A Smartphone application to provide real-time cardiopulmonary resuscitation quality feedback

a b s t r a c t

Background: Quality of cardiopulmonary resuscitation (CPR) contributes significantly to outcomes. Key determi- nants of CPR quality pertaining to chest compressions are compression rate, compression depth, duration of in- terruptions, and chest recoil. Several studies have demonstrated that Real-time audiovisual feedback improves CPR quality. We hypothesize that a mobile application using sensor data from built-in accelerometers in smartphones can provide accurate chest Compression quality feedback in real time. This study aims to develop and validate an application for smartphone which can provide real-time audiovisual and haptic feedback on determinants of CPR quality.

Methods: A mobile application was developed to detect the compression depth and compression rate in real time using data captured from a smartphone’s intrinsic accelerometer. The mobile device was placed on an adult manikin‘s chest at the point of compressions. In a simulated environment, data obtained using the application was compared directly to data obtained from a validated standard CPR quality tool.

Results: CPR quality parameters were obtained from the application and industry standard for 60, 30s-long sessions. Bland-Altman plot analysis for compression depth showed agreement between the app measurements and standard within +-4 mm (<10% error). The interclass correlation for agreement in the measurement of compression count was 0.92 (95% CI: 0.88-0.95), indicative of very strong agreement.

Conclusions: Smart device applications using acceleration sensor data derived from smart phones can accurately provide real-time CPR quality feedback. With further development and validation, they can provide a ubiqui- tously available CPR feedback tool valuable for out-of-hospital arrests and in-hospital arrests in under- privileged areas.

(C) 2022

  1. Introduction

chest compression rates and depth during cardiopulmonary resuscita- tion (CPR) affect survival after cardiac arrest [1,2]. Real-time audiovisual feedback has been associated with improved CPR technique and improved outcomes from cardiac arrest [3]. Audiovisual Feedback devices designed for the express use of improving CPR quality during resuscitation are

* Corresponding author at: M Health Fairview University of Minnesota Masonic Children’s Hospital, MN, United States of America

E-mail address: [email protected] (P. Sinha).

1 Both authors worked together and contributed equally and significantly for shared

first authorship.

readily available in emergency departments and intensive care units at resource-rich hospitals and have demonstrated reproducible favorable re- sults [3,4]. In fact, this technology is often built into manual defibrillators and can be utilized by either the person providing compressions or a ded- icated “CPR Coach” on the resuscitation team. Notably, the specific addi- tion of a CPR Coach to the resuscitation team has now been recognized as a CPR quality improvement measure, further highlighting the impor- tance of Real-time feedback of chest compression quality [5], both during real scenarios and for educational and training simulations.

When cardiac arrest occurs outside the hospital or in resource- limited settings, real-time audiovisual feedback is not routinely avail- able to the provider. This deficit probably has a disproportionate effect in the out-of-hospital cardiac arrest (OHCA) setting, where the CPR

0735-6757/(C) 2022

provider is more likely to be a lay person (as in “bystander CPR”) and an additional provider acting as a CPR coach is unlikely to be present [6,7]. OHCA is a leading cause of mortality in the United States with an inci- dence approaching 360,000 annually, and is witnessed by a lay by- stander 37% of the time [7]. Quality of CPR is a frequently studied modifiable contributing factor to OHCA morbidity and mortality. Feed- back devices designed for bystander CPR training have been shown to improve Quality of chest compressions by lay people, and other devices designed specifically for use by out-of-hospital medical personnel have been successful in improving CPR provided in the field [3]. Resource- limited areas of the world, where more CPR is administered manually, also represent an opportunity to improve resuscitation outcomes with more widely available compression feedback.

As Wearable technology has now become nearly ubiquitous, a pro- gram capable of generating real-time audiovisual feedback delineating quality of chest compressions would be a life-saving addition to the common smartphone or other wearable technology platforms. Mobile device use is rapidly growing globally and more than half of these de- vices are smartphones [8]. Though a wide range of smartphones exist on the market, most include an accelerometer to measure linear accel- eration and a gyroscope to measure angular velocity. These standard components support common smartphone features including mapping services, pedometers, and automatic screen rotation. The same technol- ogy can be harnessed to record and analyze the quality of chest com- pressions in real-time.

Access to a simple, user-friendly application on a smartphone or other wearable technology platforms has the potential to fill an impor- tant gap in both the pre-hospital treatment of OHCA, and in resuscita- tion in hospitals where validated feedback systems are not integrated into routinely available equipment.

