Help! Too Many Conflicting Reviews for Smart Scales
Original Paper
Abstract
Background: Smart scales are increasingly used at home by patients to monitor their body weight and trunk composition, only scale accuracy has not often been documented.
Objective: The goal of the research was to determine the accuracy of 3 commercially bachelor smart scales for weight and body limerick compared with dual x-ray absorptiometry (DEXA) as the gold standard.
Methods: We designed a cantankerous-sectional study in consecutive patients evaluated for DEXA in a physiology unit in a tertiary infirmary in France. At that place were no exclusion criteria except patient declining to participate. Patients were weighed with one smart calibration immediately after DEXA. Three scales were compared (scale 1: Body Partner [Téfal], scale ii: DietPack [Terraillon], and calibration three: Torso Cardio [Nokia Withings]). We determined accented error between the gilt standard values obtained from DEXA and the smart scales for body mass, fat mass, and lean mass.
Results: The sample for analysis included 53, 52, and 48 patients for each of the 3 tested smart scales, respectively. The median accented error for torso weight was 0.iii kg (interquartile range [IQR] –0.one, 0.7), 0 kg (IQR –0.iv, 0.3), and 0.25 kg (IQR –0.x, 0.52), respectively. For fat mass, accented errors were –ii.2 kg (IQR –5.8, 1.3), –4.four kg (IQR –six.vi, 0), and –three.7 kg (IQR –8.0, 0.28), respectively. For muscular mass, absolute errors were –two.ii kg (IQR –5.8, ane.3), –4.iv kg (IQR –six.6, 0), and –3.65 kg (IQR –eight.03, 0.28), respectively. Factors associated with fat mass measurement error were weight for scales i and two (P=.03 and P<.001, respectively), BMI for scales ane and 2 (P=.034 and P<.001, respectively), body fatty for scale 1 (P<.001), and muscular and os mass for scale 2 (P<.001 for both). Factors associated with muscular mass error were weight and BMI for scale one (P<.001 and P=.004, respectively), body fat for scales one and 2 (P<.001 for both), and muscular and bone mass for calibration 2 (P<.001 and P=.002, respectively).
Conclusions: Smart scales are not accurate for body limerick and should not replace DEXA in patient intendance.
Trial Registration: ClinicalTrials.gov NCT03803098; https://clinicaltrials.gov/ct2/testify/NCT03803098
doi:10.2196/22487
Keywords
Introduction
Cellular-continued scales, familiarly known as smart scales, are increasingly used at abode for weight follow-up. They have been shown to increase the frequency of self-weighing and weight loss [,]. They can exist connected to other smart objects, such equally move sensors, and thus may help subjects engage in greater concrete activeness and better nutritional habits. Most available smart scales combine a classic weight calibration with a foot-to-foot impedance meter (FFI) that can estimate body composition (ie, fat mass [FM] and fat-costless mass [FFM]) by measuring foot-to-foot impedance at different frequency. Whole body FFM is calculated from a model comprising torso impedance, superlative, weight, and age trained with dual x-ray absorptiometry (DEXA) data []. Smart scales are easier to use than medical impedance meters since they do not require a supine position and their electrodes are indefinitely reusable. But the accuracy of smart scales depends on the representativeness of the patient population used to railroad train the model and the model itself. Although some FFIs take been compared with DEXA and to medical impedance meters [,], no information are available for smart calibration FFIs, and the regression equations used are unknown. The purpose of this cross-sectional study was to assess metrologic accurateness of 3 commercially available smart scales compared with DEXA every bit the gold standard in a population of adult patients from a third hospital physiology unit.
Methods
Patients
Consecutive adult patients evaluated for body composition past DEXA at the Bichat Hospital during the report were eligible. Patients were referred for DEXA because of obesity or a chronic status that tin can affect body composition (eg, chronic kidney disease, long-term steroids). All patients were included in the assay except those who refused to participate or whose weight measured by Lunar iDXA (GE Healthcare) exceeded the maximum weight tolerated by the smart calibration (eg, 160 kg for DietPack and 180 kg for Torso Partner and Body Cardio). Weight and body limerick past DEXA was performed by a trained technician co-ordinate to current practice on a Lunar iDXA. Iii smart scales were assessed (Trunk Partner [Téfal], DietPack [Terraillon], and Body Cardio [Nokia Withings]) during three consecutive periods (ie, merely ane scale was tested per patient for feasibility reasons). Patients were not required to be fasting. DEXA and assessment by the smart scale were performed with the patient wearing a hospital dispensable gown and their undergarment. A total of 55 patients were planned for each scale.
