical emergency scenarios taken from validated PALS algorithm sequences. The two scenarios were identical for all residents. Scenario 1 was a young child that presented with sinus tachycardia and deteriorated into pulseless ventricular tachycardia necessitating basic life support and defibrillation. Scenario 2 was an infant with respiratory distress and sinus tachycardia deteriorating into sinus bradycardia necessitating basic life support and epinephrine. The two scenarios had common expected initial interventions (including airway assessment and intervention, monitor application, recognition of a critical rhythm change, and CPR) and interventions specific to the algorithms. A brief debriefing session was conducted after each scenario in which any major medical errors were discussed. No teaching around CRM principles occurred in the debriefing. Pediatric intensivists at Stollery Children’s Hospital reviewed the scenarios for content and realism. Each resident team leader was accompanied by a standardized resuscitation team consisting of a respiratory therapist and two registered PICU nurses from the Stollery Children’s Hospital who provided assistance within their scope of practice. The team members’ behavior during each scenario was practiced previously and scripted to ensure identical behavior exposure for each team leader. Real-time clarification was provided as needed via earpiece from the standardized simulator observer familiar with the scenarios. Team members only performed maneuvers when instructed to by the leader and were scripted to prompt the team leader about changes in the patient’s clinical condition in a standardized manner. Team leaders were instructed to involve the team members (respiratory therapist and RNs) as they would in a true scenario. All scenarios were videotaped for analysis with the consent of the participants. Three months after the initial scenario, the participants participated in two more simulated scenarios that were identical to the previous scenarios except for clinical history, age of the patient, and simulator size. The resuscitation team and simulation observer remained the same. www.pccmjournal.org
e169
Blackwood et al
Assessment Using the video recordings, timing to a predetermined endpoint of a priori–specified medical interventions from fixed times “zero” was determined by one of the investigators (J.B.). The initial “time zero” commenced from the end of the scenario presentation by the nurse and was used to determine times to expected basic life support maneuvers. The second “time zero” mark was 2 minutes later when the heart rhythm converted from sinus to a critical pathologic rhythm and was used to time predetermined medical interventions required for that specific PALS algorithm. Interventions were chosen that were common in sequence in the two algorithms. If a basic life support intervention did not occur prior to pathological rhythm change (second time zero), then “2 minutes” was allotted as the intervention time and “a critical error” was marked. Deviations from the PALS algorithms (wrong algorithm and medication or interventions not performed or performed out of order) were also noted. Prior to reviewing the study scenarios, the three raters standardized their understanding of the Ottawa Global Rating Scale (OGRS) descriptive anchors. The raters then independently scored previously recorded scenarios (not related to the study) demonstrating below average, average, and above average team performance and then debriefed until rater scores were consistent for practice scenarios. The raters reviewed all of the recordings, blinded to which arm of the study the participants were part of, and independently scored the resident’s nontechnical performance using the validated OGRS for crisis management skills (15). The OGRS allows the rater to rate the overall performance from 1 to 7 and also rates the subject’s performance in five other domains (leadership, problem solving, situational awareness, resource utilization, and communication) from 1 to 7 for an additional 35 points. The groups’ CRM Score Global Rating as well at the additive score of all six OGRS categories (out of 42) was analyzed. The ratings given by the raters were averaged, and in order to assess the agreement of these, we determined the intraclass correlation coefficient (ICC). The ICC was calculated for each of the four scenario-time combinations (ICC range, 0.74–0.92; p < 000.1). These values showed very good to excellent agreement between raters, and hence, averaging the ratings as is usual practice made sense. Statistics All of the pediatric residents that were available and consented were included in the study. A convenience sample of 15 first- and second-year residents was enrolled. Descriptives per scenario/ date/group are reported as medians with interquartile ranges for nonsymmetrically distributed data and means with sds for normally/symmetric distributed data. Demographic data, which were all categorical-type data, were assessed for differences via Agresti-Coull-Caffo 95% CIs for difference in proportions. A linear mixed effects model with fixed effects for intervention and date and their interaction plus random effect for subjects was used to detect differences/effects on time to maneuver due to each intervention and date. This technique models both the intervention and the time effect via models that assume an “on average” effect. As all of the scenarios e170
www.pccmjournal.org
required the same basic life support interventions, scenarios were not considered to be differential with respect to the outcome being measured, and hence, it was not included as a variable of effect in the model. Data were entered into Microsoft Excel, then cleaned, and analyzed with R: A Language and Environment for Statistical Computing program (http://www.r-project.org/) (16) and using the “nlme” package (http://cran.r-project.org/web/packages/ nlme/index.html) (17). Comparisons of the mean times to maneuvers for control and CRM groups were carried out considering the potential period/intervention statistical interaction (effect modification) and the repeated measures nature of the data. Mean times were used in the model as no method exists for using median data. Significance of the effects of intervention and date was determined using linear mixed effects models: one for times to maneuvers and another for CRM scores. Crude descriptive data per scenario, arm, and date are reported as medians with interquartile ranges. The level of significance used was set at 0.05.
RESULTS Participants All first- and second-year pediatric residents in the University of Alberta Pediatric residency training program (n = 15) participated in the study. Seven were randomized to the intervention group and eight to the control group. All residents attended both the initial and follow-up sessions. Please see Appendix 1 for the flow diagram outlining the events involved for the participants. There were no differences in gender, level of training, previous ICU training, or PALS certification between groups (Table 1). As the study period occurred early in the academic year cycle, no regularly scheduled PALS recertification or resident promotion between years of training occurred. Time to Interventions Using a linear mixed effects model, there were significant differences found between the CRM and control groups for several of the predetermined medical interventions (Table 2). Table 1. Demographic Information About the Participants Intervention Group (n = 7) (%)
Control Group (n = 8) (%)
Female
7 (100)
7 (88)
Postgraduate year 2
3 (43)
4 (50)
Previous ICU experience
5 (71)
4 (50)
Mock code leader experience
2 (29)
5 (63)
Pediatric advanced life support certified
7 (100)
8 (100)
May 2014 • Volume 15 • Number 4
Online Clinical Investigations
Table 2.
