American Journal of Emergency Medicine 33 (2015) 993–997
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Original Contribution
Does accelerometer feedback on high-quality chest compression improve survival rate? An in-hospital cardiac arrest simulation☆,☆☆,★ Min Hee Jung, MD a, Je Hyeok Oh, MD b,⁎, Chan Woong Kim, MD, PhD b, Sung Eun Kim, MD, PhD b, Dong Hoon Lee, MD, PhD b, Wen Joen Chang, MD, PhD a a b
Department of Emergency Medicine, Sungae Hospital, Seoul, Republic of Korea Department of Emergency Medicine, College of Medicine, Chung-Ang University, Seoul, Republic of Korea
a r t i c l e
i n f o
Article history: Received 3 April 2015 Received in revised form 8 April 2015 Accepted 9 April 2015
a b s t r a c t Objective: We investigated whether visual feedback from an accelerometer device facilitated high-quality chest compressions during an in-hospital cardiac arrest simulation using a manikin. Methods: Thirty health care providers participated in an in-hospital cardiac arrest simulation with 1 minute of continuous chest compressions. Chest compressions were performed on a manikin lying on a bed according to visual feedback from an accelerometer feedback device. The manikin and accelerometer recorded chest compression data simultaneously. The simulated patient was deemed to have survived when the chest compression data satisfied all of the preset high-quality chest compression criteria (depth ≥51 mm, rate N100 per minute, and ≥95% full recoil). Survival rates were calculated from the feedback device and manikin data. Results: The survival rate according to the feedback device data was 80%; however, the manikin data indicated a significantly lower survival rate (46.7%; P = .015). The difference between the accelerometer and manikin survival rates was not significant for participants with a body mass index greater than or equal to 20 kg/m 2 (93.3 vs 73.3%, respectively; P = .330); however, the difference in survival rate was significant in participants with body mass index less than 20 kg/m2 (66.7 vs 20.0%, respectively; P = .025). Conclusions: The use of accelerometer feedback devices to facilitate high-quality chest compression may not be appropriate for lightweight rescuers because of the potential for compression depth overestimation. Trial registration: Clinical Research Information Service (KCT0001449). © 2015 Elsevier Inc. All rights reserved.
1. Introduction The 2010 International Consensus on Cardiopulmonary Resuscitation (CPR) and Emergency Cardiovascular Care Science With Treatment Recommendations outlined the conditions necessary for high-quality chest compression as a compression depth of at least 2 inches (5 cm) at a rate of at least 100 compressions per minute, full chest recoil, and minimal interruptions [1]. In particular, maintaining a compression depth of more than 5 cm during CPR is associated with higher survival rates in adult and pediatric cardiac arrest patients [2,3]. Therefore, a device that provides feedback on chest compression depth, chest wall recoil, and compression rates has been developed and shown to facilitate the performance of high-quality chest compression under simulated cardiac arrest conditions [4,5]. However, in 2009, Perkins et al [6] reported that accelerometer feedback devices could overestimate compression depth when compressions ☆ Financial support: none. ☆☆ Conflict of interest statement: On behalf of all authors, the corresponding author states that there is no conflict of interest. ★ Funding: none. ⁎ Corresponding author at: Department of Emergency Medicine, Chung-Ang University Hospital, 102 Heukseok-ro, Dongjak-gu, Seoul, Republic of Korea 156-755. Tel.: +82 2 6299 1820; fax: +82 2 6299 2558. E-mail address: [email protected] (J.H. Oh). http://dx.doi.org/10.1016/j.ajem.2015.04.016 0735-6757/© 2015 Elsevier Inc. All rights reserved.
were performed on a soft surface. Such overestimation is likely to occur during in-hospital CPR because most cardiac arrest patients are lying on a bed. Although Oh et al [7] reported in 2012 that a dual accelerometer could prevent overestimation of compression depth, the dual accelerometer technique is not currently used in clinical settings. We used a manikin in an in-hospital cardiac arrest simulation to investigate whether visual feedback from an accelerometer device facilitated the performance of high-quality chest compressions.
2. Material and methods 2.1. Study design A prospective, nonrandomized single trial was carried out, with continuous chest compressions performed during 1 minute after regular CPR education at our hospital (Fig. 1). The study was approved by the institutional review board of our hospital (approval number: SA2015-07).
2.2. Study setting and study population A total of 30 health care providers who worked in the emergency department of a community hospital participated in the study.
994
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
PC SkillReporting System (Laerdal Medical) was used to collect chest compression data from the manikin. A CPRmeter (Laerdal Medical) was used to provide chest compression feedback. The CPRmeter provided visual feedback on compression depth, rate, and full recoil; and Q-CPR Review software version 3.1 (Laerdal Medical) was used to collect the data. 2.4. Study protocol
Fig. 1. Study flow diagram.
All participants were recruited voluntarily during regular CPR education and gave their verbal informed consent before the simulations. In-hospital cardiac arrest simulations using a manikin were conducted in the simulation center of the hospital after CPR training. Chest compression data were collected during the simulation. The sample size was calculated based on the mean compression depth (MCD) determined in a previous study (52.6 ± 6.7 mm) [8]. Given the SD of 6.7 mm, the expected MCD difference between the accelerometer device and manikin was postulated to be 10%, and the allowable difference was set at 5.26 mm. The 2-sided significance level was set at 0.05 with statistical power of 80%. The minimum number of participants was determined to be 13 using a Web-based program (sample size calculator: 1 sample mean) [9].
