Page 37 - JSOM Winter 2025

Methods the reliability of our findings, bones with visible fractures or
damage were discarded after visual inspection.
The study was conducted in accordance with the Declaration
of Helsinki and approved by the Institutional Review Board Our methodological approach adheres to ethical and regulatory
of East Carolina University (UMCIRB 22-000255) for stud- standards in animal research and provides comparable insights
ies involving humans. Informed consent was obtained from all into IO catheter placement techniques. Both devices instruct
16
subjects involved in the study. the user to either “Gently press needle through the skin until
the tip touches the bone” (EZ-IO) or “insert needle assembly
17
Study Design and Setting through the skin and adipose tissue, touching the periosteum
This prospective randomized comparative study was per- before attempting IO insertion” (SAM IO). The exposed por-
18
formed at several EMS stations, and EMS clinicians (Ad- cine bone presented a suitable fidelity model since insertion
vanced EMTs and paramedics) were recruited. A convenience started with bone contact. Only the needle insertion time was
sample was obtained from EMS clinicians who volunteered measured to isolate each driver’s needle insertion times through
to participate in a county EMS system comprised of one ur- the bone (differentiating from the assembly times and stabi-
ban fire-based agency, multiple rural contracted agencies, and lization procedures, which could be skewed based on device
one municipal third service agency that responds to 36,000 familiarity), and stabilization procedures were similar.
requests for service annually.
Outcome Measures
Participants watched instructional videos produced by EZ-IO Our primary outcome was the mean IO needle insertion time
and SAM IO demonstrating device use, then were given between EZ-IO and SAM IO. Time was measured in sec-
hands-on instruction and practice time with each driver and onds (s) from the start of trigger actuation until the partic-
needle size. Participants practiced insertions with both IO de- ipant informed the researcher they were finished. All of the
vices, in no particular order, based on their selection, of both post- training attempts IO insertions (100%) were confirmed
a simulated plastic bone task trainer and a porcine bone until by saline flush, and visual inspection by the researcher were
they self-reported proficiency. After completion of the practice deemed successful. Success was determined by the participant
session, each participant was randomized to either the EZ-IO flushing 1–3mL of fluid through the inserted IO needle and vi-
or SAM IO driver. Needle size selection was based on avail- sual confirmation by the researcher that the saline flush passed
ability and was not formally randomized. At the end of the through the bone marrow. Secondary outcomes compared nee-
session, participants were asked to complete a survey using dle size (length), credential level, years of experience, sex, and
Qualtrics to gather opinions regarding the use of each of the prior experience with SAM IO with driver insertion times. A
devices. The survey was constructed by EMS knowledge ex- Qualtrics survey platform using a five-point Likert scale col-
perts, then field tested and revised before use. lected participant feedback data on both devices. 19

Survey Design Statistical Analysis
The researchers anticipated that most, if not all, of the partic- Data was analyzed with IBM SPSS 28 (IBM Corp, Armonk,
ipants currently used the EZ-IO exclusively. However, some NY) using a linear mixed-effects model and a 2×3 factorial
may have had some training experience with the SAM IO. ANOVA with a Bonferroni correction for pairwise compar-
Questions were developed to determine user viewpoints for isons to examine the effects of the device types EZ-IO and
either device, but they were mainly used to compare the SAM SAM IO on needle insertion times. Age, experience, and prior
IO to the most used EZ-IO. The eight-question survey was de- use were also analyzed as contributing factors represented by
veloped by a group of experienced EMS clinicians, including frequency and descriptive statistics; the frequency of responses
board-certified EMS physicians. In a pilot trial, six paramed- represented survey results.
ics provided feedback, interpretation of questions, and revi-
sion. The study purposefully did not compare the assembly
time and flow rates since the 15mm, 25mm, and 45mm needle Results
length and diameter specifications were identical: 15 gauge One hundred and fifteen EMS clinicians (103 paramedics and
and 1.8mm. 12 Advanced EMTs) consented, participated in a comparison
study, and completed the survey, with none declining to partic-
Model Selection ipate. Data was inspected by histogram and boxplot for nor-
Using the principles of Refine, Reduce, or Reuse in animal re- mality and distribution. Visualization of the data histogram
search, this study investigated IO catheter placement and used and a significant Levene’s test (P<.001) for equality of vari-
fresh, commercially processed porcine humerus bones. Based ances was significant, demonstrating that the data between
on previous research, porcine bones were obtained from 70 groups was not homogeneous.
to 90kg porcine specimens, indicating their bone density ap-
proximates that of 20 to 39-year-old adult patients. 14,15 Our The entire intent-to-treat cohort was analyzed with a lin-
basis for choosing porcine humerus bones is their compara- ear mixed-effects model sensitive to the data distribution to
bility with transfusion rates and medication efficacy, offering examine the effect of driver type on IO insertion time while
higher-fidelity models than simulated training bones for com- controlling for needle size, sex, age, years of experience, and
parative analyses. The porcine bones were procured from a credential level. A random intercept for subject ID was included
licensed North Carolina Department of Agriculture and Spe- to account for between-subject variability. The model revealed
cial Services meat processing facility and were acquired fol- a significant main effect of driver type, F =9.8, P=.002. Es-
1, 107.0
lowing state permit regulations. Upon harvesting, the bones timated marginal means showed that the EZ-IO driver had a
were promptly stored in a cooler on ice. The bone specimens mean insertion time of 1.2s (95% CI 1.0–1.4), compared to the
were brought to room temperature before study use. To ensure SAM IO driver with a mean of 2.4s (95% CI 2.2–2.7).

SAM IO versus EZ-IO: Insertion Time and Usability | 35