Time to Predetermined Medical Interventions Initial Scenarios
Time to Medical Intervention, Median (s) (IQR)
CRM
Control
Follow-Up Scenarios CRM
Control
Model Effect (95% CI)
p
Assessment of airway Bradycardia
17.4 (15.2–78.3)
24.0 (14.6–133.8)
125.1 (28.4–269.8)
68.4 (44.4–265.2)
VT
39.2 (33.0–47.7)
35.5 (16.8–214.8)
83.8 (30.3–153.1)
139.4 (24.1–243.8)
Bradycardia
78.3 (66.8–102.0)
84.9 (71.1–129.8)
75.0 (65.7–96.2)
134.2 (122.2–152.8)
VT
172.4 (131.4–197.6)
160.5 (126.3–219.2)
52.7 (45.8–95.0)
97.4 (48.3–103.3)
Bradycardia
109.5 (60.2–137.0)
83.7 (76.6–116.6)
65.1 (20.7–96.3)
60.5 (37.8–137.2)
VT
58.8 (51.0–90.8)
35.6 (27.1–48.6)
129.6 (81.3–177.5)
82.8 (51.4–91.3)
Bradycardia
73.5 (62.0–76.5)
88.2 (54.9–118.6)
60.6 (52.0–71.4)
82.5 (67.4–125.8)
VT
70.4 (60.0–79.2)
95.8 (67.1–124.0)
56.5 (53.4–78.5)
95.3 (81.2–108.2)
Bradycardia
155.7 (137.8–184.6)
216.3 (179.7–271.4)
100.5 (79.5–132.0)
93.8 (57.3–123.4)
VT
111.2 (98.5–172.4)
160.0 (117.0–186.4)
128.6 (82.1–146.4)
189.5 (147.0–334.0)
16.2 (7.4–69.5)
185.7 (101.1–256.0)
18.3 (12.0–42.0)
20.5 (10.7–113.7)
36.0 (23.3–141.6)
58.2 (35.4–101.8)
18.5 (11.1–25.7)
46.0 (29.0–113.2)
CRM 26.7 s faster (110 s slower to 56.7 s faster)
0.09
CRM 23.3 s faster (56.7 s slower to 10.2 s faster)
0.06
CRM 8 s slower (70 s slower to 50 s faster)
0.77
CRM 24.6 s faster (3.5 s faster to 45.8 s faster)
0.03
CRM 47.1 s faster (0.8 s faster to 93.4 s faster)
0.04
CRM 50.4 s faster (6.5 s faster to 94.4 s faster)
0.03
CRM 84.9 s faster (30.4 s faster to 139.3 s faster)
0.01
CRM 27 s faster (19 s slower to 73 s faster)
0.23
CRM 14 s faster (12 s slower to 41 s faster)
0.26
Place oxygen on
Initial pulse check
Place monitors on
Placing an IV
Calling for help Bradycardia VT
Pulse check after rhythm change Bradycardia
18.6 (11.3–50.0)
59.4 (43.2–352.0)
21.2 (4.4–76.0)
343.4 (98.6–352.0)
VT
31.5 (11.7–46.9)
31.0 (17.8–74.3)
13.8 (11.4–39.5)
35.4 (9.2–76.7)
Start bag mask ventilation after apnea Bradycardia
8.4 (3.6–12.9)
6.9 (0.1–63.6)
24.3 (13.3–41.7)
42.6 (30.8–204.4)
VT
40.5 (29.8–56.0)
35.9 (16.3–53.2)
11.4 (9.5–30.1)
7.5 (5.4–50.4)
Cardiopulmonary resuscitation after pulseless Bradycardia
13.5 (10.1–24.6)
17.1 (13.1–56.9)
11.0 (8.6–26.1)
32.4 (10.5–57.7)
VT
60.6 (23.2–89.5)
30.2 (13.8–52.8)
27.3 (20.1–49.2)
19.8 (11.0–81.6)
IQR = interquartile range, CRM = crisis resource management, VT = ventricular tachycardia.
Pediatric Critical Care Medicine
www.pccmjournal.org
e171
Blackwood et al
The intervention CRM group placed monitor leads 24.6 seconds earlier (95% CI, 3.5–45.8; p = 0.02), placed an IV 47.1 seconds sooner (95% CI, 0.8–93.4; p = 0.04), and called for help 50.4 seconds faster (95% CI, 6.5–94.4; p = 0.03). Finally, the intervention group checked for a pulse after noticing a rhythm change 84.9 seconds quicker (95% CI, 30.4 s faster to 139.3 s faster; p = 0.01). There was no statistically significant difference found for initial pulse check prior to pathologic rhythm and starting bag valve mask ventilation (BMV) or CPR after the patient went pulseless. Data were also analyzed for effect modification of the time period on intervention group and none was found. Thus, when the groups were tested 3 months later, the ability for the CRM group to perform the same skills faster persisted with no improvement in the control group over time. In addition, at the 3-month retest, there was an indication of faster CPR initiation for the CRM group compared with their initial times (Table 2). CRM Performance Using the linear mixed effects model, the CRM group had an “overall OGRS” score (out of 7) that was 1.15 points (95% CI, 0.2–2.1, p = 0.02) higher than that in the control group, and this increase in overall CRM score was maintained at the 3-month retest scenario (Table 3). The summative score of all seven categories (out of 42) was 6.7 points (1.6–11.8, p = 0.01) higher in the CRM group, and this difference was not affected by retest 3 months later.
DISCUSSION This study investigated the effect of teaching CRM on the resuscitation performance of junior pediatric residents in high-fidelity simulation scenarios. We found that those that were taught CRM performed certain skills and medical interventions required during a simulated resuscitation quicker than a similarly experienced control group. In addition, the intervention group had better CRM skills, as measured by the OGRS, when compared with the control group. In our study, initial steps of the resuscitation were the skills that appeared to be most improved by teaching CRM. Residents, after only a brief CRM instruction session, called for help, placed monitors, and checked a pulse faster than our Table 3.
