Efficacy of a single‐day task trainer–based extracorporeal membrane oxygenation training course

Abstract

Background

Extracorporeal membrane oxygenation (ECMO) is an advanced medical technology used to treat respiratory and heart failure. The coronavirus pandemic has resulted in significantly more ECMO patients worldwide. However, the number of hospitals with ECMO capabilities and ECMO‐trained staff are limited. Training of personnel in ECMO could supplement this demand.

Objective

To evaluate our previously developed ECMO course using a task trainer–based training, as opposed to an existing live tissue–training model, and determine if such a program was adequate and could be expanded to other facilities.

Methods

Seventeen teams, each consisting of a physician and nurse, underwent a 5 hour accelerated ECMO course in which they watched prerecorded ECMO training lectures, primed circuit, cannulated, initiated ECMO, and corrected common complications. Training success was evaluated via knowledge and confidence assessments and observation of each team attempting to initiate ECMO while troubleshooting complications on a Yorkshire swine.

Results

Seventeen teams successfully completed the course. Sixteen teams (94%, 95% CI = 71%–100%) successfully placed the swine on veno‐arterial ECMO. Of those 16 teams, 15 successfully transitioned to veno‐arterial–venous ECMO. The knowledge assessments and confidence levels of physicians and nurses increased by 24.3% from pretest (mean of 65.3%, SD 14.4%) to posttest (mean of 89.6%, SD 10.3%), p < 0.0001.

Conclusions

An abbreviated one day lecture and hands‐on task trainer–based ECMO course resulted in a high rate of successful skill demonstration and improvement of physicians' and nurses' knowledge assessments and confidence levels, similar to our previous live tissue training program.

INTRODUCTION

Extracorporeal membrane oxygenation (ECMO) is an advanced medical technology used to treat refractory respiratory failure, heart failure, or both. When used for respiratory failure, venovenous (VV) ECMO pumps deoxygenated blood out of a central vein, oxygenates it, and returns it via a central vein. Conversely, venoarterial (VA) ECMO can support the heart by pumping blood out of a central vein and into an artery. Finally, venoarterial‐venous (VAV) ECMO can support patients with both pulmonary and cardiac failure by pumping blood out of a vein, oxygenating it, and pumping it into both a central artery and vein via separate cannulas. Studies have demonstrated the efficacy of VV ECMO for management of severe Acute Respiratory Distress Syndrome (ARDS). ¹ , ² , ³ , ⁴ Several authors have indicated the importance of ECMO use and properly trained ECMO clinicians on rescuing patients from cardiopulmonary complications, including COVID‐19 patients. ⁵ , ⁶ , ⁷ , ⁸ Of further relevance to emergency physicians is the use of VA ECMO for management of out‐of‐hospital cardiac arrest. ⁹ , ¹⁰ , ¹¹ , ¹² Recently, a study was prematurely terminated by reviewers given that survival in the ECMO arm was greater compared to therapy (33% vs. 7%) and thus, enrollment of patients into the control arm was considered unethical. ¹³ The coronavirus disease 2019 (COVID‐19) pandemic resulted in a significant increase in ECMO demand worldwide and the initiation of ECMO in emergency departments, despite a sufficient number of ECMO specialty teams. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸

When a trauma patient is received at the hospital, handoff primarily occurs in the Emergency Department from Emergency Medical Services (EMS) to Emergency Medicine (EM) and Critical Care (CC) physicians and staff. It is there that the EM team will need to identify the need for ECMO, call for the ECMO team, and act quickly to initiate ECMO, especially if the ECMO specialty team is not nearby. For this reason, our focus was on these specific participants. Additionally, emergency medicine residents are actively engaged in the management of ECMO patients during Intensive Care unit (ICU) rotations and those electing to receive EMS and critical care fellowship training will likely experience an increased prevalence of ECMO patients.

Despite evidence for the benefit of ECMO, its availability (ECMO specialty teams and equipment) remains limited to major medical centers. In our previous study, we developed and validated an abbreviated single‐day “Teach, Train, Test” ECMO course, to train teams of one EM or CC physician and either a nurse or respiratory therapist, to prime the circuit, initiate VA ECMO, transition to VAV ECMO, and troubleshoot common ECMO complications. ¹⁹ We taught participants using prerecorded audiovisual lectures and hands‐on training and testing using a live tissue model (Yorkshire swine). However, since most training facilities lack access to live tissue models, we opted to conduct this study utilizing a task trainer instead with the anticipation that task trainers could be incorporated into simulation labs for more frequent utilization in the near future.

