How To Synch Healbe Gobe2 With Phone
JMIR Mhealth Uhealth. 2020 Jul; 8(7): e16405.
Wearable Technology to Quantify the Nutritional Intake of Adults: Validation Study
Monitoring Editor: Gunther Eysenbach
Sarah M Dimitratos, BSc, RD,ane J Bruce German, DPhil,1 and Sara E Schaefer, DPhil 
1
one Foods for Health Found, University of California, Davis, CA, Us
Sara E Schaefer, Foods for Health Establish, University of California, 2141 Robert Mondavi Institute, North Building, one Shields Ave, Davis, CA, 95616, Us, Phone: ane 530 574 0797, Email: ude.sivadcu@refeahcses.
Sarah One thousand Dimitratos
one Foods for Health Institute, Academy of California, Davis, CA, United States
J Bruce German
1 Foods for Health Establish, University of California, Davis, CA, United states
Sara E Schaefer
1 Foods for Wellness Establish, University of California, Davis, CA, U.s.
Received 2019 Sep 26; Revisions requested 2019 November eighteen; Revised 2020 Jan xiii; Accepted 2020 Apr x.
Abstract
Background
Wear and mobile sensor technologies can be useful tools in precision nutrition research and practise, but few are reliable for obtaining accurate and precise measurements of diet and nutrition.
Objective
This study aimed to appraise the ability of wearable applied science to monitor the nutritional intake of adult participants. This paper describes the development of a reference method to validate the wristband's interpretation of daily nutritional intake of 25 gratis-living study participants and to evaluate the accuracy (kcal/solar day) and practical utility of the technology.
Methods
Participants were asked to utilise a nutrition tracking wristband and an accompanying mobile app consistently for ii xiv-twenty-four hours exam periods. A reference method was developed to validate the interpretation of daily nutritional intake of participants by the wristband. The research squad collaborated with a university dining facility to fix and serve calibrated report meals and record the energy and macronutrient intake of each participant. A continuous glucose monitoring system was used to measure adherence with dietary reporting protocols, but these findings are not reported. Bland-Altman tests were used to compare the reference and examination method outputs (kcal/day).
Results
A total of 304 input cases were collected of daily dietary intake of participants (kcal/mean solar day) measured past both reference and exam methods. The Banal-Altman analysis had a mean bias of −105 kcal/twenty-four hours (SD 660), with 95% limits of agreement between −1400 and 1189. The regression equation of the plot was Y=−0.3401X+1963, which was significant (P<.001), indicating a tendency for the wristband to overestimate for lower calorie intake and underestimate for higher intake. Researchers observed transient betoken loss from the sensor applied science of the wristband to be a major source of fault in computing dietary intake among participants.
Conclusions
This study documents loftier variability in the accurateness and utility of a wristband sensor to track nutritional intake, highlighting the need for reliable, effective measurement tools to facilitate accurate, precision-based technologies for personal dietary guidance and intervention.
Keywords: wearable technology, mobile wellness, mobile phone, nutrient intake, validation written report
Introduction
Diet and wellness guidelines are based on preventing or treating illnesses in the full general population. Technological advances and enhanced understanding of systems biology are guiding scientists to pursue personalized interventions for disease prevention and treatment. Every bit scientists are quantifying the elasticity of homo wellness and its diverseness, opportunities to arbitrate in human health are broadening to include precision control of phenotypic performance. Precision or personalized wellness is the approach of using quantified data on individual characteristics to develop tailored products and services aimed at guiding the underlying processes of wellness [ane-5]. The breadth of precision interventions includes the measurement of individuals' characteristics, genetics, immunity, metabolism, physiology, medical history, and more [6-8]. Personalizing the content and delivery of approaches too require alignment with individuals' behaviors, preferences, goals, and barriers to modification every bit an integral aspect of achieving lasting behavior change [ii,9,10].