2.3. Study materials The following devices were used during the simulation. The Stretcher Trolley–Paramount Model: KK-728E (Paramount Bed, Tokyo, Japan) was used for the in-hospital cardiac arrest simulation. We used a step stool to control for differences between bed height and rescuer knee height [10]. We measured the knee heights of 26 adults who did not participate in the study. The mean height of the rescuers' knees was 45.7 ± 3.0 cm (range, 39-52 cm); therefore, we set the difference between the bed and step stool height to 45 cm. The difference between bed and knee height could be adjusted within 10 cm using this method. A Resusci Anne SkillReporter manikin (Laerdal Medical, Stavanger, Norway) was used as the simulated cardiac arrest patient. The Laerdal
Cardiopulmonary resuscitation training, which did not include practice using a manikin, was conducted for 1 hour with a focus on the requirements for high-quality chest compression. Previous investigations showed that chest compression quality was significantly lower when CPR was performed by lightweight rescuers than when performed by heavier rescuers; therefore, we recorded the participants' height and body weight before the simulation [11,12]. In addition, we recorded the participants' sex, age, and knee height (distance from the floor to the tibial tuberosity in the erect position) [8]. The in-hospital cardiac arrest simulation protocol was as follows. The manikin was placed on an emergency department bed in the supine position. Continuous chest compressions were performed according to the visual feedback provided by the accelerometer for 1 minute without ventilation assuming that an advanced airway was in place. The participants practiced CPR for 1 minute before the simulation to familiarize themselves with the visual feedback provided by the accelerometer during chest compressions. The investigator supervising data collection used a stopwatch to accurately instruct the participants when to start and stop chest compressions. The practice was limited to 1 minute because rescuer fatigue and muscle strength affect the quality of chest compression [11-13]. The bed-to-step stool height was adjusted by measuring the height of the step stool (x cm), and the height of the bed (distance from the floor to the upper surface of the mattress) was then adjusted to (x + 45) cm. The same step stool and bed were used for all simulations and were fixed at the same height. The mattress used was provided by the bed manufacturer. According to the 2010 International Consensus on CPR, there is insufficient evidence to argue for or against the use of backboards during CPR [1]. Therefore, we did not use a backboard in our simulations. The chest compression data were collected simultaneously by the accelerometer feedback device and manikin. 2.5. Outcome variables The simulated patient was deemed to have survived when the chest compression data collected during the simulation satisfied all of the preset high-quality chest compression criteria. Conversely, the patient was considered to have died when not all of the criteria were met. Our criteria were based on the recommendations of the 2010 International Consensus on CPR for high-quality chest compressions and included MCD greater than or equal to 51 mm, mean compression rate (MCR) greater than or equal to 100 compressions per minute, and greater than or equal to 95% full chest recoil [1]. The “minimizing interruptions” recommendation was excluded because chest compression was continuous in our simulation. The acceptable percentage of compressions with full chest recoil was at the investigator's discretion because the recommended proportion was not stated in the 2010 International Consensus. Survival rates were calculated from both the feedback device and manikin chest compression data. Survival rate was the primary outcome variable. The secondary outcome variables were the 5 chest compression indices common to the feedback device and the manikin: MCD (millimeters), MCR (counts per minute), adequate rate (percent), adequate depth (percent), and complete release (percent). As chest compressions performed by lightweight rescuers are shallower than those performed by heavier rescuers, we assessed the outcome variables according to the participants' body mass index
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
995
(BMI) by comparing the normal (≥20 kg/m2) and low (b20 kg/m2) BMI groups [11,12]. Furthermore, we collected data regarding corrected hand position on the manikin (percent), good compressions (percent), and perfusion time (percent) from the feedback device. 2.6. Statistical analyses PASW Statistics 18.0 (SPSS Inc, Chicago, IL) was used to conduct all statistical tests. Continuous variables are expressed as means ± SD. The categorical data are expressed as percentages. The Fisher exact test was used for statistical comparisons between the survival rates calculated from the feedback device and manikin. The Wilcoxon signed rank test was used to compare the compression data from the feedback device and the manikin. The Mann-Whitney U test was used to compare the compression data of the normal and lightweight groups. In all analyses, P b .05 was taken to indicate statistical significance. 3. Results 3.1. Participant characteristics The study group included 30 health care providers (2 male, 28 female) with a mean age of 28.4 ± 5.2 years, mean height of 162.3 ± 6.7 cm, mean body weight of 56.4 ± 11.0 kg, and a mean BMI of 21.3 ± 3.0 kg/m2. The mean participant knee height was 42.2 ± 2.2 cm, and the mean difference between bed and participant knee height was –2.8 ± 2.2 cm. 3.2. Comparison of feedback device and manikin survival rates The chest compression data from the feedback device and manikin are shown in Table 1. Survival rate calculated from the accelerometer data was significantly higher than that from the manikin (80% [24/30] vs 46.7% [14/30], respectively; P = .015) (Fig. 2). 3.3. Comparison of survival rates according to BMI The chest compression data according to BMI are shown in Tables 2 and 3. The calculated survival rates of the feedback device and manikin did not differ significantly among participants with BMI greater than or equal to 20 kg/m 2 (93.3% [14/15] vs 73.3% [11/15], respectively; P = .330). However, in participants with BMI less than 20 kg/m2, the survival rate was significantly higher for the accelerometer than for the manikin (66.7% [10/15] vs 20.0% [3/15], respectively; P = .025) (Fig. 3). Moreover, although the survival rates calculated from the feedback device did not differ significantly between the normal and low BMI groups (93.3 [14/15] vs 66.7% [10/15], respectively; P = .169), the manikin survival rates were significantly different between the normal (73.3% [11/15]) and low (20.0% [3/15]) BMI groups (P = .009) (Fig. 3).
Table 1 Comparisons of the chest compression data between the feedback device and the manikin Feedback device (n = 30) Manikin (n = 30) P MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Corrected hand position (%) Good compression (%) Flow time (%)
61.1 ± 7.6 114.0 ± 7.5 76.8 ± 31.2 91.0 ± 15.5 95.0 ± 12.1 N/A 67.1 ± 33.2 98.8 ± 0.8
Abbreviation: N/A, not applicable. A P value b .05 is presented in bold.
50.7 ± 5.4 114.7 ± 7.4 76.3 ± 30.9 55.9 ± 40.2 100.0 ± 0.0 100.0 ± 0.0 N/A N/A
.000 .109 .049 .000 .001 N/A N/A N/A
Fig. 2. Comparison of feedback device and manikin survival rates.
4. Discussion The 2010 International Consensus on CPR and Emergency Cardiovascular Care Science With Treatment Recommendations noted the potential for overestimation of compression depth when using feedback devices on a soft surface [1]. Despite various attempts, no devices that address this shortcoming are commercially available [7,14-16]. Therefore, standard feedback devices continue to be used. We used in-hospital cardiac arrest simulation to determine whether an accelerometer feedback device facilitated the performance of highquality chest compression. We chose the CPRmeter feedback device because it collects chest compression data and provides visual feedback for compression depth, rate, and full recoil. Although the CPRmeter has been assessed in previous simulation studies, these studies were performed with the surrogate victim on the floor [4,5]. Therefore, we reassessed the feedback device in the in-hospital environment where compressions are most likely to be performed with the patient on a bed. We found that although the accelerometer indicated high-quality compressions, the device overestimated the chest compression depth (Table 1). Thus, whereas the proportion of adequate depth compressions recorded by the feedback device was high, those recorded by the manikin were significantly lower (91.0% ± 15.5% vs 55.9% ± 40.2%, respectively; P = .000). Moreover, the calculated survival rate was overestimated by the feedback device compared with the manikin (80% and 46.7%, respectively; P = .015) (Fig. 2). A total of 43.3% (13/30) of participants met the high-quality chest compression criteria according to both the feedback device and manikin data. This finding indicates that although some participants were able to perform high-quality chest compression despite overestimation of compression depth, more than half were unable to do so. We hypothesized that the difference was attributed to differences in body weight. A previous study showed that the force needed to compress the victim's chest was affected by the rescuer's body weight, such that heavier rescuers used less muscle strength to perform chest compressions than did lightweight rescuers. Thus, the MCD of the lightweight group decreased significantly over the CPR period [11].