With modifications a similar application could be expanded to in- clude other wearable technology platforms, expanding availability and ease of use.

    1. Objective

This study aims to develop and validate a mobile smartphone appli- cation able to provide real-time audiovisual feedback to CPR providers. The global goal is to optimize CPR quality in out-of-hospital and low re- source settings. The ideal application would measure all mechanical as- pects of chest compression, i.e., compression rate, compression depth,

compression fraction, and pauses. It would also provide visual, audio and haptic feedback for real time assessment of CPR quality, and thus provide opportunities for improvement.

The primary endpoints for the study were compression depth and compression count per session. The primary hypothesis for the study is that the designed application can measure compression depth and count of compressions per session as accurately as the validated standard.

  1. Methods

As this work does not involve any human subjects Institutional Review Board approval was not needed.

    1. Application development

An application was developed on a smart phone to record raw data describing compression depth, compression rate and pause duration using the built-in accelerometer. An algorithm was designed to trans- late the data into visual representation of compression depth and com- pression rate. The phone’s built-in accelerometer captures displacement in three dimensions (X, Y and Z axis). The application samples the dis- placement of the device at 50 Hz, or every 20 ms. To calculate the nec- essary indices, an algorithm was designed to use second order numerical integration to convert accelerometer displacement values to measures of compression depth. A moving average filter was then ap- plied to eliminate signal noise. Calculated values were adjusted using linear regression to more accurately represent compression depth. Inci- dence of compressions per second was reported as the compression rate, and the duration of uninterrupted compressions in each 30 s session was reported as the chest compression fraction (Fig. 1).

The application’s user interface (UI) was designed to provide real- time visual feedback to the CPR provider(s) incorporating the three parameters described above (Fig. 2). In addition, it supplies audio and haptic feedback which alarms for any pauses in compressions lasting longer than one second. Together, the visual and audio cues alert the provider of deficiencies in compression rate, compression depth, and compression fraction during CPR. The application was also designed to provide a comprehensive report to the user after each session for more detailed review.

Image of Fig. 1

Fig. 1. Instantaneous acceleration data and corresponding velocity and displacement values calculated by the application.

Image of Fig. 3

Image of Fig. 2Fig. 2. Application user interface displaying compression depth, compression rate, and chest compression fraction.

    1. Application validation
      1. Data collection

After development of the application, it was loaded onto a OnePlus One smart phone, which uses an Android operating system. Simulated CPR (compression only) using a validated CPR training manikin (Laerdal QCPR [9]), which is the commercially available standard was performed in 30-s-long sessions. 60 sessions were conducted in total. The smart de- vice was then placed on the anterior chest of the manikin at the site of compressions. Simulated (compression-only) CPR was again conducted while data was simultaneously obtained by both the smart phone appli- cation and the built in CPR quality assessment tool inside the manikin. (Fig. 3). CPR providers randomly provided chest compressions of opti- mal or suboptimal quality with respect to depth and rate for compres- sions through different sessions; however, the sessions lengths were kept the same (30 s long) to maintain consistency.

      1. Statistical analysis

Data from the application and the commercially available standard were analyzed using Pearson’s correlation and Bland-Altman analysis for agreement. Compression depth and number of compressions were analyzed. Mean difference (bias) between the application and commer- cially available standard was calculated with 95% confidence interval for limits of agreement. Relative bias for compression depth measurements was calculated as a percentage based on the absolute bias of the applica- tion divided by the average depth measurement. The intraclass correla- tion coefficient (ICC) was calculated with corresponding 95% confidence interval to assess agreement between the two devices regarding com- pression count. Statistical analyses were completed using Stata 15.0 (Stata Corp, College Station, TX).

Fig. 3. Data collection and validation system- Placement of smartphone on the anterior chest at the site of chest compressions allowed obtaining CPR quality data from the devel- oped application and validated standard.

  1. Results

CPR quality parameters were obtained from the application and the commercial standard for 60, 30-s-long sessions. Data obtained from both the commercial standard and the developed application was ana- lyzed for number of compressions and depth of compressions.

Pearson’s correlation revealed a strong correlation between the application and commercial standard for both depth of compressions (R2 = 0.95) (Fig. 4A) and number of compressions (R2 = 0.88) (Fig. 4B).

