Device-led versus human-led feedback on chest compressions for 1 cardiopulmonary resuscitation and providers’ experience and 2 preference: a randomised crossover study 3

Background: High cardiopulmonary resuscitation (CPR) quality is associated with better 31 patient survival from cardiac arrest. However, CPR providers may not have an accurate 32 perception of the depth and rate of their chest compressions (CC). Realtime feedback during 33 resuscitation improves CPR quality compared to no feedback. Evidence comparing audio-visual 34 feedback device (AVF) and team leader’s feedback (TLF) in improving CPR performance is 35 limited and conflicting. randomized crossover study to evaluate CC performance with TLF. CPR CPR-sensing

participants were randomized in a 1:1 allocation ratio into AVF-led CC followed by TLF-led CC 1 4 1 (Group 1), or TLF-led CC followed by AVF-led CC (Group 2). We allocated a 10-minute 1 4 2 washout interval between intervention periods to minimize participant fatigue. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 27, 2022. ;https://doi.org/10.1101/2022 doi: medRxiv preprint 6 1 4 4 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 27, 2022. ;https://doi.org/10.1101/2022 doi: medRxiv preprint 7 Equipment 1 4 5 The CPR manikin used was Little Anne™ (Laerdal, Orpington, UK) with a spring constant of 1 4 6 4.46 kg/cm and a 22.3 kg-force required to press 5 cm and maximum compression depth of 7 guidelines. The CPR Dashboard™ using Real CPR Help® technology in the device provided 1 5 0 real-time AVF on rate, depth, and release of each compression via an accelerometer. The CPR manikin was placed on the floor to avert mattress compressibility as a confounding 1 5 4 variable. Participants were instructed to perform CC as they would usually perform in real 1 5 5 practice, but with a small CPR sensor placed underneath their hands on the manikin's chest. We collected participants' demographic data comprising of age, gender, and profession. The 1 6 5 primary outcome measured were CC rate (min -1 ) and depth (cm). We extracted the CC rate and 1 6 6 depth data stored in the built-in memory storage of the defibrillator using the ZOLL CodeNet regarding their perception and experience. Each interview was audio-recorded and later 1 7 0 transcribed for analysis to look for common themes. We also asked if they felt they performed 1 7 1 better with AVF or with TLF, and their preferred method of feedback during resuscitation. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint An earlier 2 x 2 crossover study of CC with and without feedback device showed mean chest 1 7 5 compression rates of 113 min -1 (± 7) and 113 min -1 (± 13) respectively (17). Based on this and 1 7 6 the AHA-recommended CC target rate range of 100 to 120 min -1 , we defined the minimal 1 7 7 clinically significant difference in CC rate as 7 min -1 for this study (since difference of >7 min -1 1 7 8 would exceed the recommended CC target rate). Using sample size calculation for two-1 7 9 intervention crossover study, a total of 57 participants will have 80% power to detect a to include an estimated 20% dropout rate. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint students (20%). The comparison of CC rate and depth between TLF-led CC and AVF-led CC is shown in Table   2 0 3 1. The mean CC rate was higher in AVF-led CC compared with TLF-led CC (121.75 ± 17.66 2 0 4 min -1 vs. 117.43 ± 13.45 min -1 , p = 0.005). The mean CC rate with AVF was also slightly above 2 0 5 the recommended target rate (>120 min -1 ). There was no significant difference between AVF-  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint  categorized into within target and not within target. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint Those who preferred AVF generally cited "more objective and straightforward feedback" as the 2 9 2 reason. need to improve -the rate, the depth, I know exactly how much I need to improve.

9 6
Whereas with team leader, while it is also helpful, when he or she says push harder, but 2 9 7 I am not seeing how much exactly I need to improve on." . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 27, 2022. ; https://doi.org/10.1101/2022.09.25.22280283 doi: medRxiv preprint 1 5 Participants' perceived performance and preference, and chest compression quality Participants who perceived that they performed better with TLF, and those who preferred TLF, 3 0 1 were noted to perform CC at a mean rate significantly higher and above the recommended 3 0 2 target rate when performing with AVF (Table 3). Participants who perceived that they performed 3 0 3 better with AVF, and those who preferred AVF, showed similar CC performance during AVF and  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The 2015 AHA guideline states that high quality CC requires the following: (1) optimal hand 3 1 7 position, (2) compressing the lower part of the sternum by at least one-third of the anterior- achieving compression rate of 100 to 120 min -1 , and (4) allowing for complete chest recoil 3 2 0 between each CC. In this study, we focused on the CC rates and depth as these parameters may not be considered significant in real clinical settings. The proportion of participants performing CPR at CC rates within the recommended target 3 2 8 range of 100 to 120 min-1 was similar between AVF and TLF (48.6% and 51.4%, p = 0.824).

2 9
However, the mean CC depth with both TLF and AVF in this study were below the 3 3 0 recommended range, while the mean CC rate inclined toward 120 min -1 . This corroborated a 3 3 1 previous study reporting significant decrease in CC depth as the CC rate increases (25). With AVF, suboptimal compression depth can be due to difficulty in following the audio-visual . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint On the other hand, TLF would reduce the extraneous load as participants need to only listen to This suggests that team leaders had conceptual and habitual tacit knowledge of the appropriate 3 5 2 CC rate. Tacit knowledge is the implicit knowledge that one possesses based on personal 3 5 3 experience (30). It is personal, intuitive, and difficult to be coded, transferred, or taught (31, 32).

5 4
Schemata on how tacit knowledge and habitual practices influence the management of 3 5 5 resuscitation in the ED and other departments have been provided in previous studies (33, 34).

5 6
Interestingly, assessment and feedback on the CC depth by team leaders in this study were not 3 5 7 as accurate as that on CC rate. This may be due to the misidentification of CC depth as 3 5 8 adequate at higher compression rates (35).