Table 2 Comparisons of the chest compression data collected by the feedback device between the normal BMI group (≥20 kg/m2) and the light BMI group (b20 kg/m2)
MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Good compression (%) Flow time (%)
BMI ≥20 kg/m2 (n = 15)
BMI b20 kg/m2 (n = 15)
P
64.8 ± 6.8 112.2 ± 8.2 77.9 ± 34.8 98.1 ± 1.6 96.3 ± 11.3 70.5 ± 35.3 98.8 ± 1.1
57.5 ± 6.7 115.8 ± 6.5 75.8 ± 28.3 83.9 ± 19.7 93.6 ± 13.1 63.6 ± 31.7 98.7 ± 0.5
.007 .116 .389 .004 .217 .412 .285
A P value b .05 is presented in bold.
996
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
Table 3 Comparisons of the chest compression data collected by the manikin between the normal BMI group (≥20 kg/m2) and the light BMI group (b20 kg/m2)
MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Corrected hand-position (%)
BMI ≥20 kg/m2 (n = 15)
BMI b20 kg/m2 (n = 15)
P
53.3 ± 4.8 113.0 ± 8.3 76.5 ± 34.6 75.3 ± 36.0 100.0 ± 0.0 100.0 ± 0.0
48.1 ± 4.9 116.4 ± 6.4 76.0 ± 28.0 36.5 ± 35.2 100.0 ± 0.0 100.0 ± 0.0
.008 .217 .595 .006 1.000 1.000
A P value b .05 is presented in bold.
We found no significant difference between survival rates calculated from the feedback device and manikin in the normal BMI group (Fig. 3). However, the accelerometer and manikin survival rates were significantly different in the low BMI group (P = .025). Moreover, the survival rate calculated from the accelerometer data did not differ significantly according to BMI (Fig. 3), whereas the survival rate calculated from the manikin data was significantly lower in the low BMI group (P = .009). Our results have several implications. First, despite overestimation of compression depth, 43.3% of the participants were able to perform high-quality chest compressions using the accelerometer feedback, particularly those with a normal or high BMI. Second, the rescuers with a low BMI could not perform high-quality chest compressions because the feedback device overestimated the compression depth. Thus, the development of a feedback device that can measure compression depth accurately is essential. In lieu of such a device, altering the recommended compression depth (ie, 50 mm) to account for expected overestimation may provide a solution. For example, if an overestimation of 10 mm is expected, the recommended compression depth for a feedback device should be increased to 60 mm. However, this method may not be practical, as the degree of overestimation is likely to differ according to mattress and bed types as well as rescuer height and weight. Our findings are significant in that we confirmed that the overestimation of compression depth by feedback devices is associated with the rescuer's body weight. In particular, we established that feedback indices may be inappropriate for lightweight rescuers. The potential solution to developing an effective feedback device for use by any rescuer, regardless of their height, weight, and sex, is addressed by Maier et al [17]. If the upper body weight of the rescuer can be accurately enough estimated, the feedback device can be programmed to prompt the proper compression rate for each rescuer.
Men have greater upper body weight than women because they have more muscle mass; therefore, separate estimates must be made for men vs women. So that the feedback device can be used from the kneeling position on the floor or ground, a stool that was originally designed for kneeling during CPR provided on an emergency department bed, or for that matter, any elevated surface [18], can be redesigned for use on the floor or ground and be made adjustable in height, so that each rescuer can kneel at the optimal height, so that they can exert the maximal percentage of their upper body weight. Furthermore, to maximize force that is applied directly over the compression point of the lower half of the sternum, the right hand of the rescuer should be in contact with the sternum when standing or kneeling at the right side of the patient and vice versa [19], and an inclined step stool should be used [20]. When possible, the rescuer can choose the hand/body position they are most comfortable with. In addition, rescuers need to prompt proper rate and effective depth high-impulse CPR using 2 metronomes, 1 for compression and 1 for decompression [21]. Chest compressions with a hand position at the sternoxiphoid junction minimize the force requirements for effective compression and maximize compression over the left ventricular inflow tract [19,22-24]. However, whether it is best to perform high-impulse CPR at a slower rate according to the weight of the patient [25,26], or highimpulse CPR at a fast rate [17] requires further study in man. In addition, compression of the abdomen increases the risk of internal organ injury such as liver laceration [27]. Therefore, safety issue should be solved if we could use the sternoxiphoid junction as an anatomic landmark for chest compressions. In children, to avoid compression over the liver, compression should be performed with a hand position at the internipple line, which requires much more force [19,23,28]. Therefore, compressions should be performed with a 2-hand technique by someone with small hands, with the right hand in contact with the sternum. The rescuer is at the left side of the patient and vice versa, instead of right hand in contact with the sternum with rescuer at right side and vice versa, as it is recommended for adults [19,23]. The vertical 2-thumb technique [29] with more rapid and shallower compressions [25,26] might be helpful for infants (and some small children). Our study had several limitations. First, we used mechanical models, and therefore, the level of evidence was low. Second, most participants were female. Moreover, the mean BMI was low (21.3 ± 3.0 kg/m 2) because the BMI was less than 20 kg/m 2 in half of the participants. Third, we defined survival rate according to the recommended criteria for high-quality chest compression; however, the survival rate of cardiac arrest patients may not rely on these criteria. Fourth, our study did not include a nonfeedback group, so we were unable to evaluate
Fig. 3. Comparisons of survival rates according to BMI.