When comparing medical task trainers to livetissue, there are considerable differences in malleability, elasticity, anatomy, and tensile strength. ²⁰ , ²¹ For example, some task trainers use plastic tubing to mimic vascular structures, which is generally not as collapsible and lacks the pulse‐to‐pulse change in diameter seen in live tissue models and humans. Since laceration of the central artery or vein was one of the more serious and common complications encountered during cannulation in our previous study, we were concerned that using a task trainer alone would be too simple for students, resulting in a false sense of confidence. Conversely, a swine model presents a greater challenge to clinicians since the femoral artery and vein are smaller, more tortuous, and covered by a greater quantity of muscular tissue, compared to human central vessels. Therefore, we aimed to develop a task trainer–based ECMO course to use within training facilities lacking access to live tissue models and determine its efficacy.

Our objective was to evaluate a team consisting of one EM or CC physician and one nurse or respiratory therapist, who completed the accelerated training on the task trainer, on their ability to successfully initiate ECMO and troubleshoot ECMO complications on the live swine model. This study had mixed participants but predominantly EM physicians and ICU nurses (Table 1).

TABLE 1.

Student characteristics

Variable	Total sample (n = 34)	Physicians (n = 17)	RNs (n = 14) and RTs (n = 3)
Time in position (%), y
<1	3 (9)	1 (6)	2 (12)
1–5	23 (68)	12 (71)	11 (65)
6–10	6 (18)	3 (18)	3 (18)
>10	2 (6)	1 (6)	1 (6)
Department (%)
ED	21 (62)	16 (94)	5 (29)
ICU	10 (29)	0 (0)	10 (59)
Research	1 (3)	0 (0)	1 (6)
Respiratory therapy	1 (3)	0 (0)	1 (6)
Surgery	1 (3)	1 (6)	0 (0)
Time in department (%), y
<1	5 (15)	1 (6)	4 (24)
1–5	23 (68)	12 (71)	11 (65)
6–10	4 (12)	3 (18)	1 (6)
>10	2 (6)	1 (6)	1 (6)
Work experience caring for ECMO patient(s) (%), h
0	15 (44)	5 (29)	10 (59)
1–10	16 (47)	10 (59)	6 (35)
11–35	1 (3)	1 (6)	0 (0)
>36	2 (6)	1 (6)	1 (6)
Formal ECMO training	0 (0)	0 (0)	0 (0)
Other training (%)
Advanced trauma life support (task trainer lab)	15 (44)	14 (82)	1 (6)
Advanced trauma life support (animal lab)	11 (32)	11 (65)	0 (0)
Pediatric advanced life support	26 (76)	17 (100)	9 (53)
Neonatal resuscitation program	12 (35)	9 (53)	3 (18)
Emergency war surgery course	1 (3)	1 (6)	0 (0)
Swine procedure lab or research	5 (15)	5 (29)	0 (0)

Note: Values given are number (column percentage).

Abbbreviations: ECMO, extracorporeal membrane oxygenation; ED, emergency department; ICU, intensive care unit.

METHODS

Study design

As in our previous ECMO study, we continued the “Teach, Train, Test” method to evaluate the efficacy of our abbreviated ECMO course to provide the teams with the necessary skills to initiate and maintain ECMO therapy and troubleshoot common complications. ¹⁹ However, in our previous study, a live swine was used for the “Train” section as opposed to the task trainer in this study. To determine the effectiveness of simulation training, the “Test” section, was conducted on live swine as our previous study (Figure 1). The study was determined to be exempt by the 59th Medical Wing (MDW) Institutional Review Board and was approved by the 59th MDW Clinical Investigation and Research Support (CIRS) Institutional Animal Care and Use Committee (IACUC).

FIGURE 1.

Flow chart of study procedures

Study setting and selection of participants

We solicited volunteers from the San Antonio Military Medical Center (SAMMC), the Department of Defense's only Level I trauma center, via e‐mail, word of mouth, and presentations at department grand rounds. Since our current medical facility policy is that those physicians, nurses, and respiratory therapists with emergency medicine and/or critical care training are eligible to volunteer for our facility ECMO team, we established a similar criteria for our study. These participants are also primarily the first line of care when a patient enters the hospital as a trauma patient. Eligible study participants included EM and critical care physicians (attending or third year), critical care fellowship physicians, third‐year or later general surgery residents, EM and critical care nurses, and respiratory therapists, who had no previous formal ECMO training or ECMO cannulation experience. Participants in our previous study were not eligible for this study given the prior ECMO training they received. We adhered to the regulations and guidelines of the Animal Welfare Act, the National Research Council Guide for the Care and Use of Laboratory Animals, and the American Association for the Accreditation of Laboratory Animal Care. The Congressionally Directed Medical Research Program (CDMRP) Joint Program Committee 6 (JPC − 6) /Combat Casualty Care Research Program funded the study.