control group. A study by Hunt et al (18) revealed significant delays in the initiation of important resuscitation maneuvers in simulated pediatric resuscitation scenarios. They concluded that there should be focus on improving the quality of care delivered in the first 5 minutes of resuscitation efforts. As demonstrated in our study, teaching CRM may be an important component of this. It is possible that the initial organization of the team with good communication, leadership, and role assignment is critical for timely performance of these activities. This is consistent with a study by Jankouskas et al (19) in which medical and nursing students were randomized to CRM and basic life support (BLS) training or BLS training alone. In that study, the addition of CRM training had no effect to time to BLS (including b ag-mask-ventilation and chest compressions). However, in that study, the teams continued to practice their BLS skills as a team and both groups demonstrated improvement in BLS skills at the end of the training. In our study, our team members were scripted to not start CPR unless specifically directed to by the team leader. It is possible that with CRM training, team members will feel more comfortable making suggestions and potentially improve time to CPR initiation. Additional and repeated simulation training and debriefing of CRM in concert with simple and complex acute care pediatric scenarios with the in situ code team may be necessary to see improved team effectiveness in the more complex PALS algorithm above and beyond BLS skills. This study demonstrates improvement with only 1 hour of specific CRM teaching. More dramatic and lasting effects may occur with a program of regular CRM teaching and feedback. Recent reviews have outlined the importance of teamwork, effective communication, and leadership behavior in managing emergency situations (7, 20). Such studies have lead to the inclusion of CRM into PALS teaching. Our study did show some benefits of CRM training with initial assessment and management in resuscitation. This benefit was maintained at a 3-month assessment with no improvement in the control group. There is no indication that leads us to believe that despite the importance of CRM behaviors in resuscitation, residents are developing these skills over the
Summary of Ottawa Global Rating Scale Ratings Initial Scenarios
OGRS Score, Mean (sd)
CRM
Control
Follow-Up Scenarios CRM
Control
Model Effect (95% CI)
p
Overall performance (out of 7) (sd) Bradycardia
4.7 (0.9)
2.9 (1.0)
4.3 (1.1)
3.4 (1.0)
VT
4.4 (1.1)
3.8 (0.8)
4.5 (1.1)
3.3 (1.1)
CRM OGRS 1.15 points higher (0.2–2.1)
0.02
CRM OGRS 6.7 points higher (1.6–11.8)
0.01
Sum of all categories (out of 42) (sd) Bradycardia
28.6 (4.6)
18 (5.8)
26.8 (6.4)
21.2 (5.4)
VT
27.3 (5.0)
23.3 (4.4)
27.4 (6.6)
20.7 (6.6)
OGRS = Ottawa Global Rating Scale, CRM = crisis resource management, VT = ventricular tachycardia.
e172
www.pccmjournal.org
May 2014 • Volume 15 • Number 4
Online Clinical Investigations
course of their routine clinical training. We advocate that it is important to incorporate CRM into pediatric training curriculum (7, 11). The time to initiation of CPR, a marker that is known to correlate with outcome (21), was not statistically different between the groups, and it took a protracted amount of time for the leader to ask the team to initiate CPR. Our scenarios were designed such that CPR could not be initiated until asked for by the team leader. In a clinical event, CPR may be recommended to the leader by any of the team in attendance. Previous studies have demonstrated that nontechnical skills are important in the quality of CPR (8, 22, 23). In our study, the nontechnical performance of the CRM participants was improved with better communication and team dynamics. With these improvements in team dynamics, the other team members may feel comfortable suggesting CPR initiation in actual clinical scenarios and potentially improve CPR times. Although not significant, there was a trend toward an earlier start to CPR in the CRM group; perhaps with a larger study and more intense CRM training, a significant difference may be found. It should be noted that there were a number of deviations in PALS algorithms demonstrated by participants in our study, despite all being PALS certified and participating in a 1-hour PALS refresher. Deviations included failing to BMV, failing to defibrillate, or not checking for a pulse after a noted rhythm change. This is consistent with other studies (24). A recent study of second-year pediatric residents noted a rapid decline in their resuscitation skills following a PALS course (25). Limitations This study was limited in that it was a small single-center pilot study. However, even with such small numbers, significant differences in some resuscitation teamwork behaviors were noted. A larger trial is required to confirm these findings. Although the CRM raters were blinded to participant group allocation, the investigator measuring time to critical interventions was not. However, as this is objective data (time in seconds) being recorded by the same reviewer for all scenarios, the effect of differential bias is low. It must be acknowledged that our study did not assess participants at baseline and this may contribute to some of the differences noted though randomization should minimize that bias. Assessments of participants prior to and following CRM teaching would be beneficial to undertake with any further studies. The intervention arm received an extra hour of CRM training and that extra time in training may explain some of the improvements in medical maneuvers we noted. This study focused on the team leader as all of the team members’ responses were scripted. This study was a pilot study in which the leader, and not all members of the team, was instructed on CRM but did not require mastery of the skill. Larger studies in which these considerations are accounted for are warranted to investigate the effect on resuscitation measures. Pediatric Critical Care Medicine
CONCLUSIONS Pediatric resuscitation is a skill that is necessary for all pediatric medicine practitioners, yet it is difficult to both teach and experience due to the relatively rare number of events. Simulation has been shown to assist in learning these skills. CRM can help pediatric resident team leaders recognize and assess sick patients, call for appropriate help sooner, place IVs, and monitoring sooner. Our results suggest that CRM is not a skill set that is passively developed and may require specific education and feedback. This pilot study suggests that larger well-designed studies of crisis resources education should be undertaken.
REFERENCES
1. Kleinman ME, Chameides L, Schexnayder SM, et al; American Heart Association: 2010 American Heart Association (AHA) guidelines for cardiopulmonary resuscitation (CPR) and emergency cardiovascular care (ECC) of pediatric and neonatal patients: Pediatric advanced life support. Pediatrics 2010; 126:e1361–1399 2. Eppich WJ, Adler MD, McGaghie WC: Emergency and critical care pediatrics: Use of medical simulation for training in acute pediatric emergencies. Curr Opin Pediatr 2006; 18:266–271 3. Fletcher GCL, McGeorge P, Flin RH, et al: The role of non-technical skills in anaesthesia: A review of current literature. Br J Anaeseth 2002; 88:418–429 4. Gaba DM, Fish KJ, Howard SK: Chapter 1-Theory of Dynamic Decision Making and Crisis Management. In: Crisis Management in Anesthesia. New York, Churchill Livingstone, 1994, pp 5–29 5. Glavin RJ, Maran NJ: Integrating human factors into the medical curriculum. Med Educ 2003; 37(Suppl 1):59–64 6. Eppich WJ, Brannen M, Hunt EA: Team training: Implications for emergency and critical care pediatrics. Curr Opin Pediatr 2008; 20:255–260 7. Cheng A, Donoghue A, Gilfoyle E, et al: Simulation-based crisis resource management training for pediatric critical care medicine: A review for instructors. Pediatr Crit Care Med 2012; 13: 197–203 8. Hunziker S, Tschan F, Semmer NK, et al: Hands-on time during cardiopulmonary resuscitation is affected by the process of teambuilding: A prospective randomized simulator-based trial. BMC Emergency Medicine 2009; 9:3 9. Hunziker S, Bühlmann C, Tschan F, et al: Brief leadership instructions improve cardiopulmonary resuscitation in a high-fidelity simulation: A randomized controlled trial. Crit Care Med 2010; 38: 1086–1091 10. Jankouskas T, Bush MC, Murray B, et al: Crisis resource management: Evaluating outcomes of a multidisciplinary team. Simul Healthc 2007; 2:96–101 11. Gilfoyle E, Gottesman R, Razack S: Development of a leadership skills workshop in paediatric advanced resuscitation. Med Teach 2007; 29:e276–e283 12. Yee B, Naik VN, Joo HS, et al: Nontechnical skills in anesthesia crisis management with repeated exposure to simulation-based education. Anesthesiology 2005; 103:241–248 13. Reznek M, Smith-Coggins R, Howard S, et al: Emergency medicine crisis resource management (EMCRM): Pilot study of a simulationbased crisis management course for emergency medicine. Acad Emerg Med 2003; 10:386–389 14. Neilipovitz DT: Acute Resuscitation and Crisis Management. Ottawa, University of Ottawa Press, 2005 15. Kim J, Neilipovitz D, Cardinal P, et al: A pilot study using high-fidelity simulation to formally evaluate performance in the resuscitation of critically ill patients: The University of Ottawa Critical Care Medicine, High-Fidelity Simulation, and Crisis Resource Management I Study. Crit Care Med 2006; 34:2167–2174 www.pccmjournal.org