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
the positive effects of using a feedback device despite the potential overestimation of compression depth. Further studies to address this issue are warranted. 5. Conclusions Accelerometer feedback devices do not provide accurate information when compressions are performed on a bed. The proportion of actual high-quality chest compressions was significantly lower than that indicated by the device, particularly for rescuers with BMI less than 20 kg/m 2. Our findings indicate that the use of visual feedback from accelerometer devices to perform high-quality compressions is not appropriate for lightweight rescuers performing in-hospital CPR. Acknowledgments The authors thank all health care providers for their contribution in this study. References [1] Koster RW, Sayre MR, Botha M, Cave DM, Cudnik MT, Handley AJ, et al. Part 5: adult basic life support: 2010 International consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Resuscitation 2010;81(Suppl. 1):e48–70. [2] Vadeboncoeur T, Stolz U, Panchal A, Silver A, Venuti M, Tobin J, et al. Chest compression depth and survival in out-of-hospital cardiac arrest. Resuscitation 2014;85:182–8. [3] Sutton RM, French B, Niles DE, Donoghue A, Topjian AA, Nishisaki A, et al. 2010 American Heart Association recommended compression depths during pediatric in-hospital resuscitations are associated with survival. Resuscitation 2014;85: 1179–84. [4] Skorning M, Beckers SK, Brokmann JCh, Rortgen D, Bergrath S, Veiser T, et al. New visual feedback device improves performance of chest compressions by professionals in simulated cardiac arrest. Resuscitation 2010;81:53–8. [5] Bulleon C, Parienti JJ, Halbout L, Arrot X, De Facq Regent H, Chelarescu D, et al. Improvement in chest compression quality using a feedback device (CPRmeter): a simulation randomized crossover study. Am J Emerg Med 2013;31:1457–61. [6] Perkins GD, Kocierz L, Smith SC, McCulloch RA, Davies RP. Compression feedback devices over estimate chest compression depth when performed on a bed. Resuscitation 2009;80:79–82. [7] Oh J, Song Y, Kang B, Kang H, Lim T, Suh Y, et al. The use of dual accelerometers improves measurement of chest compression depth. Resuscitation 2012;83:500–4. [8] Oh JH, Kim CW, Kim SE, Lee SJ, Lee DH. Comparison of chest compressions in the standing position beside a bed at knee level and the kneeling position: a nonrandomised, single-blind, cross-over trial. Emerg Med J 2014;31:533–5. [9] Centre for Clinical Research and Biostatistics. The Chinese University of Hong Kong. Sample size estimation. [Cited 11 February 2015.] Available from URL http://www. cct.cuhk.edu.hk/stat/index.htm. [10] Cho J, Oh JH, Park YS, Park IC, Chung SP. Effects of bed height on the performance of chest compressions. Emerg Med J 2009;26:807–10.
997
[11] Hasegawa T, Daikoku R, Saito S, Saito Y. Relationship between weight of rescuer and quality of chest compression during cardiopulmonary resuscitation. J Physiol Anthropol 2014;33:16. [12] Russo SG, Neumann P, Reinhardt S, Timmermann A, Niklas A, Quintel M, et al. Impact of physical fitness and biometric data on the quality of external chest compression: a randomized, crossover trial. BMC Emerg Med 2011;11:20. [13] Ock SM, Kim YM, Chung Jh, Kim SH. Influence of physical fitness on the performance of 5-minute continuous chest compression. Eur J Emerg Med 2011;18:251–6. [14] Kang MK, Oh JH, Kim TW. Chest compression depth measurement using IRUWB for CPR (cardiopulmonary resuscitation). Proc IEEE APCCAS 2010. [Cited 15 February 2015.] Available from URL http://ieeexplore.ieee.org/xpl/articleDetails. jsp?arnumber=5774952. [15] Beesems SG, Koster RW. Accurate feedback of chest compression depth on a manikin on a soft surface with correction for total body displacement. Resuscitation 2014;85: 1439–43. [16] Wutzler A, Bannehr M, von Ulmenstein S, Loehr L, Forster J, Kuhnle Y, et al. Performance of chest compressions with the use of a new audio-visual feedback device: a randomized manikin study in health care professionals. Resuscitation 2015;87: 81–5. [17] Maier GW, Tyson Jr GS, Olsen CO, Kernstein KH, Davis JW, Conn EH, et al. The physiology of external cardiac massage: high-impulse cardiopulmonary resuscitation. Circulation 1984;70:86–101. [18] Oh J, Chee Y, Lim T, Cho Y, Kim IY. Chest compression with kneeling posture in hospital cardiopulmonary resuscitation: a randomised crossover simulation study. Emerg Med Australas 2014;26:585–90. [19] Baubin M, Kollmitzer J, Pomaroli A, Kraincuk P, Kranzl A, Sumann G, et al. Force distribution across the heel of the hand during simulated manual chest compression. Resuscitation 1997;35:259–63. [20] Yun SW, Lee BK, Jeung KW, Park SW, Choi SS, Lee CH, et al. The effect of inclined step stool on the quality of chest compression during in-hospital cardiopulmonary resuscitation. Am J Emerg Med 2014;32:851–5. [21] Chung TN, Bae J, Kim EC, Cho YK, You JS, Choi SW, et al. Induction of a shorter compression phase is correlated with a deeper chest compression during metronomeguided cardiopulmonary resuscitation: a manikin study. Emerg Med J 2013;30: 551–4. [22] Cha KC, Kim YJ, Shin HJ, Cha YS, Kim H, Lee KH, et al. Optimal position for external chest compression during cardiopulmonary resuscitation: an analysis based on chest CT in patients resuscitated from cardiac arrest. Emerg Med J 2013;30:615–9. [23] Shin J, Rhee JE, Kim K. Is the inter-nipple line the correct hand position for effective chest compression in adult cardiopulmonary resuscitation? Resuscitation 2007;75: 305–10. [24] Pickard A, Darby M, Soar J. Radiological assessment of the adult chest: implications for chest compressions. Resuscitation 2006;71:387–90. [25] Rottenberg EM. The critical need for further research and development of abdominal compressions cardiopulmonary resuscitation. Am J Emerg Med 2014;32:931–4. [26] Babbs CF, Meyer A, Nadkarni V. Neonatal CPR: room at the top—a mathematical study of optimal chest compression frequency versus body size. Resuscitation 2009;80:1280–4. [27] Park CH, Jeung KW, Min YI, Heo T. Sustained manual abdominal compression during cardiopulmonary resuscitation in a pig model: a preliminary investigation. Emerg Med J 2010;27:8–12. [28] Park YS, Park I, Kim YJ, Chung TN, Kim SW, Kim MJ, et al. Estimation of anatomical structures underneath the chest compression landmarks in children by using computed tomography. Resuscitation 2011;82:1030–5. [29] Na JU, Choi PC, Lee HJ, Shin DH, Han SK, Cho JH. A vertical two-thumb technique is superior to the two-thumb encircling technique for infant cardiopulmonary resuscitation. Acta Paediatr 2015;104:e70–5.