Materials

The CARDIOHELP ECMO pump (Maquet Getinge Group CARDIOHELP‐I), circuit (Maquet HLS Set Advanced 7.0, HLS 7050 USA, 701052794) and the Maquet arterial and venous cannulas (PAL 1523, PVS 1938, and PVL 2155) were used. ¹⁴ Maquet percutaneous insertion kits (12Fr.‐18Fr.; PIK 150‐USA) and the 20‐French Avalon Elite vascular dilator (20Fr.; #12210) were used for vessel dilation.

Course materials

The course was set up in a “Teach, Train, Test” model in which participants first reviewed a total of 1.5 hours of narrated lectures with visual slides covering the following ECMO topics: indications, contraindications, physiology, patient management, cannula selection, cannulation strategies, routine circuit management, and troubleshooting circuit emergencies. Following lecture completion, subjects were given hands‐on training on the ECMO circuit and task trainers (2–3 hours). Previously developed procedural checklists utilized in our prior study (Supplementary Table S1) were used for training and assessment for each the following tasks: set‐up and priming of the ECMO circuit, patient cannulation, and management of ECMO complications. ¹⁹

Hands‐on training with task trainers

Hands‐on training was provided to each team by an ECMO‐trained physician, a nurse, and nurse assistants. The training included verbal and hands‐on instruction to the subjects as they assembled and primed the ECMO circuit and cannulated the femoral vessels (two veins and one artery) of the ECMO Vascular Access Model task trainer (Christopher Ho, MD, Inc.). The task trainer resembles a human pelvis containing bilateral femoral artery and vein access (tubing), connected to a saline bag system allowing for continuous fluid circulation upon ECMO initiation.

Participants were reminded to use the written instructions in a stepwise fashion throughout this process. Cannulation was accomplished using an ultrasound‐guided percutaneous Seldinger technique. Serial dilations of one of the femoral veins was performed up to 20 French and then a 21‐French cannula (Maquet HLS cannula) was placed. Serial dilations of a femoral artery were performed up to 14 French and then a 15‐French cannula (Maquet arterial HLS cannula) was placed. Cannulas were de‐aired and attached to the ECMO circuit.

Following confirmation of successful VA ECMO initiation, cannulation of the second femoral vein in the task trainer was accomplished. Serial dilations of the second femoral vein were performed up to 18‐Fr and a 19‐Fr. cannula (Maquet venous HLS cannula) was used. The cannula was then de‐aired and a clamp was placed ensuring no entrainment of air or leaking of fluid (saline) from the cannula. To initiate VAV ECMO, the arterial line flowing out of the ECMO circuit was clamped and cut. A Y‐connector (NovoSci, C330S) was placed, allowing the fluid pumped out of the ECMO circuit into the task trainer's artery to be diverted into the 15‐Fr. arterial cannula and the 19‐Fr. venous cannula (Figure 2). Flow regulator clamps were placed on the tubing of the 19‐Fr. venous cannula, allowing participants to regulate the blood flow from the Cardiohelp between the arterial and venous system.

FIGURE 2.

VAV ECMO Set‐up for task trainer–based ECMO course. Left panel: Overall (Venous–Venous–Arterial) cannulation on task trainer. Middle panel: Task trainer cannulas attached to CARDIOHELP ECMO device. Right panel: Task trainer connected to fluid reservoir.

Following confirmation of successful VAV ECMO initiation in the task trainer, participants were provided with hands‐on training on identification and management of common ECMO complications including loss of electrical power, access insufficiency (often referred to as “chatter”), air in the ECMO circuit, and loss of circuit integrity (i.e., a hole in the tubing).

Additional cannulation practice took place using the ECMO Simulator Prototype task trainer (University of Illinois, Coordinated Science Laboratory, Health Care Engineering System Center, Urbana, Illinois). The Simulator is a human torso with access to a unilateral femoral artery and vein connected to a peristaltic pump providing pulsatile fluid flow, resembling human systolic and diastolic blood flow.