Statistics
Patient characteristics were described as median and interquartile range for quantitative variables and percentages for categorical variables. Accented and relative errors for each scale with respect to DEXA were reported with median and interquartile range. Bland-Altman representations were used to show systematic bias or trend in measurement error. Univariate linear models were used to estimate and examination the association between measurement error and possible associated variables. For this assay, we reported the slope estimation with its 95% confidence interval and the P value respective to the Wald statistic. We performed no imputation for missing data. The significance threshold was .05. All analyses were performed using R software version 4.0.2 (R Foundation for Statistical Computing).
Ethical Aspects
This written report is part of the Evaluation of the Metrological Reliability of Connected Objects in the Measurement of Medical Physiological Parameters (EvalExplo) study [NCT03803098]. Ethics approval was obtained from Comité de Protection des Personnes Sud Est Vi (approval number AU 1443), and written nonopposition was obtained.
Results
Patient Characteristics
The final sample for analysis included 53, 52, and 48 patients for each scale, respectively, after taking into account missing data (eg, smart scale not retrieving data despite several attempts for all merely one who was excluded because of excessive weight for the scales). Patient characteristics are presented in . In that location were no significant differences between the 3 groups.
| Characteristic | Calibration ane: Trunk Partner (n=53) | Scale 2: DietPack (n=52) | Scale 3: Torso Cardio (north=48) | |
| Sex, male, northward (%) | 17 (32) | 27 (52) | 32 (67) | |
| Age in years, median (IQRa) | 46 (37-58) | 48 (40-56) | 51 (43-62) | |
| Weight (kg), median (IQR) | 83.iv (71.4-102.1) | 80.five (69.2-99.3) | 82.2 (75-94) | |
| Height (cm), median (IQR) | 166 (162-173) | 169.5 (164-175) | 166 (159-173) | |
| BMI (kg/m2), median (IQR) | 30.3 (25.0-37.6) | 28.iv (24.8-34.ane) | 31.two (26.ane-35.5) | |
| | BMI < 25, n (%) | 13 (25) | 14 (27) | 9 (19) |
| | 25 ≤ BMI < thirty, n (%) | 11 (21) | sixteen (31) | 13 (27) |
| | thirty ≤ BMI < 35, due north (%) | 14 (26) | 11 (21) | 12 (25) |
| | BMI ≥ 35, n (%) | 15 (28) | 11 (21) | 14 (29) |
| Body fat (kg), median (IQR) | 33.v (21.5-49.viii) | 31.two (23.5-43.iv) | 28.5 (16.0-41.four) | |
| Muscular mass (kg), median (IQR) | 49.ane (42.0-56.1) | 52.0 (43.eight-57.1) | 47.half dozen (42.viii-53.8) | |
| Bone mass (kg), median (IQR) | ii.77 (two.23-three.02) | 2.74 (2.33-3.21) | 2.56 (ii.17-2.93) | |
| Excess mass (kg), median (IQR) | one.04 (0.84-i.29) | 0.98 (0.81-1.27) | 4.8 (2.2-6.2) | |
aIQR: interquartile range.
Accuracy of the Scales
All 3 scales gave rather accurate weights, with a median absolute difference of less than a kilogram compared with DEXA. The median accented error for body weight was 0.three kg (interquartile range [IQR] –0.1, 0.7), 0 kg (IQR –0.4, 0.3), and 0.25 kg (IQR –0.1, 0.5) for Body Partner (calibration one), DietPack (scale 2), and Body Cardio (scale 3), respectively.
For fatty mass, absolute errors were –2.2 kg (IQR –five.8, 1.3), –4.4 kg (IQR –six.6, 0), and –iii.seven kg (IQR –8.0, 0.3), respectively. Body fat was globally underestimated in all iii scales. For muscular mass, accented errors were four.50 kg (IQR 0.4, 7.3), –half dozen.half dozen kg (IQR –9.4, –3.6), and 4.0 kg (IQR 0.i, 7.half-dozen), respectively. Muscular mass was thus underestimated by scale 2 and overestimated by scales 1 and 3.
Bland-Altman graphs are presented in for weight, torso fat, and muscular mass for the 3 scales. They evidence a significant linear trend of body weight on measured weight, but absolute errors remain reasonable and compatible with clinical practice. However, pregnant linear trends for fat and muscular mass (scales ane and 2, respectively) lead to very high errors.
