Article Text
PDF
Abstract
Background Conflicting studies exist about the effectiveness of cardiopulmonary resuscitation (CPR) on a dental chair. In some situations, dental surgeons are obliged to perform CPR with the patient on the chair. Feedback devices are supposed to guide the compression depth in order to improve the quality of CPR, but some devices are based on an accelerometer that can theoretically report erroneous results because of the lack of rigidity of a dental chair.
Objective The aim of this study was to evaluate the accuracy of these devices to guide chest compressions on a dental chair.
Methods A prospective, randomised, crossover, equivalence/non-inferiority study was conducted to compare the values of compression depths provided by the feedback device (Real CPR Help®, delivered by Zoll© Medical Corporation, Chelmsford, MA, USA) with the real measurements provided by the manikin (Resusci Anne® Advanced SkillTrainer, Laerdal Medical AS©, Norway). Chest-compression-only CPR was performed by 15 Basic Life Support instructors who carried out two rounds of continuous CPR for 2 min each. Data were analysed with a correlation test, a Bland–Altman method and a Wilcoxon test. Statistical significance was defined as p<0.05.
Results A significant difference was found between the mean depths of compression measured by the feedback device and the manikin on a dental chair and on the floor (p<0.0001). The feedback device overestimated the depth of chest compressions, and Bland–Altman analysis demonstrated poor agreement.
Conclusion This study indicates that feedback devices with accelerometer technology are not sufficiently reliable to ensure adequate chest compression on dental chairs.
- Cardiac arrest
- resuscitation
Statistics from Altmetric.com
Introduction
Despite significant advances in resuscitation science over the last decade,1 ,2 the survival rate following a cardiac arrest remains low.3 ,4 One explanation is the suboptimal quality of cardiopulmonary resuscitation (CPR). To improve the quality of CPR, some manufacturers of CPR feedback devices have started monitoring the quality of chest compressions during training or CPR and have integrated visual and/or audio feedback during actual CPR performance, allowing real-time improvement in the quality of chest compressions. However, the reliability of CPR feedback devices on a soft surface is questionable.
Studies show that patients do undergo cardiac arrest while under the care of a dental surgeon, though with a very low incidence of 0.002 cardiac arrests per dentist per year.5–7 Although dental surgeons are medical professionals educated to perform CPR, the CPR technique is challenging even for health professionals undergoing regular training.8 Furthermore, dental surgeons who have not been able to maintain these skills at a sufficient level may experience more difficulties in performing CPR. A good CPR technique requires placing the patient on a hard surface to ensure the effectiveness of thoracic compressions; however, dental procedures are carried out on a padded dental chair, which may alter the effectiveness of CPR. Feedback devices may help dental surgeons to correctly perform CPR on dental chairs; however, to our knowledge, no study has evaluated the reliability of these feedback devices for CPR performed on a dental chair. Finally, we do not know whether CPR can be performed effectively on dental chairs or whether it must be performed only on the floor.
The primary objective of this study was to evaluate the use of a feedback device to guide chest compressions performed on a manikin on a dental chair by comparing the compression depth measured by the feedback device (Real CPR Help®, delivered by Zoll© Medical Corporation, USA) with those directly measured by the manikin on the floor or on a dental chair. The null hypothesis was that the feedback device provided the same values as the manikin, which allows the use of feedback devices on dental chairs. The secondary objective was a quality comparison of chest-compression-only CPR performed on the floor and that performed on a dental chair. The hypothesis was that median compression depths would not be different on the floor or on a dental chair.
Materials and methods
This is a prospective, randomised, crossover, equivalence/non-inferiority study. A Resusci Anne® Advanced SkillTrainer manikin (Laerdal Medical AS©, Norway), weighing 70 kg, was used during all evaluations. The data provided by the Resusci Anne® are based on a direct evaluation of chest compression depth measured by the distance between the sternum and the spine. Thus, this device provides real measurements of chest compression. The data were collected from the manikin and downloaded directly to a laptop using Laerdal© PC Skill reporter® 2.2.0.1 (Laerdal Medical AS©). The manikin was on a fully reclined chair. The height of the dental chair was adjusted for each Basic Life Support (BLS) provider.
Accelerometer data (called feedback data) were collected by placing a Zoll© CPR-D padz® (Zoll© Medical Corporation) on the centre of the manikin's chest beneath the palms of the BLS provider's hands during chest compressions. The accelerometer was linked to a Zoll© AED (Automated External Defibrillation) Plus® equipped with the Real CPR Help® technology. The data were analysed using Rescue Net Code Review 5.1.1.8® (Zoll© Medical Corporation). The chair was a planmeca Chair® (planmeca®, Finland) with memory foam, as used in the majority of dental chairs.