Precision health is made possible past modern tools, technologies, and platforms that provide increasingly diverse, mechanistic, and accurate assessments of the human body [11,12]. Health measurement research encompasses the breadth of phenotypic differences between individuals that contribute to health condition. Advancements in the -omics sciences highlight how many factors individually and interactively bear on health, including genetics, lifestyle, life phase, diet, and microbial variety. Many wellness metrics are assessed statically, only others must be captured dynamically using specific challenges, such equally with insulin sensitivity and caused amnesty. These scientific breakthroughs are guiding the development of measurement technologies that interrogate individuals beyond illness diagnostics, including mobile and wearable torso sensors that enable more than spatially and temporally specific measures of a broader range of phenotypic factors [4,xiii-15].
The most important change in the science of nutrition and health is as much philosophical as mechanistic. The focus of nutrition enquiry is shifting from the study of individual foods and ingredients and their furnishings on entire populations to the study of private humans in response to entire diets.
Precision Nutrition: Challenges and Breakthroughs
Bringing authentic wellness monitoring technologies to the market provides a public service that reduces people's doubt well-nigh how 24-hour interval-to-day choices affect their individual health [fifteen,xvi]. More precise and predictive dietary guidance follows the understanding in nutritional sciences that identifying the unmarried all-time diet for human being health is no longer scientifically defensible. It is now understood that different people respond differently to foods and nutrients, warranting personalized approaches to diet interventions and services [three,17-21]. National dietary guidelines are intended to prevent deficiency and maintain wellness for the majority of the population. Using testify-based scientific discipline to create diets for individuals requires an understanding of what humans share with regard to dietary needs too every bit how, when, and why needs differ. On a fundamental level, all people crave a nutrition sufficient in calories to support normal body weight and all essential nutrients to back up life. Notwithstanding, nutrient requirements to prevent deficiency and sustain life are just the first step in agreement the office of diet in human health [22].
A primal challenge in nutrition research is the accurate quantification of food intake and its interpretation equally precise nutrition quality. Currently, the golden standard of dietary assessment is the 24-hour in-patient written report, yet major limitations include cost, reduced physical activeness, boredom, depression, and weight loss because of reduced dietary freedom and food options deviating from one'south personal routine. In epidemiologic and clinical diet, dietary assessment typically relies on researcher-facilitated or autonomous participant recollect using methods such equally 24-hour retrieve, food frequency questionnaires, and food diary inventories. These memory-based assessment methods take demonstrated poor validity considering of man under- or overestimation of intake and intentional or unintentional alteration of intake patterns [23]. Each traditional assessment method is a reflection of the individual's perceived intake rather than an accurate mensurate of truthful intake. Furthermore, such cess methods are nonfalsifiable, as what the participant reports must be accustomed as truth, despite knowledge of likely incongruence. Moreover, other assessment methods rely on photograph analysis of foods consumed, conducted either by trained personnel or software analysis [24]. Although more closely reflective of nutrient intake in a free-living situation, the remote food photography method is yet limited by the disability to record in true real time, difficulty in estimating portion sizes, the necessity for simultaneous use of a backup analysis method, difficulty analyzing culturally unique foods, and analyzing mixed dishes via photographs lonely. The United States Department of Agriculture (USDA) Nutrient Limerick Database is the gold standard for nutrient analysis; admitting comprehensive in scope, this tool cannot mayhap account for inherent variability in climate, soil quality, geographic location, particular ripeness, and cooking method, all of which may significantly modify the nutrient composition of food. Even nutrient and energy quantification past way of nutrition facts label analysis is fault prone, as the Food and Drug Assistants allows for certain margins of error in food reporting on packaged nutrient labels. Diet fact labels, therefore, provide an educated estimation of packaged food nutrient content. Consumer-focused dietary tracking methods apply databases that are often crowdsourced. These errors, compounded with the aforementioned human misreporting of dietary intake, demonstrate that more precise methods of dietary assessment and analysis are needed.