Animal preparation

Given the inability to evaluate mortality and physiology using a task trainer model and the lack of practical and ethical means of enrolling human volunteers, validation was performed using a live animal model. Seventeen Yorkshire swine (Sus scrofa) weighing between 70 and 90 kg (one swine per team) were used based on a priori power analysis. Each animal was sedated with intramuscular ketamine (10 mg/kg), and endotracheal intubation was performed. Mechanical ventilation was commenced and adjusted to maintain the arterial partial pressure of CO₂ (PCO₂) between 38–42 mmHg using a volume‐limited, time‐cycled ventilator (Fabius GS anesthesia machine, Drager‐Siemens). Preemptive analgesia was used in each animal. Anesthesia was maintained by a qualified surgical technician throughout the procedure using isoflurane mixed with room air (1%–3.5%). The publication from our previous study can be referenced for further specifics regarding animal preparation, instrumentation, and sedation. ¹⁹ Following stabilization, a baseline arterial blood gas (ABG) was collected to obtain measurements of oxygen saturation, PaO₂, PaCO₂, hemoglobin (Hb), pH, bicarbonate, base excess, and lactate (ABL 800 Flex blood gas analyzer, Radiometer America).

Training validation and testing in swine model

Upon completion of the hands‐on training, teams were provided the supplies necessary to prepare the circuit, access and cannulate the required blood vessels, initiate and maintain ECMO therapy, and troubleshoot common complications.

Two research assistants independently observed participants to evaluate and document performance and time to completion of each major task taught during the hands‐on training. Any difference in documented information between the research assistants were clarified via consensus; if there was no consensus, the primary investigator made the final determination. If participants were unsuccessful in initiating ECMO, an ECMO specialist would intervene and initiate ECMO. Failed ECMO initiation was documented and participants proceeded with the next steps.

First, participants prepared and primed the ECMO circuit. Then, cannulation was accomplished using an ultrasound‐guided percutaneous Seldinger technique and placed guidewires in each femoral vein and in one femoral artery; placement was verified via fluoroscopy. In the event the guidewires were not located in the inferior vena cava and descending aorta, participants removed the improperly placed wire and reattempted appropriate placement. Following confirmation, heparin (10,000 units) was administered intravenously. Participants then performed the aforementioned serial dilations of one of the femoral veins and the femoral artery, placed the appropriate cannulas, and connected the cannulas to the ECMO circuit.

After initiation of VA ECMO, participants were excused from the room and the research team induced a common complication. Upon their return they were instructed to “assess the patient” and the Cardiohelp system, state the complication and perform the necessary steps to correct the problem. This process was repeated for each complication: loss of electrical power, excessive blood flow turbulence (“chatter”), diagnosis and removal of air from the venous side of the circuit, diagnosis and removal of air from the arterial side of the circuit, and loss of circuit integrity. Research assistants documented success vs. failure and time to completion for each task.

After completion of the troubleshooting phase, participants were instructed to place the patient on VAV ECMO. Participants cannulated the second femoral vein and connected the femoral vein cannula to the arterial ECMO line. Flow regulator clamps were used to establish the balance of blood flow necessary to ensure swine's adequate oxygenation (pulse oximetry >92%) and appropriate blood perfusion (mean arterial pressure >55 mmHg). Research assistants documented success vs. failure and time to completion. Following completion of the study, subjects were debriefed and lessons learned as well as feedback regarding the team's performance was conducted. Swine were euthanized (IV pentobarbital, 100 mg/kg) under veterinary guidance in accordance with the American Veterinary Medical Association Guidelines for the Euthanasia of Animals, 2020.

Data collection

At the beginning of each ECMO course, participants completed a survey to obtain baseline demographic information, prior ECMO training, and ECMO comfort level (Table 1). Each participant completed a written assessment to determine their ECMO knowledge, prior to receiving any course materials (Figure 1). At the completion of the hands‐on skills validation, students completed posttraining written assessments of ECMO knowledge and comfort level.

Measurements

For comparison of efficacy of our previous animal‐based training to the new task trainer–based training, we used identical task completion assessments and comfort and knowledge assessment documents as used in the previous study.

The confidence assessment consisted of 10 items describing the necessary steps to initiate and maintain ECMO (Table 2). Participants rated their confidence and experience level for each item on a scale from 0 (“no experience”) to 5 (“expert”). The knowledge assessment consisted of 20 questions (multiple choice and true/false) focused on ECMO procedures, indications, complications, and troubleshooting.

TABLE 2.

