Validation and Reliability of PerformTek Earbud Heart Rate Sensor Utilizing 12 Lead ECG

Transcription

1 Sumner Blvd Raleigh, NC Validation and Reliability of PerformTek Earbud Heart Rate Sensor Utilizing 12 Lead ECG White paper prepared by: Dr. Chris Eschbach, CSCS, HFS Exercise Physiologist Valencell, Inc Independent data collection completed by: Dr. Meir Magal Fellow, American College of Sports Medicine

2 PURPOSE This report examines the accuracy and repeatability of Valencell, Inc. s PerformTek earbud sensor relative to a 12 lead electrocardiogram for measuring heart rate during exercise. INTRODUCTION Healthy living has seen a global resurgence as people place more value on staying active to live more fulfilling, healthier and longer lives. In this environment, fitness applications enjoy a meteoric rise across mobile phones, gaming and consumer fitness products. Workout efforts are optimized when receiving feedback on target heart rate zones and physical activity (Jeukendrup and Van Diemen, 1998). Additionally, heart rate monitoring of training intensity provides more accurate information than self-reports of training intensity (Gilman and Wells, 1993). The American College of Sports Medicine reports that heart rate monitoring has been shown to result in improvements in cardiorespiratory fitness when used for exercise prescription (ACSM, 2011). New technology has allowed the collection of metabolic data in a variety of ways including an earbud sensor module. The module obtains measures from the ear region using an approach akin to reflective-mode photoplethysmography (PPG). Highly miniaturized, the sensor module is designed to fit inside audio earbuds while still providing sufficient room for an audio cavity, speaker driver, and the other elements of a highperformance sports audio headset. Unlike pulse oximetry ear clips, this sensor module is completely embedded in the audio earbud, with no ear clips or ear supports required in order to function appropriately. The embedded sensors allow for real-time analysis of exercise data such as heart rate, distance traveled, calories burned, and VO2max (aka aerobic fitness). These sensors operate with light emitting diodes (LED) and are harmless to the users. The use of heart rate monitors is highly popular with individuals who exercise. Currently continuous heart rate monitors utilize a chest strap that transmits data to a specialized watch. These chest strap units are highly accurate as examined by Goodie et. al. (2000) but somewhat invasive and uncomfortable (Schonfelder et. al., 2011) in that the strap must be applied below any clothing at the level that the rib cage comes together. Earbud technology would eliminate the need for the chest strap. And because a growing number of individuals listen to music via headphones while exercising, earbud sensors are seamless and compatible with existing consumer behavior. A solution for mobile monitoring of heart rate during exercise must not only be seamless but also accurate and repeatable. However, PPG sensing of heart rate is notoriously limited by motion artifacts resulting in unsatisfactory performance (Burke and Whelan, 1987; Gehring et all, 2002; Yamaya et. al, 2002; Webb et. al., 2005; Comtois et. al, 2007). To the best of the author s knowledge there have been no reports with sufficient cohort data examining accuracy and reliability of heart rate sensors that are truly in the form-factor of audio earbuds that people are already used to wearing. METHODS Clinical trials examining validity and reliability of Valencell s PerformTek earbud sensing technology was completed using 41 healthy participants (22 males, 19 females, 29.7 ± 10.6 years, 76.5 ± 13.2 kg, ± 9.7 meters, 25.9 ± 3.7 BMI). Participants completed one of three 15.5 minute protocols involving sitting, standing, walking, and running while heart rate was monitored via a 12 lead electrocardiogram (ECG) and the Performtek sensor both devices recorded heart rate at 5-second intervals. Additionally, 36 participants returned to the laboratory and repeated the protocol within two weeks of the first trial to examine reliability.

3 Table 1. Protocols (all at 1% grade) Time (min) 0:00 0:30 0:30 1:00 1:00 2:00 2:00 4:00 4:00 6:00 6:00 7:30 7:30 9:30 9:30 11:30 11:30 13:30 13:30 14:30 14:30 15:00 15:00 15:30 Protocol #1 (mph, n=32) Sitting in chair Protocol #2 (mph, n=5) Sitting in chair Protocol #3 (mph, n=4) Sitting in chair Standing Standing Standing Sitting in chair Sitting in chair Sitting in chair Standing Standing Standing sensors provide an accurate and valid measure of heart rate. Additionally, test-retest reliability was high (r = 0.93) and similar to that of the ECG (r = 0.96). Figure 2. Correlation/regression analysis for trial 1 PerformTek vs. ECG (7626 data points). Figure 3. Frequency distribution for trial 1 PerformTek vs. ECG. (7626 data points). Figure 1. PerformTek placement in participant s ear. Figure 4. Bland-Altman plot for trial 1 PerformTek vs. ECG. (7626 data points). RESULTS Results indicated high correlation between the ECG and PerformTek (r = 0.99, r 2 = 0.98, SEE 4.43, 95% confidence intervals = Lower: Upper: 8.18, bias = -0.68). It was found that 87% of PerformTek HR measurements fell within ± 5 BPM of ECG HR measurements. This data indicates that earbud CONCLUSION High correlations in combination with low SEE indicate the PerformTek is a valid instrument for the measurements of HR during various activity levels. Additionally, test-retest reliability correlations similar to the benchmark measure