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Original Contribution
Does accelerometer feedback on high-quality chest compression improve survival rate? An in-hospital cardiac arrest simulation☆,☆☆,★ Min Hee Jung, MD a, Je Hyeok Oh, MD b,⁎, Chan Woong Kim, MD, PhD b, Sung Eun Kim, MD, PhD b, Dong Hoon Lee, MD, PhD b, Wen Joen Chang, MD, PhD a a b
Department of Emergency Medicine, Sungae Hospital, Seoul, Republic of Korea Department of Emergency Medicine, College of Medicine, Chung-Ang University, Seoul, Republic of Korea
a r t i c l e
i n f o
Article history: Received 3 April 2015 Received in revised form 8 April 2015 Accepted 9 April 2015
a b s t r a c t Objective: We investigated whether visual feedback from an accelerometer device facilitated high-quality chest compressions during an in-hospital cardiac arrest simulation using a manikin. Methods: Thirty health care providers participated in an in-hospital cardiac arrest simulation with 1 minute of continuous chest compressions. Chest compressions were performed on a manikin lying on a bed according to visual feedback from an accelerometer feedback device. The manikin and accelerometer recorded chest compression data simultaneously. The simulated patient was deemed to have survived when the chest compression data satisfied all of the preset high-quality chest compression criteria (depth ≥51 mm, rate N100 per minute, and ≥95% full recoil). Survival rates were calculated from the feedback device and manikin data. Results: The survival rate according to the feedback device data was 80%; however, the manikin data indicated a significantly lower survival rate (46.7%; P = .015). The difference between the accelerometer and manikin survival rates was not significant for participants with a body mass index greater than or equal to 20 kg/m 2 (93.3 vs 73.3%, respectively; P = .330); however, the difference in survival rate was significant in participants with body mass index less than 20 kg/m2 (66.7 vs 20.0%, respectively; P = .025). Conclusions: The use of accelerometer feedback devices to facilitate high-quality chest compression may not be appropriate for lightweight rescuers because of the potential for compression depth overestimation. Trial registration: Clinical Research Information Service (KCT0001449). © 2015 Elsevier Inc. All rights reserved.
1. Introduction The 2010 International Consensus on Cardiopulmonary Resuscitation (CPR) and Emergency Cardiovascular Care Science With Treatment Recommendations outlined the conditions necessary for high-quality chest compression as a compression depth of at least 2 inches (5 cm) at a rate of at least 100 compressions per minute, full chest recoil, and minimal interruptions [1]. In particular, maintaining a compression depth of more than 5 cm during CPR is associated with higher survival rates in adult and pediatric cardiac arrest patients [2,3]. Therefore, a device that provides feedback on chest compression depth, chest wall recoil, and compression rates has been developed and shown to facilitate the performance of high-quality chest compression under simulated cardiac arrest conditions [4,5]. However, in 2009, Perkins et al [6] reported that accelerometer feedback devices could overestimate compression depth when compressions ☆ Financial support: none. ☆☆ Conflict of interest statement: On behalf of all authors, the corresponding author states that there is no conflict of interest. ★ Funding: none. ⁎ Corresponding author at: Department of Emergency Medicine, Chung-Ang University Hospital, 102 Heukseok-ro, Dongjak-gu, Seoul, Republic of Korea 156-755. Tel.: +82 2 6299 1820; fax: +82 2 6299 2558. E-mail address: [email protected] (J.H. Oh). http://dx.doi.org/10.1016/j.ajem.2015.04.016 0735-6757/© 2015 Elsevier Inc. All rights reserved.
were performed on a soft surface. Such overestimation is likely to occur during in-hospital CPR because most cardiac arrest patients are lying on a bed. Although Oh et al [7] reported in 2012 that a dual accelerometer could prevent overestimation of compression depth, the dual accelerometer technique is not currently used in clinical settings. We used a manikin in an in-hospital cardiac arrest simulation to investigate whether visual feedback from an accelerometer device facilitated the performance of high-quality chest compressions.
2. Material and methods 2.1. Study design A prospective, nonrandomized single trial was carried out, with continuous chest compressions performed during 1 minute after regular CPR education at our hospital (Fig. 1). The study was approved by the institutional review board of our hospital (approval number: SA2015-07).
2.2. Study setting and study population A total of 30 health care providers who worked in the emergency department of a community hospital participated in the study.
994
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
PC SkillReporting System (Laerdal Medical) was used to collect chest compression data from the manikin. A CPRmeter (Laerdal Medical) was used to provide chest compression feedback. The CPRmeter provided visual feedback on compression depth, rate, and full recoil; and Q-CPR Review software version 3.1 (Laerdal Medical) was used to collect the data. 2.4. Study protocol
Fig. 1. Study flow diagram.
All participants were recruited voluntarily during regular CPR education and gave their verbal informed consent before the simulations. In-hospital cardiac arrest simulations using a manikin were conducted in the simulation center of the hospital after CPR training. Chest compression data were collected during the simulation. The sample size was calculated based on the mean compression depth (MCD) determined in a previous study (52.6 ± 6.7 mm) [8]. Given the SD of 6.7 mm, the expected MCD difference between the accelerometer device and manikin was postulated to be 10%, and the allowable difference was set at 5.26 mm. The 2-sided significance level was set at 0.05 with statistical power of 80%. The minimum number of participants was determined to be 13 using a Web-based program (sample size calculator: 1 sample mean) [9].
