Page 35 - JSOM Spring 2024

a total of 14 responses were obtained, like in previous stud- responses was discarded, and a final checklist with at least
ies. 15,16 All expert participants and study investigators were 70% expert consensus was yielded.
blinded to the identity of the other participants to minimize
bias.
Results
Initial Checklist Development This study was conducted from February 1 to June 1, 2023.
First, four investigators (MK, EP, DA, and RK) developed This was a blinded study; the exact specialties of the 14 respon-
an initial checklist of 36 items they felt were crucial to dents are not known. The specialty backgrounds surveyed are
ultrasound-guided bifemoral VV ECMO cannulation in oper- listed in Table 1. Each round was stopped after 14 responses
ational, resource-limited environments. This 36-item checklist were obtained. The initial checklist developed by investiga-
was created based on the experience of the investigators as well tors was distributed with a 9-point Likert scale (Table 2). No
as a literature review. Multiple guidelines were used for the items in either rounds 1 or 2 had mean values <3 (Figure 1).
development of the checklist. 12,17,18 These guidelines differed Twenty-three items from round 1 and 28 items from round
somewhat in their recommendations, and efforts were made 2 had scores ≥7 and were included. Ten items from round 1
to represent the areas of agreement in the initial checklist. No and 7 items from round 2 had values >3 but <7. These items
specific guidelines existed for resource-limited considerations were reviewed by investigators for necessity and clarity (EP,
in bifemoral VV ECMO cannulation. After the creation of the DA, and RK). One item from round 1 (item 9) and two items
initial checklist, expert review was conducted in three rounds. from round 2 (items 10, 15) were removed. During the first
two rounds, nine edits were made to the existing items (items
Round 1 1, 5, 6, 14, 16, 19, 24, 27, and 28). There were no additional
The initial checklist was distributed using Research Electronic checklist items that were added, as all edits were incorporated
Data Capture (REDCap), a secure web-based software pro- into existing items.
gram designed to support data capture for research studies. 19,20
Each expert participant received an individualized link to the TABLE 1 Specialty Background of Expert Participants (n=14)
survey by email. The participants were asked to rank each item Specialty No. of participants
on the checklist according to how important they felt it was Emergency medicine–critical care 7
for inclusion on a procedural checklist. Each item in the check-
list was ranked with a 9-item Likert-type scale used to rank Pulmonary–critical care 1
each item according to its importance: irrelevant (must be dis- Surgery–critical care 3
carded), extremely unimportant, very unimportant, unimport- Trauma surgery 7
ant, neutral, important, very important, extremely important, Vascular surgery 4
and mandatory (must be included). Cardiac surgery 2
Interventional cardiology 2
Expert participants were allowed to provide their own items
for inclusion in the checklist. Participants were also able to After three items were removed and the suggested edits were
make suggestions for item removal. Finally, participants had made, the 33-item checklist was distributed to the expert par-
an opportunity to comment on any of the items and provide ticipants to rate each item as either include or discard (round
suggested edits or necessary discussion points. 3). One item received less than 70% consensus for inclusion
(item 7) and it was discarded (Figure 2). This step resulted in a
The composite results were collected by one investigator (EP) 32-item checklist (Table 3).
who calculated mean scores for each item. The group of re-
searchers then reviewed each response. An item with a mean Discussion
score of ≤3 was discarded from the checklist. An item with
a mean score of ≥7 was included, while a mean score of <7 Using the modified Delphi method, we created a procedural
but >3 was discussed by the investigators to reach a consensus performance checklist based on expert consensus for VV
regarding exclusion, inclusion, or the need for modification. ECMO cannulation in austere environments. To our knowl-
These items were edited as indicated by participants’ responses edge, this is the only checklist currently available for this
and were included or excluded based on the investigators’ procedure, and it will help standardize training for bifemoral
consensus. Any additional items created by participants were cannulation in resource-limited situations.
added to the checklist for inclusion in the second round.
Specific Inclusions, Exclusions, and Edits
Round 2 The investigators removed three items. First, local anesthesia
Round 2 of the survey was then sent to the expert participants at the cannula insertion site was removed (item 9). Though
until 14 responses were collected. They were again asked to regional anesthesia is important, patients requiring VV ECMO
rank each item according to the above scale. As in round 1, in the operational environment will likely be unresponsive and
21
participants had the option to include their items, make sug- receiving intravenous analgesia and sedation. Local anesthe-
22
gestions for removal, and make edits to the existing items. The sia may also not be readily available. Given the variability
results were similarly reviewed, and items were edited accord- of the availability and use of local anesthesia, this item was
ing to participants’ comments. removed. Next, site-specific cannula assignments was removed
(item 10). Though the right femoral vein may be the preferred
Round 3 location for the return cannula, given the inferior vena cava
Finally, a third version of the checklist was sent to the partic- anatomy and linear approach to the right atrium, this place-
ipants for review. In this version, experts classified an item as ment was not essential to the performance of cannula place-
“include” or “discard.” Any item with more than 30% discard ment so it was removed. Finally, placement of the return
18
Checklist for Bifemoral VV ECMO Cannula Placement in Operational Environments | 33