Factors Associated With Measurement Fault
and show the factors associated with fatty and muscular mass measurement error. Factors associated with fat mass measurement mistake were weight for scales 1 and 2 (P=.03 and P<.001, respectively), BMI for scales 1 and 2 (P=.034 and P<.001, respectively), torso fatty for scale ane (P<.001), and muscular and bone mass for scale 2 (P<.001 and P<.001, respectively). Factors associated with muscular mass error were weight and BMI for scale 1 (P<.001 and P=.004, respectively), torso fat for scales one and two (P<.001 and P<.001, respectively), and muscular and os mass for calibration 2 (P<.001 and P=.002, respectively).
We found no gene associated with measurement mistake of fat or muscular mass for calibration iii. Sexual practice did not evidence a meaning influence on measurement error.
| Characteristic | Calibration one: Body Partner (n=53) | Scale 2: DietPack (n=52) | Calibration iii: Body Cardio (n=48) | |||||||||
| | Approximate | 95% CI | P value | Estimate | 95% CI | P value | Guess | 95% CI | P value | |||
| Sex | —a | — | .96 | — | — | .24 | — | — | .57 | |||
| | Female | i (ref) | — | — | 1 (ref) | — | — | one (ref) | — | — | ||
| | Male | 0.095 | –3.5 to 3.seven | — | 2.one | –one.4 to 5.7 | — | 1.3 | –3.4 to half dozen.0 | — | ||
| Historic period in years | 0.023 | –0.091 to 0.14 | .69 | 0.042 | –0.088 to 0.17 | .52 | –0.064 | –0.25 to 0.12 | .49 | |||
| Weight (kg) | –0.065 | –0.12 to –0.0 | .03 | 0.18 | 0.11 to 0.24 | <.001 | 0.079 | –0.032 to 0.190 | .16 | |||
| Height (cm) | 0.024 | –0.15 to 0.20 | .79 | 0.054 | –0.17 to 0.28 | .64 | 0.2 | –0.035 to 0.43 | .09 | |||
| BMI (kg/one thousand²) | — | — | .03 | — | — | <.001 | — | — | .83 | |||
| | BMI<25 | 1 (ref) | — | — | i (ref) | — | — | 1 (ref) | — | — | ||
| | 25<BMI<30 | ii | –ii.8 to 6.7 | — | i.two | –2.viii to 5.2 | — | 0.97 | –v.eight to –viii.four | — | ||
| | thirty<BMI<35 | –2.8 | –7.ii to 1.vi | — | iii.8 | –0.55 to 8.ii | — | –ane.5 | –eight.4 to 5.iv | — | ||
| | BMI>35 | –four.4 | –8.7 to –0.013 | — | 9.8 | 5.42 to 14.2 | — | –1.four | –8.0 to v.iii | — | ||
| Torso fat (kg) | –0.27 | –0.40 to –0.14 | <.001 | –0.053 | –0.23 to 0.12 | .54 | 0.08 | –0.07 to 0.23 | .29 | |||
| Muscular mass (kg) | 0.0081 | –0.16 to 0.17 | .92 | 0.42 | 0.29 to 0.54 | <.001 | 0.xvi | –0.084 to 0.41 | .19 | |||
| Bone mass (kg) | –0.45 | –three.4 to 2.5 | .76 | 6.two | 3.2 to 9.3 | <.001 | 2.iv | –ane.half dozen to half dozen.3 | .24 | |||
aNot applicable.