Proportion of students rating themselves as “competent” or better on the confidence assessment

Question	Pretest (%)	Posttest (%)	Percent difference/increase (95% CI)
1. Determining which patients would benefit from ECMO initiation.	2 (6)	16 (47)	41 (24–59)
2. Initiation of IV anticoagulation.	19 (56)	25 (74)	18 (0–36)
3. Placement of percutaneous cannula using Seldinger technique.	12 (35)	19 (56)	21 (2–39)
4. Preparing cannulas for connection to ECMO circuit.	0 (0)	22 (65)	65 (48–82)
5. Connecting patient to ECMO circuit.	0 (0)	22 (65)	65 (48–82)
6. Securing cannulas.	5 (15)	21 (62)	47 (29–65)
7. Achieving respiratory and hemodynamic goals.	9 (26)	18 (53)	27 (9–44)
8. Maintaining ECMO during patient transport.	1 (3)	16 (47)	44 (27–62)
9. Trouble shooting and managing issues with circuit/equipment.	0 (0)	15 (44)	44 (27–62)
10. Initiation of ECMO in a critical patient (impending death) in a deployed setting.	0 (0)	15 (44)	44 (27–62)

Note: Values given are number (percentage out of n = 34). Differences are statistically significant if the 95% confidence interval does not include or cross zero.

On the assessment, “competent” is defined as “able to mostly recognize and complete the skillset or problem using my own judgment and able to achieve most tasks without additional input.”

Abbreviation: ECMO, extracorporeal membrane oxygenation.

During the hands‐on procedures, blood from the carotid artery was collected at each time point: baseline, post‐ECMO initiation, and study completion, and used for ABG analysis.

Data analysis

To compare the results between the two studies, we performed the same type of analysis with the data collected from this study. We reported categorical variables as frequencies and percentages and continuous variables as means with standard deviations. The primary endpoint was a binary outcome (successful ECMO initiation, yes vs. no). Other variables of interest included successful priming and preparation of the ECMO circuit (yes vs. no), preparation time (from start of lab to completion of circuit preparation), procedure time (from start to end of each procedure, i.e., cannulation and each complication), and lab time (from start of lab to completion of lab).

We performed descriptive analyses of swine physiological parameters. We evaluated interrater agreement on the procedure instruction checklist using percent agreement over all items and teams (calculated as number of items in agreement across all teams divided by the total number of items across all teams) and Cronbach's alpha as a measure of internal consistency for the 10‐item confidence assessment. Scores from each item on the confidence assessment were summed and a percentage score was calculated out of the 50 possible points. We also calculated the proportion of students who rated themselves as “competent” (a score of 3) or better on each item. For the knowledge assessment, we calculated the percentage of correct answers (out of 20 items). Proportions of successful ECMO initiation and troubleshooting were compared to that of the teams in our previous study. Additionally, the students' comfort and knowledge assessments were compared, 1) before and after the task trainer–based ECMO training course and 2) after animal‐based vs. task trainer–based training. Descriptive summaries of participant demographics and backgrounds were conducted. Analysis of secondary outcomes included paired t‐tests (continuous variables) and McNemar tests (nominal variables) comparing pretraining scores to posttraining scores to determine improvements on the confidence and knowledge assessments. These results were compared to the findings in our live tissue–based training study using chi‐square and independent t‐tests.

Power analysis

To ensure ease of study comparison, we matched the prior sample size of 17 teams (two clinicians per animal). ¹⁹ That study can be referenced for further details of our power analysis. In brief, a precision analysis was performed a priori and determined that a sample size of 17 teams (2 clinicians per animal) could produce a 95% confidence interval of 90% to 100% with the assumption of 99.9% success rate in ECMO initiation. Analyses were performed using SAS v9.4 (SAS Institute, Inc. Cary, North Carolina). Statistical tests are two‐sided with a significance level of 5%.

RESULTS

Characteristics of study subjects

A total of 34 clinicians (17 teams) completed the ECMO course. These teams consisted of 16 EM physicians, one surgery department physician, 5 emergency department nurses, 10 intensive care unit (ICU) nurses, one research nurse, and one respiratory therapist. Most physicians (77%) and nurses/respiratory therapists (89%) had been in their current position for 5 years or fewer. Most participants had less than 10 hours of experience caring for a patient on ECMO (91%) and none had ever completed formal ECMO training prior to this course (Table 1).