Bland-Altman plot analysis was performed to evaluate the accuracy of compression depth via comparison of the application data to the commercially available commercial standard (Laerdal QCPR). Measure- ments for compression depth showed agreement between application measurements and the standard within +-4 mm (<10% error; 95% CI:

-7.9% to 7%). Interclass correlation was performed for compression count by the application and the standard tool. (Fig. 5). Bland-Altman plot analysis to evaluate the number of compressions between the ap- plication and commercial standard also showed agreement between the two methods and was 0.92 (95% CI: 0.88-0.95). (Fig. 6).

  1. Discussion

The components of high-quality CPR as described by the American Heart Association are: minimal interruptions (<20% of CPR time), chest compression rate between 100 and 120 per minute, chest com- pression depth (>5 cm in adults or >1/3rd the anterior-posterior di- mension of the chest in infants and children), and full chest recoil [1]. There is also evidence to suggest that real-time feedback on CPR quality improves CPR delivery and outcomes after cardiac arrest [3]. However, accurate feedback tools are available only in specialized hospital or EMS settings. They are particularly lacking in instances of bystander CPR as well as in the majority of the world where low- and middle- income countries cannot afford the healthcare infrastructure to incorpo- rate expensive feedback systems.

Smart phone use is rising rapidly with over 6 billion users worldwide and rapidly growing [10]. Widespread use of these devices is having a positive impact on multiple public health issues [11-16] This is the first report of the use of Smartphone application to assess and impact CPR quality in settings that do not have sophisticated CPR quality mea- suring tools (bystander CPR and Low- and middle-income countries). These settings experience the majority of instances of CPR performed worldwide.

Our results indicate that built-in hardware in smartphones is sensi- tive and accurate, and in conjunction with our application can provide

Image of Fig. 4

Fig. 4. Graph showing correlation between data obtained from the application and commercial standard with regards to (A)depth of compressions and (B) number of compressions.

accurate real time feedback on chest compression quality. A strong cor- relation between our application and the commercial standard with regards to number and depth of compression as well as good agreement between measured compression depth suggest that the ubiquitous smartphone technology is readily adaptable to this indication. Commer- cially available standard feedback systems are routinely used for CPR training for both lay people and advanced practitioners, with the goal of improving compression rate, compression depth, and compression fraction during In-hospital CPR.

Our data strongly suggests that a smartphone-based application can afford the same advantage to the out-of-hospital setting and to low- resource settings. An ideal application designed for use during CPR should be easy to operate by users with a wide breadth of medical expe- rience, readily available in unpredictable emergency situations, not

limited by the availability of internet access, and validated using state- of-the-art standard technology. Currently, bystander CPR is initiated in under half of bystander witnessed OHCA. Lay persons, regardless of prior CPR training, may be more likely to initiate chest compressions if equipped with guidance in the form of feedback during these stressful and unpredictable events. The application described here, when pre- loaded onto a user’s smartphone, will provide life-saving real-time qual- ity data to the rescuer without the need for special training, additional equipment, or the presence of an additional provider. In addition, the ability of this application to provide data summaries after each session has the potential to incorporate real-time audio-visual feedback into in-hospital CPR instruction for medical professionals in resource- limited settings without ample access to standard devices such as the Laerdal QCPR system.

Image of Fig. 5

Fig. 5. Bland-Altman plot analysis showing agreement between the application and the available commercial standard with regards to compression depth.

Image of Fig. 6

Fig. 6. Bland-Altman plot analysis showing agreement between the application and the available commercial standard with regards to number of compressions.

As similar hardware is available in other wearable technology such as smart watches and fitness trackers in the future may further expand its application, with the advantage of the tracker on the CPR providers’ wrists rather than on the patient’s chest, minimizing the chance of dam- age to the equipment itself.

  1. Limitations

This is a pilot study conducted with the intent to demonstrate feasi- bility of a smart device-based CPR quality feedback system. This study used smartphone-based technology as a proof of concept to measure CPR quality. The location of the accelerometer should ideally be opti- mized in relation to the compression site, which may be best achieved by integrating the app into wearable technology on the provider’s wrist. This could ostensibly be done by adapting the technology to a smart watch platform. This adaptation would also prevent placing the device directly on the chest during compressions which would safe- guard against damage to the device or injury to the patient. In addition, the study was performed using an adult manikin only, as adult parame- ters of good CPR quality are standardized. Our developed application could be easily scaled down to allow validation using Pediatric manikins in the future. Further validation of this technology using wearable devices in clinical settings is required.