| Characteristic | Calibration 1: Torso Partner (northward=53) | Scale 2: DietPack (n=52) | Scale 3: Body Cardio (n=48) | ||||||||||||||||||
| | Estimate | 95% CI | P value | Approximate | 95% CI | P value | Estimate | 95% CI | P value | ||||||||||||
| Sexual practice | —a | — | .52 | — | — | .07 | — | — | .47 | ||||||||||||
| | Male | 1 (ref) | — | — | 1 (ref) | — | — | one (ref) | — | | |||||||||||
| | Female | 1.i | –2.4 to 4.seven | — | –two.6 | –v.47 to 0.26 | — | –ane.vii | –6.3 to 2.9 | | |||||||||||
| Age in years | –0.039 | –0.150 to 0.072 | .49 | 0.098 | –0.0054 to 0.20 | .06 | 0.073 | –0.11 to 0.25 | .42 | ||||||||||||
| Weight (kg) | 0.xi | 0.054 to 0.16 | <.001 | –0.048 | –0.12 to 0.02 | .xvi | –0.068 | –0.18 to 0.042 | .22 | ||||||||||||
| Top (cm) | 0.085 | –0.087 to 0.26 | .33 | –0.fifteen | –0.331 to 0.032 | .10 | –0.2 | –0.43 to 0.025 | .08 | ||||||||||||
| BMI (kg/chiliad²) | — | — | .003 | — | — | .87 | — | — | .77 | ||||||||||||
| | BMI<25 | 1 (ref) | — | — | 1 (ref) | — | — | 1 (ref) | — | — | |||||||||||
| | 25<BMI<30 | –i.9 | –6.three to 2.5 | — | –0.43 | –iv.4 to 3.5 | — | –0.41 | –seven.0 to 6.2 | — | |||||||||||
| | thirty<BMI<35 | 1.9 | –2.2 to 6.0 | — | –1.67 | –6.0 to 2.seven | — | ii | –iv.vii to eight.seven | — | |||||||||||
| | BMI>35 | 5.8 | one.8 to 9.nine | — | –1.eleven | –5.5 to 3.2 | — | 2.fourteen | –4.4 to eight.7 | — | |||||||||||
| Body fatty (kg) | 0.28 | 0.fifteen to 0.xl | <.001 | 0.23 | 0.10 to 0.36 | <.001 | –0.063 | –0.21 to 0.085 | .40 | ||||||||||||
| Muscular mass (kg) | 0.xi | –0.047 to 0.27 | .17 | –0.32 | –0.42 to –0.21 | <.001 | –0.15 | –0.40 to 0.089 | .21 | ||||||||||||
| Bone mass (kg) | 2.two | –0.64 to 4.94 | .13 | –four.2 | –vi.nine to –1.6 | .002 | –two.1 | –6.0 to i.9 | .30 | ||||||||||||
aNot applicable.
Discussion
Primary Findings
To our knowledge, this is the first written report to assess metrologic accuracy of commercially bachelor smart scales (scale 1: Body Partner, calibration ii: DietPack, scale 3: Trunk Cardio). We show that all scales were reasonably accurate for torso weight but not body composition. Total body weight was associated with fatty mass and muscular measurement error for scales i and 2, simply we were not able to find factors associated with mistake measurement for calibration three.
Possible Explanations for the Lack of Accuracy
Smart scales combine a classic weight scale and an FFI, which has been widely available for almost 3 decades and has been compared with medical impedance meters and DEXA in several publications [-]. They have been shown to be more than sensitive to differences in morphology than whole trunk impedance measurements, since their information depend upon an extrapolation of measurements fabricated on the lower part of the body to the entire torso. Indeed, Bousbiat et al [] conducted an extensive technical study on FFIs. They plant that measurement can be affected by surface contact (ie, foot size and width) and sweat simply also by human foot position on the scale and flexion of the legs. Surface contact with the electrodes will vary depending on the discipline'southward pes length and width, and thus can be affected past total body weight and full height. Since at that place is no precise guidance on the scales for subjects on where to put their feet during body composition estimation, this may partly explain the differences between DEXA and scales. In clinical settings, clinicians or technicians should pay attention to the field of study'southward position during measurement. In the same style, subjects should be advised non to bend their legs. At dwelling house, subjects should try to follow directions given by the scale as closely as possible and continue the aforementioned position on the scale for follow-up.
In FFIs, whole torso FFM is calculated from a regression equation expressing resistance more often than not as a function of height, weight, and age determined past comparison with DEXA information, while FM is calculated as the weight of FFM. Each smart scale has its own regression equation. It is thus plausible that BMI tin can touch mistake measurement. Nosotros did not find whatever factor associated with mistake measurement for calibration 3 despite a very loftier dispersion, which can take different explanations: our study may be underpowered to detect such association or unobserved variables not part of the undercover regression implemented in the smart scale may explicate the residual error.
Clinical Relevance of These Data
Weight was accurately measured by all 3 scales, only body fat was underestimated. For scales 1 and 2, nosotros plant a meaning effect of higher trunk weight on fat mass error; this error remains pocket-sized compared with total body weight in patients with obesity merely can be of importance in patients with normal or underweight. Ross et al [] compared the weight from in-person visits and BodyTrace brand smart scales in 58 patients and found a hateful bias of 1.one (SD 0.8) kg, 95% CI 0.five to 2.6, but understanding seemed lower for patients weighing to a higher place 110 kg.