Subjects
BLS instructors of the French Red Cross were selected on a voluntary basis and contacted using the French Red Cross (Paris area) email system. As the objective of the study was to evaluate the feedback device, BLS instructors were used to ensure reproducible CPR.
A criterion for exclusion was known pregnancy.
Each subject received written information about the study and was required to complete a consent form.
Measures
For the purpose of comparison, simultaneous measurements of feedback compression depth (via the accelerometer placed on the sternum) and actual chest compression depth (via manikin sensors) were recorded in all experiments. The target of compression depth was between 50 and 60 mm.9
Protocol
Each BLS instructor performed two rounds of continuous chest-compression-only CPR for 2 min each. The sequence of participants' passages and the orders of the surface type for each participant were defined by a computer-generated randomised list (figure 1). Each series was separated by 20 min of rest, with each round alternating between the floor and the dental chair.10 Chest compressions were timed by a metronome set to 100 beats per minute. Rescuers used a CPR technique allowing complete chest recoil.11 Performers and the research team were blinded to the feedback device information during recording. Thus, for each participant, we aimed to record 200 chest compressions for each 2-min round. Given the type of study, no institutional agreement was necessary.
Flow diagram of the study.
Statistics
On the basis of former studies dealing with chest compression measurements, we included 15 rescuers per group.
Data were obtained directly by the Laerdal© PC Skill reporter® 2.2.0.1 and by the Rescue Net Code Review 5.1.1.8® (Zoll© Medical Corporation).
Given the crossover design, all subjects' interventions were paired. For the primary objective, we performed a three-step analysis. First, Pearson's test was carried out to verify the expected correlation of values obtained by both methods. Second, we compared the medians of these values with a Wilcoxon test. Finally, a Bland–Altman assessment for agreement was used to compare the two depth measurements. A chest compression depth bias of 5 mm was considered to be acceptable. The range of agreement was defined as a mean bias of ±2 SD. Results with a normal distribution were expressed as a mean±SD, and values with a non-normal distribution were expressed as a median, 25/75 percentile, and as percentages for categorical data. Normal distribution was verified with a Kolmogorov–Smirnov test. Statistical significance was defined as p<0.05. Statistical analyses were performed with SPSS® Statistics V.19.0.
Results
Subjects (n=15; 9 men) had an average age of 30±1 years (range, 25–42 years). The mean height was 171±2 cm and the weight was 68±3 kg. The 15 BLS instructors were certified in CPR and their last class was less than a year before the experiment. Thirty rounds of chest-compression-only CPR were available for analysis. A total of 6128 compressions were available for analysis. The mean CPR rate was 101 beats per minute.
As expected, the results given by the manikin and the feedback device were correlated (table 1 and figure 2). Nevertheless, we found a significant difference between the results given by the manikin and the feedback device on the chair (manikin median=44 mm (36–51), device median=54.8 mm (47.8–68.2); p<0.0001) and on the floor (manikin median=48 mm (44–51), device median=51.4 mm (45.9–59.4); p<0.0001), which indicates that the feedback device overestimates compression depth. Figure 2 shows that overestimation increases with compression depth. The Bland–Altman analysis indicates that the mean bias was 10.6 mm, with a 95% limit of agreement between the two methods ranging from −7.4 to 28.6 mm. Through the Bland–Altman method, we found poor agreement between the measures (figure 3), with a bias superior to the defined acceptable limit of 5 mm. The mean depth of 30% of the compression rounds was in an appropriate range (table 2).
Correlation of mean compression depth by the two devices according to surface
Percentages of rounds with mean compression depth in the target range (50–60 mm)
Point cloud of mean compression depth by the two devices for each surface (the lines at 50 and 60 mm represent the target range).
Bland–Altman analysis plotting the differences between simultaneous values of compression depth of manikin and feedback versus their average for each surface. Round forms: chair; square forms: floor; solid line: bias (mean difference between manikin and feedback); dotted lines: upper and lower limits of agreement.
For the secondary objective of the study, a Wilcoxon test found a difference of compression depth between the floor (median=48 mm (44–51)) and the dental chair (44 mm (36–51)) according to the manikin (p=0.025).