2.3. Study materials The following devices were used during the simulation. The Stretcher Trolley–Paramount Model: KK-728E (Paramount Bed, Tokyo, Japan) was used for the in-hospital cardiac arrest simulation. We used a step stool to control for differences between bed height and rescuer knee height [10]. We measured the knee heights of 26 adults who did not participate in the study. The mean height of the rescuers' knees was 45.7 ± 3.0 cm (range, 39-52 cm); therefore, we set the difference between the bed and step stool height to 45 cm. The difference between bed and knee height could be adjusted within 10 cm using this method. A Resusci Anne SkillReporter manikin (Laerdal Medical, Stavanger, Norway) was used as the simulated cardiac arrest patient. The Laerdal
Cardiopulmonary resuscitation training, which did not include practice using a manikin, was conducted for 1 hour with a focus on the requirements for high-quality chest compression. Previous investigations showed that chest compression quality was significantly lower when CPR was performed by lightweight rescuers than when performed by heavier rescuers; therefore, we recorded the participants' height and body weight before the simulation [11,12]. In addition, we recorded the participants' sex, age, and knee height (distance from the floor to the tibial tuberosity in the erect position) [8]. The in-hospital cardiac arrest simulation protocol was as follows. The manikin was placed on an emergency department bed in the supine position. Continuous chest compressions were performed according to the visual feedback provided by the accelerometer for 1 minute without ventilation assuming that an advanced airway was in place. The participants practiced CPR for 1 minute before the simulation to familiarize themselves with the visual feedback provided by the accelerometer during chest compressions. The investigator supervising data collection used a stopwatch to accurately instruct the participants when to start and stop chest compressions. The practice was limited to 1 minute because rescuer fatigue and muscle strength affect the quality of chest compression [11-13]. The bed-to-step stool height was adjusted by measuring the height of the step stool (x cm), and the height of the bed (distance from the floor to the upper surface of the mattress) was then adjusted to (x + 45) cm. The same step stool and bed were used for all simulations and were fixed at the same height. The mattress used was provided by the bed manufacturer. According to the 2010 International Consensus on CPR, there is insufficient evidence to argue for or against the use of backboards during CPR [1]. Therefore, we did not use a backboard in our simulations. The chest compression data were collected simultaneously by the accelerometer feedback device and manikin. 2.5. Outcome variables The simulated patient was deemed to have survived when the chest compression data collected during the simulation satisfied all of the preset high-quality chest compression criteria. Conversely, the patient was considered to have died when not all of the criteria were met. Our criteria were based on the recommendations of the 2010 International Consensus on CPR for high-quality chest compressions and included MCD greater than or equal to 51 mm, mean compression rate (MCR) greater than or equal to 100 compressions per minute, and greater than or equal to 95% full chest recoil [1]. The “minimizing interruptions” recommendation was excluded because chest compression was continuous in our simulation. The acceptable percentage of compressions with full chest recoil was at the investigator's discretion because the recommended proportion was not stated in the 2010 International Consensus. Survival rates were calculated from both the feedback device and manikin chest compression data. Survival rate was the primary outcome variable. The secondary outcome variables were the 5 chest compression indices common to the feedback device and the manikin: MCD (millimeters), MCR (counts per minute), adequate rate (percent), adequate depth (percent), and complete release (percent). As chest compressions performed by lightweight rescuers are shallower than those performed by heavier rescuers, we assessed the outcome variables according to the participants' body mass index
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
995
(BMI) by comparing the normal (≥20 kg/m2) and low (b20 kg/m2) BMI groups [11,12]. Furthermore, we collected data regarding corrected hand position on the manikin (percent), good compressions (percent), and perfusion time (percent) from the feedback device. 2.6. Statistical analyses PASW Statistics 18.0 (SPSS Inc, Chicago, IL) was used to conduct all statistical tests. Continuous variables are expressed as means ± SD. The categorical data are expressed as percentages. The Fisher exact test was used for statistical comparisons between the survival rates calculated from the feedback device and manikin. The Wilcoxon signed rank test was used to compare the compression data from the feedback device and the manikin. The Mann-Whitney U test was used to compare the compression data of the normal and lightweight groups. In all analyses, P b .05 was taken to indicate statistical significance. 3. Results 3.1. Participant characteristics The study group included 30 health care providers (2 male, 28 female) with a mean age of 28.4 ± 5.2 years, mean height of 162.3 ± 6.7 cm, mean body weight of 56.4 ± 11.0 kg, and a mean BMI of 21.3 ± 3.0 kg/m2. The mean participant knee height was 42.2 ± 2.2 cm, and the mean difference between bed and participant knee height was –2.8 ± 2.2 cm. 3.2. Comparison of feedback device and manikin survival rates The chest compression data from the feedback device and manikin are shown in Table 1. Survival rate calculated from the accelerometer data was significantly higher than that from the manikin (80% [24/30] vs 46.7% [14/30], respectively; P = .015) (Fig. 2). 3.3. Comparison of survival rates according to BMI The chest compression data according to BMI are shown in Tables 2 and 3. The calculated survival rates of the feedback device and manikin did not differ significantly among participants with BMI greater than or equal to 20 kg/m 2 (93.3% [14/15] vs 73.3% [11/15], respectively; P = .330). However, in participants with BMI less than 20 kg/m2, the survival rate was significantly higher for the accelerometer than for the manikin (66.7% [10/15] vs 20.0% [3/15], respectively; P = .025) (Fig. 3). Moreover, although the survival rates calculated from the feedback device did not differ significantly between the normal and low BMI groups (93.3 [14/15] vs 66.7% [10/15], respectively; P = .169), the manikin survival rates were significantly different between the normal (73.3% [11/15]) and low (20.0% [3/15]) BMI groups (P = .009) (Fig. 3).
Table 1 Comparisons of the chest compression data between the feedback device and the manikin Feedback device (n = 30) Manikin (n = 30) P MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Corrected hand position (%) Good compression (%) Flow time (%)
61.1 ± 7.6 114.0 ± 7.5 76.8 ± 31.2 91.0 ± 15.5 95.0 ± 12.1 N/A 67.1 ± 33.2 98.8 ± 0.8
Abbreviation: N/A, not applicable. A P value b .05 is presented in bold.
50.7 ± 5.4 114.7 ± 7.4 76.3 ± 30.9 55.9 ± 40.2 100.0 ± 0.0 100.0 ± 0.0 N/A N/A
.000 .109 .049 .000 .001 N/A N/A N/A
Fig. 2. Comparison of feedback device and manikin survival rates.
4. Discussion The 2010 International Consensus on CPR and Emergency Cardiovascular Care Science With Treatment Recommendations noted the potential for overestimation of compression depth when using feedback devices on a soft surface [1]. Despite various attempts, no devices that address this shortcoming are commercially available [7,14-16]. Therefore, standard feedback devices continue to be used. We used in-hospital cardiac arrest simulation to determine whether an accelerometer feedback device facilitated the performance of highquality chest compression. We chose the CPRmeter feedback device because it collects chest compression data and provides visual feedback for compression depth, rate, and full recoil. Although the CPRmeter has been assessed in previous simulation studies, these studies were performed with the surrogate victim on the floor [4,5]. Therefore, we reassessed the feedback device in the in-hospital environment where compressions are most likely to be performed with the patient on a bed. We found that although the accelerometer indicated high-quality compressions, the device overestimated the chest compression depth (Table 1). Thus, whereas the proportion of adequate depth compressions recorded by the feedback device was high, those recorded by the manikin were significantly lower (91.0% ± 15.5% vs 55.9% ± 40.2%, respectively; P = .000). Moreover, the calculated survival rate was overestimated by the feedback device compared with the manikin (80% and 46.7%, respectively; P = .015) (Fig. 2). A total of 43.3% (13/30) of participants met the high-quality chest compression criteria according to both the feedback device and manikin data. This finding indicates that although some participants were able to perform high-quality chest compression despite overestimation of compression depth, more than half were unable to do so. We hypothesized that the difference was attributed to differences in body weight. A previous study showed that the force needed to compress the victim's chest was affected by the rescuer's body weight, such that heavier rescuers used less muscle strength to perform chest compressions than did lightweight rescuers. Thus, the MCD of the lightweight group decreased significantly over the CPR period [11].