Primary outcome

All 17 teams successfully primed and prepared the ECMO circuit with a mean time of 27:01 min (SD 5:06, 95% CI = 24:23–29:38 min) (Supplementary Table S2). Sixteen of the 17 teams (94%, 95% CI = 71%–100%) were able to complete the cannulation procedures (mean time from start of the lab: 65:34 min, SD 16:26, 95% CI 57:07–74:01 min). One team was unable to initiate ECMO due to an arterial laceration caused during the cannulation process. One team successfully initiated VA ECMO but lacerated the second femoral vein and was unable to transition to VAV ECMO. The mean total lab time (from start to finish) was 148:30 ± 26:11 min (SD 26:11, 95% CI = 135:00–161:54 min).

There were no significant differences in success rates between the previous live tissue–based study and the current study (Table 4). Participants were able to prime and prepare the circuit quicker in this study than in the previous study (difference of 4:26 min, 95% CI of the difference 0:26–8:26 min), but otherwise there were no differences in the time to initiate ECMO, convert to VAV ECMO, or in the total lab time.

TABLE 4.

Comparison of outcomes between live tissue study and current study

Variable	Live tissue study (n = 17)	Current study (n = 17)	Difference (95% CI)
Primary outcomes
Prime and prepare ECMO circuit	17 (100%)	17 (100%)	0%
Time to prime circuit, min	31:26 (28:10–34:42)	27:01 (24:23–29:38)	4:26 (0:26 to 8:26) a
Initiate VA ECMO	15 (88%)	16 (94%)	−6% (−30% to 19%)
Time to initiate VA ECMO, min	59:11 (44:31–73:52)	65:34 (57:07–74:01)	‐6:23 (−21:51 to 9:05)
Convert to VAV ECMO	14 (82%)	15 (88%)	‐6% (−34% to 22%)
Time to convert to VAV ECMO, min	20:59 (16:49–25:09)	17:51 (13:45–21:58)	3:08 (−2:54 to 9:09)
Total lab time, min	145:24 (132:00–158:54)	148:30 (135:00–161:54)	3:03 (−21:14 to 15:09)
Assessments
Confidence assessment
Pretest	17.3 (13.5–21.0)	17.5 (12.5–22.5)	−0.2 (−6.4 to 5.9)
Posttest	57.6 (51.3–63.9)	52.6 (45.6–59.7)	5.0 (−4.3 to 14.2)
Change from pretest to posttest	40.3 (35.6–45.0)	35.1 (29.1–41.2)	5.2 (−2.4 to 12.7)
Knowledge assessment
Pretest	64.7 (60.8–68.6)	65.3 (60.3–70.3)	−0.6 (−6.8 to 5.6)
Posttest	85.9 (82.8–88.9)	89.6 (86.0–93.1)	−3.7 (−8.3 to 0.9)
Change from pretest to posttest	21.2 (16.5–25.8)	24.3 (18.6–29.9)	−3.1 (−10.3 to 4.1)
Troubleshooting
Venous air	15 (88%)	14 (82%)	6% (−22% to 34%)
Time to troubleshoot venous air, min	4:30 (2:34–5:44)	5:32 (4:24–6:39)	−0:57 (−2:24 to 0:31)
Arterial air	15 (88%)	15 (88%)	0%
Time to troubleshoot arterial air, min	4:19 (3:23–5:52)	5:06 (4:08–5:05)	0:12 (−2:15 to 2:39)
Loss of circuit integrity	15 (88%)	16 (94%)	−6% (−30% to 19%)
Time to troubleshoot loss of integrity, min	5:00 (2:47–5:46)	5:01 (4:01–6:01)	−0:20 (−1:59 to 1:18)
Loss of power	15 (88%)	15 (88%)	0%
Time to troubleshoot loss of power, min	3:26 (2:42–3:43)	3:11 (2:43–3:40)	−0:14 (−1:06 to 0:38)
Chatter	15 (88%)	16 (94%)	−6% (−30% to 19%)
Time to troubleshoot chatter, min	1:08 (1:03–2:49)	2:04 (1:34–2:32)	9:46 (−5:08 to 24:40)

Note: Values are count (column percentage) or mean (95% confidence interval).

^a Comparisons are significant if the 95% confidence interval of the difference does not include or cross 0.

Secondary outcomes

The procedure validation checklist had 93% interrater agreement across all items and teams, and the confidence assessment showed good reliability (Cronbach's alpha = 0.92). All students showed an increase in confidence completing ECMO tasks after the course. On average, their overall confidence scores improved by 35.1% from pretest (mean 17.5%, SD 14.4%) to posttest (mean 52.6%, SD 20.1%), with 95% CI of the increase 29.1%–41.2%, p <0.0001. The proportion of students rating themselves as “competent” or “better” significantly improved for each item on the confidence assessment after the training course (Table 2). Post‐course confidence scores ranged from 44% on two items (initiation of ECMO in a critical patient in a deployed setting and trouble shooting and managing issues with circuit/equipment) to 74% (initiation of IV anticoagulation).

