Why do I put this item up?


10.2196/mhealth.6893


Original Paper
Clinical Evaluation of the Measurement Performance of the Philips Health Watch: A Within-Person Comparative Study

• Jos Hendrikx1*, MSc ;
• Loes S Ruijs2*, MSc, PDEng ;
• Lieke GE Cox2*, PhD ;
• Paul MC Lemmens2*, PhD ;
• Erik GP Schuijers2*, MSc ;
• Annelies HC Goris1*, PhD

1Innovation Site Eindhoven, Philips, Eindhoven, Netherlands

2Philips Research, Eindhoven, Netherlands

*all authors contributed equally

Corresponding Author:
Lieke GE Cox, PhD



Philips Research

High Tech Campus 34

Eindhoven, 5656AE

Netherlands

Phone: 31 40 27 91111

Fax:31 40 27 91111

Email: lieke.cox [at] philips.com

ABSTRACT
Background: Physical inactivity is an important modifiable risk factor for chronic diseases. A new wrist-worn heart rate and activity monitor has been developed for unobtrusive data collection to aid prevention and management of lifestyle-related chronic diseases by means of behavioral change programs.

Objective: The objective of the study was to evaluate the performance of total energy expenditure and resting heart rate measures of the Philips health watch. Secondary objectives included the assessment of accuracy of other output parameters of the monitor: heart rate, respiration rate at rest, step count, and activity type recognition.

Methods: A within-person comparative study was performed to assess the performance of the health watch against (medical) reference measures. Participants executed a protocol including 15 minutes of rest and various activities of daily life. A two one-sided tests approach was adopted for testing equivalence. In addition, error metrics such as mean error and mean absolute percentage error (MAPE) were calculated.

Results: A total of 29 participants (14 males; mean age 41.2, SD 14.4, years; mean weight 77.2, SD 10.2, kg; mean height 1.8, SD 0.1, m; mean body mass index 25.1, SD 3.1, kg/m2) completed the 81-minute protocol. Their mean resting heart rate in beats per minute (bpm) was 64 (SD 7.3). With a mean error of −10 (SD 38.9) kcal and a MAPE of 10% (SD 8.7%), total energy expenditure estimation of the health watch was found to be within the 15% predefined equivalence margin in reference to a portable indirect calorimeter. Resting heart rate determined during a 15-minute rest protocol was found to be within a 10% equivalence margin in reference to a wearable electrocardiogram (ECG) monitor, with a mean deviation of 0 bpm and a maximum deviation of 3 bpm. Heart rate was within 10 bpm and 10% of the ECG monitor reference for 93% of the duration of the protocol. Step count estimates were on average 21 counts lower than a waist-mounted step counter over all walking activities combined, with a MAPE of 3.5% (SD 2.4%). Resting respiration rate was on average 0.7 (SD 1.1) breaths per minute lower than the reference measurement by the spirometer embedded in the indirect calorimeter during the 15-minute rest, resulting in a MAPE of 8.3% (SD 7.0%). Activity type recognition of walking, running, cycling, or other was overall 90% accurate in reference to the activities performed.

Conclusions: The health watch can serve its medical purpose of measuring resting heart rate and total energy expenditure over time in an unobtrusive manner, thereby providing valuable data for the prevention and management of lifestyle-related chronic diseases.

Trial Registration: Netherlands trial register NTR5552; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=5552 (Archived by WebCite at http://www.webcitation.org/6neYJgysl)


I put it up to show how medical grade devices better be accurate. They will get tested, and they just have to perform or they will NOT be adopted into a health care system. So if a Tier 1 changes a sensor- good for them. So they test it with 100 people. Hopefully they do it as in this study-compare the accuracy to a gold standard. Just like this study....these devices will be tested before adopted...