Table 2 Comparisons of the chest compression data collected by the feedback device between the normal BMI group (≥20 kg/m2) and the light BMI group (b20 kg/m2)
MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Good compression (%) Flow time (%)
BMI ≥20 kg/m2 (n = 15)
BMI b20 kg/m2 (n = 15)
P
64.8 ± 6.8 112.2 ± 8.2 77.9 ± 34.8 98.1 ± 1.6 96.3 ± 11.3 70.5 ± 35.3 98.8 ± 1.1
57.5 ± 6.7 115.8 ± 6.5 75.8 ± 28.3 83.9 ± 19.7 93.6 ± 13.1 63.6 ± 31.7 98.7 ± 0.5
.007 .116 .389 .004 .217 .412 .285
A P value b .05 is presented in bold.
996
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
Table 3 Comparisons of the chest compression data collected by the manikin between the normal BMI group (≥20 kg/m2) and the light BMI group (b20 kg/m2)
MCD (mm) MCR (/min) Adequate rate (%) Adequate depth (%) Complete release (%) Corrected hand-position (%)
BMI ≥20 kg/m2 (n = 15)
BMI b20 kg/m2 (n = 15)
P
53.3 ± 4.8 113.0 ± 8.3 76.5 ± 34.6 75.3 ± 36.0 100.0 ± 0.0 100.0 ± 0.0
48.1 ± 4.9 116.4 ± 6.4 76.0 ± 28.0 36.5 ± 35.2 100.0 ± 0.0 100.0 ± 0.0
.008 .217 .595 .006 1.000 1.000
A P value b .05 is presented in bold.
We found no significant difference between survival rates calculated from the feedback device and manikin in the normal BMI group (Fig. 3). However, the accelerometer and manikin survival rates were significantly different in the low BMI group (P = .025). Moreover, the survival rate calculated from the accelerometer data did not differ significantly according to BMI (Fig. 3), whereas the survival rate calculated from the manikin data was significantly lower in the low BMI group (P = .009). Our results have several implications. First, despite overestimation of compression depth, 43.3% of the participants were able to perform high-quality chest compressions using the accelerometer feedback, particularly those with a normal or high BMI. Second, the rescuers with a low BMI could not perform high-quality chest compressions because the feedback device overestimated the compression depth. Thus, the development of a feedback device that can measure compression depth accurately is essential. In lieu of such a device, altering the recommended compression depth (ie, 50 mm) to account for expected overestimation may provide a solution. For example, if an overestimation of 10 mm is expected, the recommended compression depth for a feedback device should be increased to 60 mm. However, this method may not be practical, as the degree of overestimation is likely to differ according to mattress and bed types as well as rescuer height and weight. Our findings are significant in that we confirmed that the overestimation of compression depth by feedback devices is associated with the rescuer's body weight. In particular, we established that feedback indices may be inappropriate for lightweight rescuers. The potential solution to developing an effective feedback device for use by any rescuer, regardless of their height, weight, and sex, is addressed by Maier et al [17]. If the upper body weight of the rescuer can be accurately enough estimated, the feedback device can be programmed to prompt the proper compression rate for each rescuer.
Men have greater upper body weight than women because they have more muscle mass; therefore, separate estimates must be made for men vs women. So that the feedback device can be used from the kneeling position on the floor or ground, a stool that was originally designed for kneeling during CPR provided on an emergency department bed, or for that matter, any elevated surface [18], can be redesigned for use on the floor or ground and be made adjustable in height, so that each rescuer can kneel at the optimal height, so that they can exert the maximal percentage of their upper body weight. Furthermore, to maximize force that is applied directly over the compression point of the lower half of the sternum, the right hand of the rescuer should be in contact with the sternum when standing or kneeling at the right side of the patient and vice versa [19], and an inclined step stool should be used [20]. When possible, the rescuer can choose the hand/body position they are most comfortable with. In addition, rescuers need to prompt proper rate and effective depth high-impulse CPR using 2 metronomes, 1 for compression and 1 for decompression [21]. Chest compressions with a hand position at the sternoxiphoid junction minimize the force requirements for effective compression and maximize compression over the left ventricular inflow tract [19,22-24]. However, whether it is best to perform high-impulse CPR at a slower rate according to the weight of the patient [25,26], or highimpulse CPR at a fast rate [17] requires further study in man. In addition, compression of the abdomen increases the risk of internal organ injury such as liver laceration [27]. Therefore, safety issue should be solved if we could use the sternoxiphoid junction as an anatomic landmark for chest compressions. In children, to avoid compression over the liver, compression should be performed with a hand position at the internipple line, which requires much more force [19,23,28]. Therefore, compressions should be performed with a 2-hand technique by someone with small hands, with the right hand in contact with the sternum. The rescuer is at the left side of the patient and vice versa, instead of right hand in contact with the sternum with rescuer at right side and vice versa, as it is recommended for adults [19,23]. The vertical 2-thumb technique [29] with more rapid and shallower compressions [25,26] might be helpful for infants (and some small children). Our study had several limitations. First, we used mechanical models, and therefore, the level of evidence was low. Second, most participants were female. Moreover, the mean BMI was low (21.3 ± 3.0 kg/m 2) because the BMI was less than 20 kg/m 2 in half of the participants. Third, we defined survival rate according to the recommended criteria for high-quality chest compression; however, the survival rate of cardiac arrest patients may not rely on these criteria. Fourth, our study did not include a nonfeedback group, so we were unable to evaluate
Fig. 3. Comparisons of survival rates according to BMI.