Students' mean knowledge assessment scores increased after training completion by 24.3% from pretest (65.3% correct, SD 14.4%) to posttest (89.6% correct, SD 10.3%), 95% CI of the increase 18.6%–29.9%, p <0.0001. Nearly all teams who successfully initiated ECMO could troubleshoot ECMO complications (Table 3). Additionally, in those 16 animals successfully placed on ECMO, ABG values and vital signs demonstrated no critical changes (Supplementary Table S3).

TABLE 3.

Results of troubleshooting complications

Complication	Teams successful (%)	Time to troubleshoot complications a
Venous air	14 (82)	5:32 [4:24–6:39]
Arterial air	15 (88)	5:06 [4:08–5:05]
Loss of circuit integrity	16 (94)	5:01 [4:01–6:01]
Loss of power	15 (88)	3:11 [2:43–3:40]
Chatter	16 (94)	2:04 [1:34–2:32]

Note: Values given are n (percentage out of n = 17) or median time in minutes:seconds [interquartile range].

^aThe single team that was unsuccessful in initiating VA ECMO was excluded from the time analysis. Teams who were unsuccessful in troubleshooting the complication were also excluded from the respective time analysis.

We found no significant differences between this study and the previous live tissue study when comparing pretest scores, posttest scores, and improvements for the confidence and knowledge assessments (Table 4). Similarly, there were no differences between the two studies with regards to the troubleshooting tasks.

DISCUSSION

Most teams receiving a brief one‐day lecture and task trainer–based hands‐on ECMO training course successfully initiated ECMO on a swine model in this follow‐on study. All teams were successful at setting up and priming the ECMO circuit and troubleshooting common ECMO complications. The current study adds to findings from our prior live tissue study by validating the use of a lecture and task trainer–based ECMO course, for use in facilities lacking live tissue training. Our study highlights the ability to train teams with emergency or critical care experience in the basics of ECMO initiation and troubleshooting. These findings offer an avenue for increased scale of ECMO training, to address the increased demand for this specialized skill set due to the COVID‐19 pandemic and the expansion of prehospital cardiac arrest ECMO. Additionally, our training program could be incorporated into further training necessary to complete the ELSO Adult ECMO Practitioner Certification [https://www.elso.org/ecmo‐eduction/certificationexam.aspx].

In this hands‐on training study, we continue to demonstrate improvement in the student's confidence level for each task evaluated. When reviewing the student's confidence assessment from the two studies, similar percent differences were found between pretest and posttest assessment for each task evaluated. However, there was some degree of variability due to the different clinicians enrolled in both studies. Proportions of procedural success of ECMO initiation and transition were comparable to the live tissue training study. The inability to initiate ECMO or transition to VAV ECMO was due to fatal arterial lacerations. This same complication was observed in our previous study. It is likely this complication can be reduced by further practicing cannulation skills for better performance. Additionally, swine anatomy makes placement of large ECMO cannulas more challenging than in humans for multiple reasons. First, the femoral vessels of 70–90 kg swine are smaller than those of human adults and follow a less linear path as they transition to the iliac vessels. Second, large muscles overlying the femoral vessels of swine results in significant resistance to placement of the large ECMO cannulas. The authors have noted vascular injury due to deviation of the cannulas path from the intended route despite use of the guide wire. Finally, swine hypercoagulability results in rapid intravascular thrombus formation if vessel puncture occurs without successful wire placement, making additional attempts to access the vessel significantly challenging. While not a hypothesis of this or the preceding study, it is the authors' opinion that obtaining central vascular access and placing cannulas in swine models is a greater challenge than doing so in human patients. Thus, the procedural success found in these studies may underestimate the effectiveness of procedural success on human patients.

All serious adverse events in both studies resulted from injuries caused during cannulation. ECMO programs should place emphasis on ensuring physicians cannulating ECMO patients have significant experience accessing the central vasculature. This is not to state that this procedure should be limited to vascular surgeons, as previous research has demonstrated high rates of successful cannulation performed by nonsurgical physicians with adequate experience. ²² Thus, the authors recommend implementing this type of training to other medicine specialties that may need to use ECMO.