e173
Blackwood et al 16. R Development Core Team: R: A Language and Environment for Statistical Computing. Vienna, Austria, R Foundation for Statistical Computing, 2011 17. Pinheiro J, Bates D, DebRoy S, et al; the R Development Core Team: nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1–102, 2011 18. Hunt EA, Walker AR, Shaffner DH, et al: Simulation of in-hospital pediatric medical emergencies and cardiopulmonary arrests: Highlighting the importance of the first 5 minutes. Pediatrics 2008; 121: e34–e43 19. Jankouskas TS, Haidet KK, Hupcey JE, et al: Targeted crisis resource management training improves performance among randomized nursing and medical students. Simul Healthc 2011; 6:316–326 20. Hunziker S, Johansson AC, Tschan F, et al: Teamwork and leadership in cardiopulmonary resuscitation. J Am Coll Cardiol 2011; 57:2381–2388 21. Ali B, Zafari AM: Narrative review: Cardiopulmonary resuscitation and emergency cardiovascular care: Review of the current guidelines. Ann Intern Med 2007; 147:171–179 22. Thomas EJ, Sexton JB, Lasky RE, et al: Teamwork and quality during neonatal care in the delivery room. J Perinatol 2006; 26:163–169 23. Marsch SC, Müller C, Marquardt K, et al: Human factors affect the quality of cardiopulmonary resuscitation in simulated cardiac arrests. Resuscitation 2004; 60:51–56 24. Hunt EA, Vera K, Diener-West M, et al: Delays and errors in cardiopulmonary resuscitation and defibrillation by pediatric residents during simulated cardiopulmonary arrests. Resuscitation 2009; 80:819–825 25. Roy KM, Miller MP, Schmidt K, et al: Pediatric residents experience a significant decline in their response capabilities to simulated life-threatening events as their training frequency in cardiopulmonary resuscitation decreases. Pediatr Crit Care Med 2011; 12:e141–e144
e174
www.pccmjournal.org
Flow diagram outlining the events involved for the participants of the study.
Appendix 1.
PALS = pediatric advanced life support, CRM = crisis resource management.
May 2014 • Volume 15 • Number 4
Objective: The effect of teaching crisis resource management skills on the resuscitation performance of pediatric residents is unknown. The primary objective of this pilot study was to determine if teaching crisis resource management to residents leads to improved clinical and crisis resource management performance in simulated pediatric resuscitation scenarios. Design: A prospective, randomized control pilot study. Setting: Simulation facility at tertiary pediatric hospital. Subjects: Junior pediatric residents. Interventions: Junior pediatric residents were randomized to 1 hour of crisis resource management instruction or no additional training. Measurements and Main Results: Time to predetermined resuscitation tasks was noted in simulated resuscitation scenarios
*See also p. 382. 1 Division of Critical Care, Department of Pediatrics, Alberta Children’s Hospital, University of Calgary, Calgary, AB, Canada. 2 Division of Critical Care, Department of Pediatrics, Stollery Children’s Hospital, University of Alberta, Edmonton, AB, Canada. 3 Departments of Pediatrics and Community Health Sciences, University of Calgary, Alberta Children’s Hospital Research Institute for Maternal and Child Health, Calgary, AB, Canada. 4 Division of Critical Care, University of Alberta Hospital, University of Alberta, Edmonton, AB, Canada. 5 Division of Neonatology, Department of Pediatrics, Stollery Children’s Hospital, University of Alberta, Edmonton, AB, Canada. This study was performed at Stollery Children’s Hospital, University of Alberta, Edmonton, Alberta, Canada. Supported, in part, by the Women and Children’s Health Research Institute Grant (Edmonton, AB, Canada) and the Royal College of Physicians and Surgeons Medical Education Grant. Dr. Duff received a research grant (as co-applicant, site investigator of Stollery Children’s Hospital) from the Heart and Stroke Foundation of Canada, grant number HSFC PG-10-0482. Dr. Nettel-Aguirre served as a board member for CREBA. Dr. Joynt received grant support from the Women and Children's Health Research Institute. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2014 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0000000000000100
e168
www.pccmjournal.org
immediately after intervention and again 3 months post intervention. Crisis resource management skills were evaluated using the Ottawa Global Rating Scale. Fifteen junior residents participated in the study, of which seven in the intervention group. The intervention crisis resource management group placed monitor leads 24.6 seconds earlier (p = 0.02), placed an IV 47.1 seconds sooner (p = 0.04), called for help 50.4 seconds faster (p = 0.03), and checked for a pulse after noticing a rhythm change 84.9 seconds quicker (p = 0.01). There was no statistically significant difference in time to initiation of cardiopulmonary resuscitation (p = 0.264). The intervention group had overall crisis resource management performance scores 1.15 points higher (Ottawa Global Rating Scale [out of 7]) (p = 0.02). Three months later, these differences between the groups persisted. Conclusions: A 1-hour crisis resource management teaching session improved time to critical initial steps of pediatric resuscitation and crisis resource management performance as measured by the Ottawa Global Rating Scale. The control group did not develop these crisis resource management skills over 3 months of standard training indicating that obtaining these skills requires specific education. Larger studies of crisis resource education are required. (Pediatr Crit Care Med 2014; 15:e168–e174) Key Words: cardiopulmonary resuscitation; crisis resource management; pediatrics; Pediatric Advanced Life Support
E
ducating pediatricians-in-training about resuscitation of acutely ill children presents many barriers. Guidelines and algorithms exist for the resuscitation of children during cardiopulmonary resuscitation (CPR) (1). As resuscitation events in children are rare, it is a challenge to educate pediatric residents in the skills required to manage these events, yet appropriately managing resuscitation is a core competency required by all pediatric trainees (2). It is becoming increasingly recognized that truly successful management of critically ill patients requires strong technical and cognitive skills, as well as complementary strong nontechnical skills (3, 4). May 2014 • Volume 15 • Number 4
Online Clinical Investigations
Crisis resource management (CRM) refers to the nontechnical skills, including leadership, problem solving, situational awareness, communication skills, and resource management (5–7). Recent studies have demonstrated that even brief team building or leadership instruction can improve a resuscitation team’s performance (8, 9). In a pediatric study, CRM teaching led to an increase in perceived team collaboration, satisfaction with care, and observed teamwork skills (10); however, there is a paucity of literature measuring CRM’s effect on medical outcome or changes in ability to perform medical resuscitation maneuvers. These nontechnical skills, although important in the management of acutely ill children, are not acquired passively during residency and hence may require specific education (7, 11). Simulation has been demonstrated to be a useful approach to teach CRM skills (12, 13). To better prepare pediatric residents for caring for acutely ill children, there may be benefit in providing instruction in CRM. We designed this pilot study to determine if a brief small group CRM training session would improve resuscitation performance of junior residents. Our primary objective was to determine if teaching CRM skills leads to improvement in the delivery of pediatric advanced life support (PALS) by junior pediatric residents, measured by a faster initiation of important resuscitation maneuvers. We also aimed to determine if a brief CRM teaching session would improve residents’ CRM skills and finally if teaching CRM had a long-term effect on resuscitation and teamwork behaviors.