4 demonstrated accuracy with repeated use. Heart rate monitoring using the PerformTek system may be used effectively as an alternative to ECG or chest strap monitors. REFERENCES American College of sports Medicine, Quantity and Quality of Exercise for Developing and Maintaining Cardiorespiratory, Musculoskeletal, and Neuromotor Fitness in Apparently Healthy Adults: Guidance for Prescribing Exercise. Med. Sci. Sports Exerc 2011; 43(7): Burke MJ, Whelan MV. The Accuracy and Reliability of Commercial Heart Rate Monitors. Br I Sports Med 1987; 21: Jeukendrup A, Van Diemen AV. Heart rate monitoring during training and competition in cyclists. J. Sports Sci 1998; 16(S): Schonfelder M. Hinterseher G., Philipp P., and Spitzenpfeil P. Scientific Comparison of Different Online Heart Rate Monitoring Systems. J. Telemed and App 2011; 1-6 Webb KM, Hanson KJ, Sheely JM, Berry MJ. Validity of Heart Rate Obtained Via Pulse Oximetry Med. Sci. Sports Exerc 2005; 37(5): S232 S233 Yamaya, Y, Bogaard HJ., Wagner PD, Niizeki K, and Hopkins SR. Validity of Pulse Oximetry During Maximal Exercise in Normoxia, Hypoxia, and Hyperoxia. J Appl Physiol 2002; 92: Comtois G, Mendelson Y, Ramuka P. A Comparative Evaluation of Adaptive Noise Cancellation Algorithms for Minimizing Motion Artifacts in a Forehead-Mounted Wearable Pulse Oximeter. Proceedings of the 29th Annual International Conference of the IEEE 2007; Lyon, France Gehring H, Hornberger C, Matz H, Konecny E, Schmucker P. The Effects of Motion Artifact and Low Perfusion on the Performance of a New Generation of Pulse Oximeters in Volunteers Undergoing Hypoxemia. Resp. Care 2002; 47(1): Gilman MB, Wells CL. The Use of Heart Rates to Monitor Exercise Intensity in Relation to Metabolic Variables. Int J Sports Med 1993; 14(6): Goodie JL, Larkin KT, Schauss S. Validation of Polar Heart Rate Monitor for Assessing Heart Rate During Physical and Mental Stress. J. Psychophysiology 2001; 14(3):

5 APPENDIX A: Summary Statistics and Individual Data Plots of PerformTek Earbud Heart Rate Sensor Utilizing 12 Lead ECG Heart Rate SUMMARY STATISTICS Table A1. Validity summary Trial 1 Device Heart Rate Mean ± SD n=7440 ECG (bpm) ±28.70 PerformTek (bpm) Coefficient of determination (r 2 ) (vs. ECG) Standard error of estimate (SEE) Bias 95% limits of agreement ± Lower: Upper: 8.18 Table A2. Test-retest reliability Trial 1 vs. Trial 2 Device Correlation (R) between Trial 1 and 2 ECG (bpm) 0.95 PerformTek 0.93 (bpm)

6 INDIVIDUAL DATA PLOTS S1 S2 S3 S4

7 S5 S6 S7 Poor ECG data participant, hardware issues during T1 Participant unable to attend trial 2

8 S8 S9

9 S10 S11 S12

10 S13 Participant unable to attend trial 2 S14 S15

11 S16 S17 Participant unable to attend trial 2 S18

12 S19- Participant unable to attend trial 2 S20 Participant had small ears and subsequent poor fit of earbud. Trial 2 ended due to inability to keep earbud in ear. S21

13 S22 S23 S24

14 S25 S26 S27

15 S28 Participant unable to attend trial 2 S29 S30

16 S31 S32 S33

17 S34 S35 S36

18 S37 S38 S39

19 S40 S41 S42 S43 Participants ear anatomy did not allow for fit of earbud