Standard approaches for recording dietary intake do not account for inherent nutrient losses in absorption and metabolism, the transformative processes past which food becomes usable energy for the trunk. Realistic and precise quantitative cess remains challenging because of free energy losses involved at every step of transforming a food matrix into bioavailable energy: absorption, distribution, metabolism, and excretion. The rate of breakdown and net usable free energy vary depending on macronutrient composition (ie, a mixed meal high in cobweb, protein, and fat will digest much more than slowly than a meal high in simple carbohydrates) [25]. Furthermore, interindividual differences in metabolic rate, gastrointestinal health, and previous meals consumed all contribute to discrepancies between measured intake and bioavailable energy.
Emerging commercial and medical technologies designed to detect a person's physiological fluctuations claim to capture more dynamic aspects of cardiometabolic health [14]. For example, continuous glucose monitors are designed to provide more than precise tracking of glucose levels for diabetic patients compared with standard blood sampling methods, the goal being to more precisely guide affliction management [26,27]. No technologies are available that can effectively assess dietary intake directly, although some methods are claiming the ability to estimate dietary intake by assessing the physiological response of the body to food intake and bioavailable energy. In all cases, rigorous testing is necessary to decide the accuracy, precision, utility, and validity of candidate devices. We sought to answer the question, "can wear technologies measure aspects of metabolic performance and cardiometabolic health of a normal range of developed human being phenotypes?" The objectives of this paper were to describe (1) the development and implementation of a reference method to estimate the nutritional intake of free-living study participants and (2) the accuracy and utility of a wristband technology for tracking nutritional intake (kcal/solar day).
Methods
Overview
A study was designed to assess the power of wearable technology to judge the nutritional intake of individuals. The wristband (GoBe2; Healbe Corp) intends to provide users with automatic tracking of daily energy intake (calories) and macronutrient intake (grams of protein, fat, and carbohydrates). The technology uses computational algorithms to catechumen bioimpedance signals into measured patterns of extracellular and intracellular fluids associated with the influx of glucose and essential nutrients into the body. From changes in fluid concentration, the technology estimates calories coinciding with glucose absorption into the bloodstream. Time serial data such as these, which capture postprandial processes, take the potential to inform phenotypic discernment of digestion, assimilation, metabolism of foods, and their influence on health.
A sample of free-living adult participants (N=25) was sought to validate the engineering over 2 information collection periods of 14 days each (28 days total). A reference method was designed to mensurate dietary intake; all meals were prepared, calibrated, and served at a campus dining facility and consumed nether the direct ascertainment of a trained enquiry team. Approving for the research written report and protocol was obtained from the University of California, Davis (UC Davis), institutional review board.
Participants
Participants aged 18 to 50 years were recruited from the UC Davis campus using emails and flyers. Those interested were screened by phone for inclusionary and exclusionary criteria. The exclusion criteria included historical or electric current diagnosis of chronic disease (including diabetes or prediabetes, cancer, asthma, hypertension, cardiovascular affliction, stroke, kidney, thyroid, or autoimmune illness), known food allergies, electric current dieting or restricted dietary habits (ie, vegetarian, ketogenic, reduced calorie), pregnancy or lactation, smoking, drug or alcohol addiction, excessive exercise or athletic grooming, and taking medications impacting digestion or metabolism. In-person screenings were conducted at the Ragle Human Diet Center on the UC Davis campus. Participants who qualified subsequently the phone screening were invited for in-person screening to complete a fasting blood draw, blood force per unit area, and anthropometric measurements. Copies of approved, signed consent forms were obtained from all participants at screening. All female participants completed urine pregnancy tests. Claret pressure measurements were obtained using a Nellcor pulse oximeter with OxiMax technology from Welch Allyn. For anthropometry, a digital scale by Scale-Tronix was used to weigh participants to the nearest 0.1 kg, and a wall stadiometer was used to measure peak to the nearest 0.i cm. Anthropometric measurements were used to calculate baseline BMI (weight [kg]/[superlative (m)]2). Every bit the wristband was intended to measure nutrient intake in a weight-stable population over the study elapsing, individuals with fluctuating weight (>5 lbs over the previous month) were excluded. All anthropometric measurements were conducted by the master investigator (PI) using methods defined in the anthropometric standardization reference transmission [28]. Participants were assigned a study ID on enrollment, and all data collected were maintained private and deidentified. Monetary compensation was offered to each participant who completed the screening (U.s. $10), phase 1 (US $125), and phase ii (U.s.a. $150).