MATERIAL AND METHODS Study Population First- and second-year pediatric residents at the University of Alberta were invited to voluntarily participate in this study. Exclusion criteria consisted of extensive acute care experience as defined as more than 1 month in an ICU setting, previous CRM training, or lack of PALS certification. No first- or second-year pediatric residents met exclusion criteria. Written informed consent was obtained from all participants. A prestudy questionnaire determined residents’ level of training, gender, and previous intensive care experience. This information was then used to stratify residents for level of training and critical care experience. Strata were placed in an opaque envelope, and names were randomly drawn to alternately be assigned to CRM and control group from each strata by an assistant not involved in the study. The residents remained in these groups for the duration of the study. The Health Research Ethics Board at the University of Alberta approved this study. Intervention On the day of the research scenario, all participants took part in a 1-hour didactic refresher session, reviewing the knowledge, technical skills, and algorithms important in PALS. Directly after the PALS refresher session, the intervention group had a 1-hour small group CRM training session. This Pediatric Critical Care Medicine
session included didactic teaching, videos of CRM skills, and hands-on interactive practice utilizing CRM skills in an unrelated resuscitation scenario. The format was based on the recent PALS curriculum and the Acute Critical Event Simulation Course (14). Following this session, the group was asked to keep the CRM teaching and intervention content confidential. The control group had no further instruction beyond the PALS refresher session. A standardized orientation to the pediatric high-fidelity patient simulators (BabySim ECS and PediSim ECS; formally METI, now CAE, Saint-Laurent, Quebec, Canada) and crash carts took place for all participants. The assessment scenarios were conducted immediately following the PALS refresher and CRM teaching. Scenarios Each resident assumed the role of “team leader” for two standardized simulated pediatric medical emergency scenarios taken from validated PALS algorithm sequences. The two scenarios were identical for all residents. Scenario 1 was a young child that presented with sinus tachycardia and deteriorated into pulseless ventricular tachycardia necessitating basic life support and defibrillation. Scenario 2 was an infant with respiratory distress and sinus tachycardia deteriorating into sinus bradycardia necessitating basic life support and epinephrine. The two scenarios had common expected initial interventions (including airway assessment and intervention, monitor application, recognition of a critical rhythm change, and CPR) and interventions specific to the algorithms. A brief debriefing session was conducted after each scenario in which any major medical errors were discussed. No teaching around CRM principles occurred in the debriefing. Pediatric intensivists at Stollery Children’s Hospital reviewed the scenarios for content and realism. Each resident team leader was accompanied by a standardized resuscitation team consisting of a respiratory therapist and two registered PICU nurses from the Stollery Children’s Hospital who provided assistance within their scope of practice. The team members’ behavior during each scenario was practiced previously and scripted to ensure identical behavior exposure for each team leader. Real-time clarification was provided as needed via earpiece from the standardized simulator observer familiar with the scenarios. Team members only performed maneuvers when instructed to by the leader and were scripted to prompt the team leader about changes in the patient’s clinical condition in a standardized manner. Team leaders were instructed to involve the team members (respiratory therapist and RNs) as they would in a true scenario. All scenarios were videotaped for analysis with the consent of the participants. Three months after the initial scenario, the participants participated in two more simulated scenarios that were identical to the previous scenarios except for clinical history, age of the patient, and simulator size. The resuscitation team and simulation observer remained the same. www.pccmjournal.org
e169
Blackwood et al
Assessment Using the video recordings, timing to a predetermined endpoint of a priori–specified medical interventions from fixed times “zero” was determined by one of the investigators (J.B.). The initial “time zero” commenced from the end of the scenario presentation by the nurse and was used to determine times to expected basic life support maneuvers. The second “time zero” mark was 2 minutes later when the heart rhythm converted from sinus to a critical pathologic rhythm and was used to time predetermined medical interventions required for that specific PALS algorithm. Interventions were chosen that were common in sequence in the two algorithms. If a basic life support intervention did not occur prior to pathological rhythm change (second time zero), then “2 minutes” was allotted as the intervention time and “a critical error” was marked. Deviations from the PALS algorithms (wrong algorithm and medication or interventions not performed or performed out of order) were also noted. Prior to reviewing the study scenarios, the three raters standardized their understanding of the Ottawa Global Rating Scale (OGRS) descriptive anchors. The raters then independently scored previously recorded scenarios (not related to the study) demonstrating below average, average, and above average team performance and then debriefed until rater scores were consistent for practice scenarios. The raters reviewed all of the recordings, blinded to which arm of the study the participants were part of, and independently scored the resident’s nontechnical performance using the validated OGRS for crisis management skills (15). The OGRS allows the rater to rate the overall performance from 1 to 7 and also rates the subject’s performance in five other domains (leadership, problem solving, situational awareness, resource utilization, and communication) from 1 to 7 for an additional 35 points. The groups’ CRM Score Global Rating as well at the additive score of all six OGRS categories (out of 42) was analyzed. The ratings given by the raters were averaged, and in order to assess the agreement of these, we determined the intraclass correlation coefficient (ICC). The ICC was calculated for each of the four scenario-time combinations (ICC range, 0.74–0.92; p < 000.1). These values showed very good to excellent agreement between raters, and hence, averaging the ratings as is usual practice made sense. Statistics All of the pediatric residents that were available and consented were included in the study. A convenience sample of 15 first- and second-year residents was enrolled. Descriptives per scenario/ date/group are reported as medians with interquartile ranges for nonsymmetrically distributed data and means with sds for normally/symmetric distributed data. Demographic data, which were all categorical-type data, were assessed for differences via Agresti-Coull-Caffo 95% CIs for difference in proportions. A linear mixed effects model with fixed effects for intervention and date and their interaction plus random effect for subjects was used to detect differences/effects on time to maneuver due to each intervention and date. This technique models both the intervention and the time effect via models that assume an “on average” effect. As all of the scenarios e170
www.pccmjournal.org
required the same basic life support interventions, scenarios were not considered to be differential with respect to the outcome being measured, and hence, it was not included as a variable of effect in the model. Data were entered into Microsoft Excel, then cleaned, and analyzed with R: A Language and Environment for Statistical Computing program (http://www.r-project.org/) (16) and using the “nlme” package (http://cran.r-project.org/web/packages/ nlme/index.html) (17). Comparisons of the mean times to maneuvers for control and CRM groups were carried out considering the potential period/intervention statistical interaction (effect modification) and the repeated measures nature of the data. Mean times were used in the model as no method exists for using median data. Significance of the effects of intervention and date was determined using linear mixed effects models: one for times to maneuvers and another for CRM scores. Crude descriptive data per scenario, arm, and date are reported as medians with interquartile ranges. The level of significance used was set at 0.05.