It is unlikely that trunk composition will be followed by DEXA in the clinical setting, since DEXA uses x-rays. Since smart scales are widely available, information technology is possible that patients will follow their body composition at home. Thus, if the first body limerick is assessed by DEXA, clinicians and patients should be aware that in that location might be a difference in body composition, which can reach 1 kg. Likewise, follow-up in the clinical setting should use the same connected device to ensure repeatability of measurements.
Interest for Follow-Up
Despite very poor accuracy for estimating torso limerick, some authors reported potential interests of in-domicile utilise of smart scales. Indeed, in several randomized studies, when compared with commercial weight management programs or standard weight loss counseling, smart scale utilise was found to let a greater proportion of participants to achieve significant weight loss after several months (3 to 12 months) [,,]. In these studies, no information were available on torso limerick and its development. Likewise, several studies written report a greater weight loss in patients using smart scales than in patients self-reporting weight loss or being weighted merely during visits [,]. Last, data exist on the importance of weight variability in terminal weight loss and weight maintenance [-]; using smart scales at habitation with automated data treatment could assist health care professionals and patients in achieving and maintaining greater weight loss [].
Strengths and Limitations
To our cognition, this is the first cantankerous-sectional study assessing metrologic accuracy of commercially available smart scales. We compared these scales to DEXA, which is considered the aureate standard for body composition assessment.
Included patients were evaluated for body limerick either because of obesity or because of a chronic condition (steroids, chronic renal failure, etc) without obesity. Thus, our sample covers a wide range of trunk weight: underweight, lean, and obese patients. The setting of the study in a third hospital explains the relatively small proportion of patients with obesity since the master reason for prescribing DEXA was long-term treatment with steroids. A specific report focusing on patients with obesity and farthermost obesity would be required since they are likely to do good the most from torso composition evaluation and follow-upwardly and since the error on fat mass is affected by total weight. This should exist done keeping in listen the maximum weight supported past the scales (180 kg for Body Cardio and Body Partner and 160 kg for DietPack), while DEXA supports higher weight (230 kg on our auto).
Finally, although torso limerick interpretation is quick with the smart scales, total experimental time was 15 to xxx minutes for each calibration due to scale and app setup and patient information. Thus, it was not possible to evaluate the 3 scales on the same patient or replicate measures. Even so, due to the very poor accurateness of these scales shown in this study, comparisons between scales and estimation of intraindividual variability would accept limited relevance. This written report included only a limited number of patients. All the same, while information technology is e'er possible that the studied scales would perform better on an independent sample, the written report results are so clear that it is also unlikely that a higher sample size could change our conclusion relative to the accuracy of body composition estimation.
Conclusion
Our study shows that although smart scales are accurate for total trunk weight, they should non be used routinely to assess body composition, especially in patients with severe obesity. Further studies are needed to clarify their utility in patient follow-up.
Acknowledgments
The authors are thankful to Najat Benlahmar and Fedja Kerzabi for information conquering and Chanez Chemani, Caroline Quintin, and Camille Couffignal for helpful methodologic word. The EvalExplo study was funded through a donation from MSD France to the Assistance Publique Hôpitaux de Paris Foundation. The authors are thankful to François Crémieux for his support in creating the Digital Medical Hub.
Conflicts of Involvement
JFM has received travel, accommodation, and meeting expenses unrelated to this work from Vitalaire, LVL Medical, Oxyvie, Ixxair Medical, and Boehringer-Ingelheim. No conflicts were declared for other authors.
Abbreviations
| DEXA: dual x-ray absorptiometry |
| FFI: foot-to-human foot impedance meter |
| FFM: fat-gratuitous mass |
| FM: fat mass |
| IQR: interquartile range |
Edited past G Eysenbach; submitted 13.07.20; peer-reviewed by J Rausch, M Amini; comments to author 22.09.twenty; revised version received xiii.11.20; accepted 12.12.twenty; published 30.04.21
Copyright©Justine Frija-Masson, Jimmy Mullaert, Emmanuelle Vidal-Petiot, Nathalie Pons-Kerjean, Martin Flamant, Marie-Pia d'Ortho. Originally published in JMIR mHealth and uHealth (https://mhealth.jmir.org), 30.04.2021.
This is an open up-access article distributed nether the terms of the Creative Eatables Attribution License (https://creativecommons.org/licenses/past/four.0/), which permits unrestricted use, distribution, and reproduction in whatever medium, provided the original piece of work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, besides equally this copyright and license information must be included.
Source: https://mhealth.jmir.org/2021/4/e22487
0 Response to "Help! Too Many Conflicting Reviews for Smart Scales"
Post a Comment