Discussion
In this study evaluating the use of a feedback device to guide chest compressions, we observed that the feedback device overestimates chest compression depth during CPR performed on a dental chair and on the floor. Most of the compressions measured in the correct range by the manikin were found to be too deep by the feedback device. The extent of this overestimation is greater on a dental chair (44 vs 54.8 mm) than on the floor (48 vs 51.4 mm).
Feedback devices have been shown to be an effective method for improving the efficiency of chest compressions performed on the floor.12 Our results warn of the potentially inappropriate use of feedback devices with accelerometers on other surfaces.
The observed overestimation seems to increase for the deepest compressions. Therefore, we hypothesise that deep compressions may push the patient into the foam of the chair or even move the whole seat, causing undetected loss of the thoracic-pump effect. This could explain how accelerometers overestimate compression depth more on dental chairs than on the floor. Accelerometers analyse the movement as a whole. They cannot distinguish between chest and chair movements. A previous study described this phenomenon on mattresses, which led to significant undercompression during CPR.13
The effectiveness of thoracic compressions carried out on dental chairs was the subject of three publications with conflicting results. In an original article from 2003, no significant difference was found between the ground and a dental chair.14 A first letter published in Resuscitation in 2008 reached the same conclusion.15 A second one, published in 2010, found conflicting results.16 However, our study includes more subjects and carries greater weight. It confirms the latest study by indicating that compression performed on dental chairs is not as effective as on the floor. Nevertheless, the clinical relevance of this difference of only 4 mm is questionable. Furthermore, this is the first study that compares accelerometer and direct measurement, provides interesting information explaining at least part of the observed differences, and allows us to advise against the use of feedback devices with accelerometers for the purpose of comparing chest compression depths performed on soft surfaces.
Studies also show that the performance of chest compressions was significantly poorer when performed on a bed than on the floor and that compression depth decreased when the bed height was 20 cm higher than the knee height of the rescuer—another factor that could alter the quality of CPR not performed on the floor.17 ,18
In our study, the quality of CPR performed by BLS instructors was not optimal. The more probable explanation is the difficulty for these BLS instructors to adapt their CPR techniques to the recent transition from the 2005 to the 2010 guidelines that modified the compression depth ranges from 40–50 mm to 50–60 mm.
The alternative to performing CPR on dental chairs is to move the patient to the floor. However, this creates a risk of fall and traumatic injury. In addition, moving the patient out of the chair could be impossible if the practitioner is alone or if the patient is obese. Furthermore, despite their training in CPR technique during the undergraduate dental curriculum, dental surgeons may have difficulties performing CPR in real life because of ineffective training and the scarcity of such events.6 ,19 ,20
Given these factors, we should further investigate tools such as feedback devices that allow an improvement in the quality of CPR performed by dental surgeons. Feedback devices using technologies other than accelerometers could be evaluated in this setting.
Previous studies have demonstrated that the use of feedback devices provides an effective means of monitoring and improving the quality of CPR performed on the floor.21–23 A recent review summarises articles demonstrating the efficacy of these devices.12 These elements led to the 2010 Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care underlining the fact that ‘effective’ chest compressions are essential for providing blood flow during CPR and that ‘rescuers can be assisted to achieve the recommended compression rate and depth by prompt/feedback devices.’9 However, as some of the devices using an accelerometer can theoretically report erroneous results because of displacement of the patient if s/he is lying on a soft surface, the 2010 Guidelines warn rescuers about the risk of overestimating compression depths.9
Some study limitations should be mentioned. First, we did not evaluate a dentist population. Second, our study was performed on only one chair model and all dental chairs have different padding. To our knowledge, there is no standard firmness for dental chairs. Finally, we conducted a manikin study, which cannot be extrapolated to a human population.
If the present study highlights the limitations of feedback devices using accelerometers to help practitioners ensure correct CPR on dental chairs, further studies should be conducted to confirm these results and to evaluate other technologies for monitoring chest compression depth. However, no studies to date have demonstrated a significant improvement in long-term survival related to the use of CPR feedback devices during actual cardiac arrest events. Those devices must enter a system-based approach.9 ,24
Conclusions
Accelerometer-equipped feedback devices are not as effective on a dental chair as they are on the floor for CPR. Owing to the lack of rigidity of the dental chair, the devices cannot accurately measure the depth of chest compressions and do not effectively guide the practitioner. As such, other feedback technologies should be evaluated in this setting.
Chest-compression-only CPR performed on dental chairs may not be as effective as that performed on the floor, and the clinical relevance of this significant but slight difference is uncertain.
References
Footnotes
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Primary survey