RESULTS Participants All first- and second-year pediatric residents in the University of Alberta Pediatric residency training program (n = 15) participated in the study. Seven were randomized to the intervention group and eight to the control group. All residents attended both the initial and follow-up sessions. Please see Appendix 1 for the flow diagram outlining the events involved for the participants. There were no differences in gender, level of training, previous ICU training, or PALS certification between groups (Table 1). As the study period occurred early in the academic year cycle, no regularly scheduled PALS recertification or resident promotion between years of training occurred. Time to Interventions Using a linear mixed effects model, there were significant differences found between the CRM and control groups for several of the predetermined medical interventions (Table 2). Table 1. Demographic Information About the Participants Intervention Group (n = 7) (%)
Control Group (n = 8) (%)
Female
7 (100)
7 (88)
Postgraduate year 2
3 (43)
4 (50)
Previous ICU experience
5 (71)
4 (50)
Mock code leader experience
2 (29)
5 (63)
Pediatric advanced life support certified
7 (100)
8 (100)
May 2014 • Volume 15 • Number 4
Online Clinical Investigations
Table 2.
Time to Predetermined Medical Interventions Initial Scenarios
Time to Medical Intervention, Median (s) (IQR)
CRM
Control
Follow-Up Scenarios CRM
Control
Model Effect (95% CI)
p
Assessment of airway Bradycardia
17.4 (15.2–78.3)
24.0 (14.6–133.8)
125.1 (28.4–269.8)
68.4 (44.4–265.2)
VT
39.2 (33.0–47.7)
35.5 (16.8–214.8)
83.8 (30.3–153.1)
139.4 (24.1–243.8)
Bradycardia
78.3 (66.8–102.0)
84.9 (71.1–129.8)
75.0 (65.7–96.2)
134.2 (122.2–152.8)
VT
172.4 (131.4–197.6)
160.5 (126.3–219.2)
52.7 (45.8–95.0)
97.4 (48.3–103.3)
Bradycardia
109.5 (60.2–137.0)
83.7 (76.6–116.6)
65.1 (20.7–96.3)
60.5 (37.8–137.2)
VT
58.8 (51.0–90.8)
35.6 (27.1–48.6)
129.6 (81.3–177.5)
82.8 (51.4–91.3)
Bradycardia
73.5 (62.0–76.5)
88.2 (54.9–118.6)
60.6 (52.0–71.4)
82.5 (67.4–125.8)
VT
70.4 (60.0–79.2)
95.8 (67.1–124.0)
56.5 (53.4–78.5)
95.3 (81.2–108.2)
Bradycardia
155.7 (137.8–184.6)
216.3 (179.7–271.4)
100.5 (79.5–132.0)
93.8 (57.3–123.4)
VT
111.2 (98.5–172.4)
160.0 (117.0–186.4)
128.6 (82.1–146.4)
189.5 (147.0–334.0)
16.2 (7.4–69.5)
185.7 (101.1–256.0)
18.3 (12.0–42.0)
20.5 (10.7–113.7)
36.0 (23.3–141.6)
58.2 (35.4–101.8)
18.5 (11.1–25.7)
46.0 (29.0–113.2)
CRM 26.7 s faster (110 s slower to 56.7 s faster)
0.09
CRM 23.3 s faster (56.7 s slower to 10.2 s faster)
0.06
CRM 8 s slower (70 s slower to 50 s faster)
0.77
CRM 24.6 s faster (3.5 s faster to 45.8 s faster)
0.03
CRM 47.1 s faster (0.8 s faster to 93.4 s faster)
0.04
CRM 50.4 s faster (6.5 s faster to 94.4 s faster)
0.03
CRM 84.9 s faster (30.4 s faster to 139.3 s faster)
0.01
CRM 27 s faster (19 s slower to 73 s faster)
0.23
CRM 14 s faster (12 s slower to 41 s faster)
0.26
Place oxygen on
Initial pulse check
Place monitors on
Placing an IV
Calling for help Bradycardia VT
Pulse check after rhythm change Bradycardia
18.6 (11.3–50.0)
59.4 (43.2–352.0)
21.2 (4.4–76.0)
343.4 (98.6–352.0)
VT
31.5 (11.7–46.9)
31.0 (17.8–74.3)
13.8 (11.4–39.5)
35.4 (9.2–76.7)
Start bag mask ventilation after apnea Bradycardia
8.4 (3.6–12.9)
6.9 (0.1–63.6)
24.3 (13.3–41.7)
42.6 (30.8–204.4)
VT
40.5 (29.8–56.0)
35.9 (16.3–53.2)
11.4 (9.5–30.1)
7.5 (5.4–50.4)
Cardiopulmonary resuscitation after pulseless Bradycardia
13.5 (10.1–24.6)
17.1 (13.1–56.9)
11.0 (8.6–26.1)
32.4 (10.5–57.7)
VT
60.6 (23.2–89.5)
30.2 (13.8–52.8)
27.3 (20.1–49.2)
19.8 (11.0–81.6)
IQR = interquartile range, CRM = crisis resource management, VT = ventricular tachycardia.