With the current study, we were also able to demonstrate that a task trainer–based ECMO training produced similar increases in post‐course confidence and ability to initiate and manage ECMO in a swine model, compared to our initial study utilizing a swine model for the training phase. This suggests that task trainer–based ECMO training may provide the necessary skills for participants to initiate, prime, and troubleshoot this technique before using it on a live model. Our findings are consistent with previous research demonstrating the superiority of simulation‐based ECMO training in skill development and error elimination. ²³ , ²⁴ , ²⁵ , ²⁶ Brum et al demonstrated similar improvements where 93% of subjects reported they felt more confident in dealing with ECMO following simulation‐based training. ²⁷

Given the surge of patients with cardiopulmonary failure, it is evident there is a need to implement this type of training on clinicians that are in the front line of this pandemic so they have the necessary skills to successfully initiate ECMO procedures without delay and alleviate the already overwhelmed ECMO teams in hospitals and trauma centers. ²⁸ The authors have initiated a subsequent study using identical methods in addition to telemedicine support, to aid the team during the various ECMO stages. If the addition of telemedicine is found to be effective, the expansion of ECMO training in combination with telemedicine support from ECMO centers may allow the expansion of ECMO capabilities to areas lacking this life‐saving therapy.

LIMITATIONS

First, this study was conducted using an animal model, which is not completely representative of humans. However, the validation of the efficacy of the ECMO lectures and hands‐on task trainer–based training on humans is impractical due to logistical and ethical concerns. The swine model, with similar anatomical and physiological characteristics compared to humans, is a feasible alternative to assess the efficacy of this training model. Additionally, the authors consider ECMO cannula placement on swine more challenging than on humans for the aforementioned reasons. Second, the knowledge and confidence assessments used to validate the efficacy of the training were performed immediately after the training, so the duration of knowledge and skill retention is unknown. As with the development of other clinical knowledge and skills, the authors believe that repetition and refresher training will likely result in greater performance and skill sustainment. Additional practice of cannulation procedures could easily be incorporated into routine procedure labs and grand rounds sessions. The intent of this course is to train personnel to initiate ECMO, stabilize a critical patient, and then admit or transfer the patient to an inpatient ECMO team. This curriculum does not provide training for the long‐term management of patients on ECMO or on the weaning of ECMO therapy. Our study also did not evaluate the feasibility and costs of implementing this training program at a medical facility. Finally, our study was aimed at the emergency medicine and general surgery physicians whose respective residences require extensive procedural training. Further studies may be needed to assess successful initiation and troubleshooting of the ECMO circuit and cannulation by internal medicine physicians and intensivists who might be in the position to utilize ECMO.

CONCLUSIONS

We developed and validated an abbreviated one‐day course consisting of lectures and hands‐on task trainer–based ECMO training for a physician and nurse or respiratory therapist team. Students achieved a high rate of successful skill demonstration in a live tissue model, with increased scores on post‐course knowledge and confidence assessments, similar to our previous live tissue training program.

AUTHOR CONTRIBUTIONS

JKM: Principal Investigator, obtained funding, designed the study, developed the lectures, provided training to study participants, drafted and edited the manuscript. RMP: acquisition of data, drafted, edited and provided critical revision of the manuscript. TA: drafted and edited the manuscript. WTD: edited and provided critical revision of the manuscript. JAP: assisted in study design, and acquisition of data. MC: acquisition of data and edited the manuscript. AAA: assisted with study design, performed statistical analysis and provided critical revision of the manuscript. CAP: acquisition of data and edited the manuscript. DSB: acquisition of data and edited the manuscript. DCR; acquisition of data and edited the manuscript. KLM: acquisition of data and edited the manuscript. PN: edited and provided critical revision of the manuscript.

FUNDING INFORMATION

This project was funded by the Department of Defense, Congressionally Directed Medical Research Program (CDMRP), Joint Program Committee 6 (JPC−6) /Combat Casualty Care Research Program, Award number: DM167042, granted to Joseph K. Maddry, MD.

CONFLICT OF INTEREST

All authors declare no conflicts of interest.

ETHICS STATEMENT

The experiments reported herein were conducted according to the principles set forth in the National Institute of Health Publication No. 80‐23, Guide for the Care and Use of Laboratory Animals and the Animal Welfare Act of 1966, as amended.

Supporting information

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Maddry JK, Paredes RM, Abdurashidov T, et al. Efficacy of a single‐day task trainer–based extracorporeal membrane oxygenation training course. AEM Educ Train. 2022;6:e10806. doi: 10.1002/aet2.10806