  1. Conclusions

Applications designed for smartphones which utilize pre-installed hardware to generate acceleration data can accurately provide CPR quality feedback in real-time. With further development and validation, as well as extrapolation to the smartwatch platform, they may function as a ubiquitously available CPR feedback tool with the potential to im- prove outcomes in out of hospital cardiac arrests and cardiac arrest in hospitals serving under-privileged areas worldwide for both clinical use and educational simulations.

Source of funding

This work was generously funded by Children’s National Hospital, Clinical and Translational Science Institute Winter Voucher Award Program awarded to PS.


Authors have no disclosures or conflicts of interest pertaining to this work.


This work was generously funded by Children’s National Hospital, Clinical and Translational Science Institute Winter Voucher Award Program awarded to Pranava Sinha, MD.

Credit authorship contribution statement

Emily Stumpf: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation. Ravi Tej Ambati: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation. Raj Shekhar: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Project administration, Methodol- ogy, Investigation, Funding acquisition, Conceptualization. Steven

J. Staffa: Formal analysis. David Zurakowski: Formal analysis. Pranava Sinha: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization.


  1. Vadeboncoeur T, Stolz U, Panchal A, Silver A, Venuti M, Tobin J, et al. Chest compres- sion depth and survival in out-of-hospital cardiac arrest. Resuscitation. 2014;85(2): 182-8.
  2. Idris AH, Guffey D, Pepe PE, Brown SP, Brooks SC, Callaway CW, et al. Chest compression rates and survival following out-of-hospital cardiac arrest. Critical Care Medicine. 2015;43(4):840-8. 0000000000000824.
  3. Lakomek F, Lukas RP, Brinkrolf P, Mennewisch A, Steinsiek N, Gutendorf P, et al. Real-time feedback improves chest compression quality in out-of-hospital cardiac arrest: a prospective cohort study. PLOS ONE. 2020;15(2):e0229431. https://doi. org/10.1371/JOURNAL.PONE.0229431.
  4. Cheng A, Overly F, Kessler D, Nadkarni VM, Lin Y, Doan Q, et al. Perception of CPR quality: influence of CPR feedback, just-in-time CPR training and provider role. Resuscitation. 2015;87:44-50.
  5. Nadkarni VM, Nolan JP, Billi JE, Bossaert L, Bottiger BW, Chamberlain D, et al. Part 2: international collaboration in resuscitation science. 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science With Treatment Recommendations, 122; 2010. p. S276-82.
  6. Wissenberg M, Lippert FK, Folke F, et al. Association of national initiatives to improve Cardiac arrest management with rates of bystander intervention and patient survival after out-of-hospital cardiac arrest. JAMA. 2013;310:1377-84.
  7. Berdowski J, Berg RA, Tijssen JG, Koster RW. Global incidences of out-of-hospital cardiac arrest and survival rates: systematic review of 67 prospective studies. Resus- citation. 2010;81:1479-87.
  8. Taylor K, Silver L. Smartphone ownership is growing rapidly around the world, but not always equally. Pew Res Cent. 2019.;47(February).
  9. Laerdal. CPR Meter. 2017; 2017..
  10. Number of smartphone subscriptions worldwide from 2016 to 2027. S. O’Dea; 2022. worldwide (Accessed on June 7, 2022).
  11. Cohen IG, Gostin LO, Weitzner DJ. Digital smartphone tracking for COVID-19: public health and civil liberties in tension. JAMA. 2020;323:2371-2.
  12. Plaza Roncero A, Marques G, Sainz-De-Abajo B, et al. Mobile health apps for Medical emergencies: systematic review. JMIR Mhealth Uhealth. 2020;8(12):e18513.
  13. Rathbone AL, Prescott J. The use of Mobile apps and SMS messaging as physical and mental health interventions: systematic review. J Med Internet Res. 2017;19(8): e295.
  14. Nicholas J, Larsen ME, Proudfoot J, Christensen H. Mobile apps for bipolar disorder: a systematic review of features and content quality. J Med Internet Res. 2015;17(8): e198.
  15. Tabi K, Randhawa AS, Choi F, et al. Mobile apps for medication management: review and analysis. JMIR Mhealth Uhealth. 2019;7(9):e13608. 13608.
  16. Collado-Borrell R, Escudero-Vilaplana V, Villanueva-Bueno C, Herranz-Alonso A, Sanjurjo-Saez M. Features and functionalities of smartphone apps related to COVID-19: systematic search in app stores and content analysis. J Med Internet Res. 2020;22(8):e20334.

Leave a Reply

Your email address will not be published. Required fields are marked *