Pediatric Critical Care Medicine
www.pccmjournal.org
e171
Blackwood et al
The intervention CRM group placed monitor leads 24.6 seconds earlier (95% CI, 3.5–45.8; p = 0.02), placed an IV 47.1 seconds sooner (95% CI, 0.8–93.4; p = 0.04), and called for help 50.4 seconds faster (95% CI, 6.5–94.4; p = 0.03). Finally, the intervention group checked for a pulse after noticing a rhythm change 84.9 seconds quicker (95% CI, 30.4 s faster to 139.3 s faster; p = 0.01). There was no statistically significant difference found for initial pulse check prior to pathologic rhythm and starting bag valve mask ventilation (BMV) or CPR after the patient went pulseless. Data were also analyzed for effect modification of the time period on intervention group and none was found. Thus, when the groups were tested 3 months later, the ability for the CRM group to perform the same skills faster persisted with no improvement in the control group over time. In addition, at the 3-month retest, there was an indication of faster CPR initiation for the CRM group compared with their initial times (Table 2). CRM Performance Using the linear mixed effects model, the CRM group had an “overall OGRS” score (out of 7) that was 1.15 points (95% CI, 0.2–2.1, p = 0.02) higher than that in the control group, and this increase in overall CRM score was maintained at the 3-month retest scenario (Table 3). The summative score of all seven categories (out of 42) was 6.7 points (1.6–11.8, p = 0.01) higher in the CRM group, and this difference was not affected by retest 3 months later.
DISCUSSION This study investigated the effect of teaching CRM on the resuscitation performance of junior pediatric residents in high-fidelity simulation scenarios. We found that those that were taught CRM performed certain skills and medical interventions required during a simulated resuscitation quicker than a similarly experienced control group. In addition, the intervention group had better CRM skills, as measured by the OGRS, when compared with the control group. In our study, initial steps of the resuscitation were the skills that appeared to be most improved by teaching CRM. Residents, after only a brief CRM instruction session, called for help, placed monitors, and checked a pulse faster than our Table 3.
control group. A study by Hunt et al (18) revealed significant delays in the initiation of important resuscitation maneuvers in simulated pediatric resuscitation scenarios. They concluded that there should be focus on improving the quality of care delivered in the first 5 minutes of resuscitation efforts. As demonstrated in our study, teaching CRM may be an important component of this. It is possible that the initial organization of the team with good communication, leadership, and role assignment is critical for timely performance of these activities. This is consistent with a study by Jankouskas et al (19) in which medical and nursing students were randomized to CRM and basic life support (BLS) training or BLS training alone. In that study, the addition of CRM training had no effect to time to BLS (including b ag-mask-ventilation and chest compressions). However, in that study, the teams continued to practice their BLS skills as a team and both groups demonstrated improvement in BLS skills at the end of the training. In our study, our team members were scripted to not start CPR unless specifically directed to by the team leader. It is possible that with CRM training, team members will feel more comfortable making suggestions and potentially improve time to CPR initiation. Additional and repeated simulation training and debriefing of CRM in concert with simple and complex acute care pediatric scenarios with the in situ code team may be necessary to see improved team effectiveness in the more complex PALS algorithm above and beyond BLS skills. This study demonstrates improvement with only 1 hour of specific CRM teaching. More dramatic and lasting effects may occur with a program of regular CRM teaching and feedback. Recent reviews have outlined the importance of teamwork, effective communication, and leadership behavior in managing emergency situations (7, 20). Such studies have lead to the inclusion of CRM into PALS teaching. Our study did show some benefits of CRM training with initial assessment and management in resuscitation. This benefit was maintained at a 3-month assessment with no improvement in the control group. There is no indication that leads us to believe that despite the importance of CRM behaviors in resuscitation, residents are developing these skills over the
Summary of Ottawa Global Rating Scale Ratings Initial Scenarios
OGRS Score, Mean (sd)
CRM
Control
Follow-Up Scenarios CRM
Control
Model Effect (95% CI)
p
Overall performance (out of 7) (sd) Bradycardia
4.7 (0.9)
2.9 (1.0)
4.3 (1.1)
3.4 (1.0)
VT
4.4 (1.1)
3.8 (0.8)
4.5 (1.1)
3.3 (1.1)
CRM OGRS 1.15 points higher (0.2–2.1)
0.02
CRM OGRS 6.7 points higher (1.6–11.8)
0.01
Sum of all categories (out of 42) (sd) Bradycardia
28.6 (4.6)
18 (5.8)
26.8 (6.4)
21.2 (5.4)
VT
27.3 (5.0)
23.3 (4.4)
27.4 (6.6)
20.7 (6.6)
OGRS = Ottawa Global Rating Scale, CRM = crisis resource management, VT = ventricular tachycardia.
e172
www.pccmjournal.org
May 2014 • Volume 15 • Number 4
Online Clinical Investigations
course of their routine clinical training. We advocate that it is important to incorporate CRM into pediatric training curriculum (7, 11). The time to initiation of CPR, a marker that is known to correlate with outcome (21), was not statistically different between the groups, and it took a protracted amount of time for the leader to ask the team to initiate CPR. Our scenarios were designed such that CPR could not be initiated until asked for by the team leader. In a clinical event, CPR may be recommended to the leader by any of the team in attendance. Previous studies have demonstrated that nontechnical skills are important in the quality of CPR (8, 22, 23). In our study, the nontechnical performance of the CRM participants was improved with better communication and team dynamics. With these improvements in team dynamics, the other team members may feel comfortable suggesting CPR initiation in actual clinical scenarios and potentially improve CPR times. Although not significant, there was a trend toward an earlier start to CPR in the CRM group; perhaps with a larger study and more intense CRM training, a significant difference may be found. It should be noted that there were a number of deviations in PALS algorithms demonstrated by participants in our study, despite all being PALS certified and participating in a 1-hour PALS refresher. Deviations included failing to BMV, failing to defibrillate, or not checking for a pulse after a noted rhythm change. This is consistent with other studies (24). A recent study of second-year pediatric residents noted a rapid decline in their resuscitation skills following a PALS course (25). Limitations This study was limited in that it was a small single-center pilot study. However, even with such small numbers, significant differences in some resuscitation teamwork behaviors were noted. A larger trial is required to confirm these findings. Although the CRM raters were blinded to participant group allocation, the investigator measuring time to critical interventions was not. However, as this is objective data (time in seconds) being recorded by the same reviewer for all scenarios, the effect of differential bias is low. It must be acknowledged that our study did not assess participants at baseline and this may contribute to some of the differences noted though randomization should minimize that bias. Assessments of participants prior to and following CRM teaching would be beneficial to undertake with any further studies. The intervention arm received an extra hour of CRM training and that extra time in training may explain some of the improvements in medical maneuvers we noted. This study focused on the team leader as all of the team members’ responses were scripted. This study was a pilot study in which the leader, and not all members of the team, was instructed on CRM but did not require mastery of the skill. Larger studies in which these considerations are accounted for are warranted to investigate the effect on resuscitation measures. Pediatric Critical Care Medicine
CONCLUSIONS Pediatric resuscitation is a skill that is necessary for all pediatric medicine practitioners, yet it is difficult to both teach and experience due to the relatively rare number of events. Simulation has been shown to assist in learning these skills. CRM can help pediatric resident team leaders recognize and assess sick patients, call for appropriate help sooner, place IVs, and monitoring sooner. Our results suggest that CRM is not a skill set that is passively developed and may require specific education and feedback. This pilot study suggests that larger well-designed studies of crisis resources education should be undertaken.