M.H. Jung et al. / American Journal of Emergency Medicine 33 (2015) 993–997
the positive effects of using a feedback device despite the potential overestimation of compression depth. Further studies to address this issue are warranted. 5. Conclusions Accelerometer feedback devices do not provide accurate information when compressions are performed on a bed. The proportion of actual high-quality chest compressions was significantly lower than that indicated by the device, particularly for rescuers with BMI less than 20 kg/m 2. Our findings indicate that the use of visual feedback from accelerometer devices to perform high-quality compressions is not appropriate for lightweight rescuers performing in-hospital CPR. Acknowledgments The authors thank all health care providers for their contribution in this study. References [1] Koster RW, Sayre MR, Botha M, Cave DM, Cudnik MT, Handley AJ, et al. Part 5: adult basic life support: 2010 International consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Resuscitation 2010;81(Suppl. 1):e48–70. [2] Vadeboncoeur T, Stolz U, Panchal A, Silver A, Venuti M, Tobin J, et al. Chest compression depth and survival in out-of-hospital cardiac arrest. Resuscitation 2014;85:182–8. [3] Sutton RM, French B, Niles DE, Donoghue A, Topjian AA, Nishisaki A, et al. 2010 American Heart Association recommended compression depths during pediatric in-hospital resuscitations are associated with survival. Resuscitation 2014;85: 1179–84. [4] Skorning M, Beckers SK, Brokmann JCh, Rortgen D, Bergrath S, Veiser T, et al. New visual feedback device improves performance of chest compressions by professionals in simulated cardiac arrest. Resuscitation 2010;81:53–8. [5] Bulleon C, Parienti JJ, Halbout L, Arrot X, De Facq Regent H, Chelarescu D, et al. Improvement in chest compression quality using a feedback device (CPRmeter): a simulation randomized crossover study. Am J Emerg Med 2013;31:1457–61. [6] Perkins GD, Kocierz L, Smith SC, McCulloch RA, Davies RP. Compression feedback devices over estimate chest compression depth when performed on a bed. Resuscitation 2009;80:79–82. [7] Oh J, Song Y, Kang B, Kang H, Lim T, Suh Y, et al. The use of dual accelerometers improves measurement of chest compression depth. Resuscitation 2012;83:500–4. [8] Oh JH, Kim CW, Kim SE, Lee SJ, Lee DH. Comparison of chest compressions in the standing position beside a bed at knee level and the kneeling position: a nonrandomised, single-blind, cross-over trial. Emerg Med J 2014;31:533–5. [9] Centre for Clinical Research and Biostatistics. The Chinese University of Hong Kong. Sample size estimation. [Cited 11 February 2015.] Available from URL http://www. cct.cuhk.edu.hk/stat/index.htm. [10] Cho J, Oh JH, Park YS, Park IC, Chung SP. Effects of bed height on the performance of chest compressions. Emerg Med J 2009;26:807–10.
997
[11] Hasegawa T, Daikoku R, Saito S, Saito Y. Relationship between weight of rescuer and quality of chest compression during cardiopulmonary resuscitation. J Physiol Anthropol 2014;33:16. [12] Russo SG, Neumann P, Reinhardt S, Timmermann A, Niklas A, Quintel M, et al. Impact of physical fitness and biometric data on the quality of external chest compression: a randomized, crossover trial. BMC Emerg Med 2011;11:20. [13] Ock SM, Kim YM, Chung Jh, Kim SH. Influence of physical fitness on the performance of 5-minute continuous chest compression. Eur J Emerg Med 2011;18:251–6. [14] Kang MK, Oh JH, Kim TW. Chest compression depth measurement using IRUWB for CPR (cardiopulmonary resuscitation). Proc IEEE APCCAS 2010. [Cited 15 February 2015.] Available from URL http://ieeexplore.ieee.org/xpl/articleDetails. jsp?arnumber=5774952. [15] Beesems SG, Koster RW. Accurate feedback of chest compression depth on a manikin on a soft surface with correction for total body displacement. Resuscitation 2014;85: 1439–43. [16] Wutzler A, Bannehr M, von Ulmenstein S, Loehr L, Forster J, Kuhnle Y, et al. Performance of chest compressions with the use of a new audio-visual feedback device: a randomized manikin study in health care professionals. Resuscitation 2015;87: 81–5. [17] Maier GW, Tyson Jr GS, Olsen CO, Kernstein KH, Davis JW, Conn EH, et al. The physiology of external cardiac massage: high-impulse cardiopulmonary resuscitation. Circulation 1984;70:86–101. [18] Oh J, Chee Y, Lim T, Cho Y, Kim IY. Chest compression with kneeling posture in hospital cardiopulmonary resuscitation: a randomised crossover simulation study. Emerg Med Australas 2014;26:585–90. [19] Baubin M, Kollmitzer J, Pomaroli A, Kraincuk P, Kranzl A, Sumann G, et al. Force distribution across the heel of the hand during simulated manual chest compression. Resuscitation 1997;35:259–63. [20] Yun SW, Lee BK, Jeung KW, Park SW, Choi SS, Lee CH, et al. The effect of inclined step stool on the quality of chest compression during in-hospital cardiopulmonary resuscitation. Am J Emerg Med 2014;32:851–5. [21] Chung TN, Bae J, Kim EC, Cho YK, You JS, Choi SW, et al. Induction of a shorter compression phase is correlated with a deeper chest compression during metronomeguided cardiopulmonary resuscitation: a manikin study. Emerg Med J 2013;30: 551–4. [22] Cha KC, Kim YJ, Shin HJ, Cha YS, Kim H, Lee KH, et al. Optimal position for external chest compression during cardiopulmonary resuscitation: an analysis based on chest CT in patients resuscitated from cardiac arrest. Emerg Med J 2013;30:615–9. [23] Shin J, Rhee JE, Kim K. Is the inter-nipple line the correct hand position for effective chest compression in adult cardiopulmonary resuscitation? Resuscitation 2007;75: 305–10. [24] Pickard A, Darby M, Soar J. Radiological assessment of the adult chest: implications for chest compressions. Resuscitation 2006;71:387–90. [25] Rottenberg EM. The critical need for further research and development of abdominal compressions cardiopulmonary resuscitation. Am J Emerg Med 2014;32:931–4. [26] Babbs CF, Meyer A, Nadkarni V. Neonatal CPR: room at the top—a mathematical study of optimal chest compression frequency versus body size. Resuscitation 2009;80:1280–4. [27] Park CH, Jeung KW, Min YI, Heo T. Sustained manual abdominal compression during cardiopulmonary resuscitation in a pig model: a preliminary investigation. Emerg Med J 2010;27:8–12. [28] Park YS, Park I, Kim YJ, Chung TN, Kim SW, Kim MJ, et al. Estimation of anatomical structures underneath the chest compression landmarks in children by using computed tomography. Resuscitation 2011;82:1030–5. [29] Na JU, Choi PC, Lee HJ, Shin DH, Han SK, Cho JH. A vertical two-thumb technique is superior to the two-thumb encircling technique for infant cardiopulmonary resuscitation. Acta Paediatr 2015;104:e70–5.