REFERENCES
1. Kleinman ME, Chameides L, Schexnayder SM, et al; American Heart Association: 2010 American Heart Association (AHA) guidelines for cardiopulmonary resuscitation (CPR) and emergency cardiovascular care (ECC) of pediatric and neonatal patients: Pediatric advanced life support. Pediatrics 2010; 126:e1361–1399 2. Eppich WJ, Adler MD, McGaghie WC: Emergency and critical care pediatrics: Use of medical simulation for training in acute pediatric emergencies. Curr Opin Pediatr 2006; 18:266–271 3. Fletcher GCL, McGeorge P, Flin RH, et al: The role of non-technical skills in anaesthesia: A review of current literature. Br J Anaeseth 2002; 88:418–429 4. Gaba DM, Fish KJ, Howard SK: Chapter 1-Theory of Dynamic Decision Making and Crisis Management. In: Crisis Management in Anesthesia. New York, Churchill Livingstone, 1994, pp 5–29 5. Glavin RJ, Maran NJ: Integrating human factors into the medical curriculum. Med Educ 2003; 37(Suppl 1):59–64 6. Eppich WJ, Brannen M, Hunt EA: Team training: Implications for emergency and critical care pediatrics. Curr Opin Pediatr 2008; 20:255–260 7. Cheng A, Donoghue A, Gilfoyle E, et al: Simulation-based crisis resource management training for pediatric critical care medicine: A review for instructors. Pediatr Crit Care Med 2012; 13: 197–203 8. Hunziker S, Tschan F, Semmer NK, et al: Hands-on time during cardiopulmonary resuscitation is affected by the process of teambuilding: A prospective randomized simulator-based trial. BMC Emergency Medicine 2009; 9:3 9. Hunziker S, Bühlmann C, Tschan F, et al: Brief leadership instructions improve cardiopulmonary resuscitation in a high-fidelity simulation: A randomized controlled trial. Crit Care Med 2010; 38: 1086–1091 10. Jankouskas T, Bush MC, Murray B, et al: Crisis resource management: Evaluating outcomes of a multidisciplinary team. Simul Healthc 2007; 2:96–101 11. Gilfoyle E, Gottesman R, Razack S: Development of a leadership skills workshop in paediatric advanced resuscitation. Med Teach 2007; 29:e276–e283 12. Yee B, Naik VN, Joo HS, et al: Nontechnical skills in anesthesia crisis management with repeated exposure to simulation-based education. Anesthesiology 2005; 103:241–248 13. Reznek M, Smith-Coggins R, Howard S, et al: Emergency medicine crisis resource management (EMCRM): Pilot study of a simulationbased crisis management course for emergency medicine. Acad Emerg Med 2003; 10:386–389 14. Neilipovitz DT: Acute Resuscitation and Crisis Management. Ottawa, University of Ottawa Press, 2005 15. Kim J, Neilipovitz D, Cardinal P, et al: A pilot study using high-fidelity simulation to formally evaluate performance in the resuscitation of critically ill patients: The University of Ottawa Critical Care Medicine, High-Fidelity Simulation, and Crisis Resource Management I Study. Crit Care Med 2006; 34:2167–2174 www.pccmjournal.org
e173
Blackwood et al 16. R Development Core Team: R: A Language and Environment for Statistical Computing. Vienna, Austria, R Foundation for Statistical Computing, 2011 17. Pinheiro J, Bates D, DebRoy S, et al; the R Development Core Team: nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1–102, 2011 18. Hunt EA, Walker AR, Shaffner DH, et al: Simulation of in-hospital pediatric medical emergencies and cardiopulmonary arrests: Highlighting the importance of the first 5 minutes. Pediatrics 2008; 121: e34–e43 19. Jankouskas TS, Haidet KK, Hupcey JE, et al: Targeted crisis resource management training improves performance among randomized nursing and medical students. Simul Healthc 2011; 6:316–326 20. Hunziker S, Johansson AC, Tschan F, et al: Teamwork and leadership in cardiopulmonary resuscitation. J Am Coll Cardiol 2011; 57:2381–2388 21. Ali B, Zafari AM: Narrative review: Cardiopulmonary resuscitation and emergency cardiovascular care: Review of the current guidelines. Ann Intern Med 2007; 147:171–179 22. Thomas EJ, Sexton JB, Lasky RE, et al: Teamwork and quality during neonatal care in the delivery room. J Perinatol 2006; 26:163–169 23. Marsch SC, Müller C, Marquardt K, et al: Human factors affect the quality of cardiopulmonary resuscitation in simulated cardiac arrests. Resuscitation 2004; 60:51–56 24. Hunt EA, Vera K, Diener-West M, et al: Delays and errors in cardiopulmonary resuscitation and defibrillation by pediatric residents during simulated cardiopulmonary arrests. Resuscitation 2009; 80:819–825 25. Roy KM, Miller MP, Schmidt K, et al: Pediatric residents experience a significant decline in their response capabilities to simulated life-threatening events as their training frequency in cardiopulmonary resuscitation decreases. Pediatr Crit Care Med 2011; 12:e141–e144
e174
www.pccmjournal.org
Flow diagram outlining the events involved for the participants of the study.
Appendix 1.
PALS = pediatric advanced life support, CRM = crisis resource management.
May 2014 • Volume 15 • Number 4