For metabolic screening, blood was drawn into ethylenediaminetetraacetic acid and plasma separation lithium heparin blood drove tubes and immediately placed on ice. Within 2 hours of collection, blood samples were centrifuged at 1800×g for 15 min at four°C to split up blood from plasma and frozen at −twenty°C until laboratory analyses were performed in batches. Blood samples were analyzed, and individuals who tested abnormally for metabolic health indicators including complete blood count, fasting claret glucose, hemoglobin A1c, erythrocyte sedimentation rate, serum poly peptide, creatine, alkaline phosphatase, potassium, and carbon dioxide were excluded. Tests were performed according to the manufacturer's instructions and quality controls by UC Davis Health System Medical Diagnostics.
Between August 2018 and September 2018, 76 adults were screened, and 35 met the inclusion criteria for enrollment in phase 1 of the study that would take place from September 25 to October nine, 2018. The initial sample included 20 women and 15 men, with an indigenous distribution of 38% white, 41% Asian, and 21% Hispanic, an average age of 25.iii (SD six.4) years, and a mean BMI of 24.2 (SD 5.1) kg/m2. 3 participants dropped out during the outset week of phase ane considering of time constraints that prohibited multiple visits to the campus dining facility each day. Stage 1 was completed by 32 participants, of which 24 enrolled in stage 2 (October 30 to November 13, 2018). During phase 2, 2 participants completed 10 of 14 days because of scheduling conflicts and were included in the analyses.
Data Collection
Participants were assigned a GoBe2 (Figure 1) and instructed to use the latest version of the accompanying app synchronized to the wrist unit. The engineering science translates sensor signals into energy intake and expenditure outputs over a 24-hour period, in accordance with the rate of food absorption. Participants received an explanation on how the wristband estimates personal calorie intake and expenditure throughout the solar day and over the week likewise as its other functions, including heart rate, slumber, hydration, and stress measurement. Participants were instructed to synchronize the wrist unit with the app twice daily, in the morning time and at nighttime, and to collect screenshots from within the app that captured the previous twenty-four hour period's final energy (kcal) estimations. The screenshots were collected by inquiry staff as records of daily caloric outputs, including daily intake, expenditure, and full residual.
Quantification of Dietary Intake
A reference method was developed to quantify the daily food, calorie, and macronutrient intake of participants during the 2 written report periods. The project team collaborated with UC Davis Dining Eatables (DC), a serial of dining facilities where campus residents primarily consume but are as well open up to the campus customs and public. A strategy was developed to carry out the nutrition report within the academy dining facility. In this approach, a specific project card was created in coordination with the facility'south existing cycle menu serving all dining patrons. In this style, the dining facility'due south normal operations were minimally perturbed, and the study squad used the facility'due south existing food prepared in accordance with standardized recipes from which nutritional data was readily derived. Repast cards were purchased for study participants and swiped on their arrival at each meal to deduct the repast price from the card. Student research assistants were trained to carry out food measurement at each meal, nutrient analysis, and information entry.
Menu Planning
A registered dietitian (RD) on the research team collaborated with the dining facility's primary chef to design the project menu. Carte items were selected to serve to study participants at breakfast, lunch, and dinner, using the following criteria: balanced macronutrients at each meal per USDA MyPlate guidelines and minimal multi-ingredient mixed dishes (ie, no casseroles, lasagna, pizza, etc). Mixed dishes were avoided to reduce error in computing calories and macronutrients that were served at each meal. When necessary, menu modifications were requested to fit the study carte du jour criteria (ie, sauces served on the side and sandwich ingredients served separately). Separating ingredients immune the staff to weigh foods more precisely and calculate energy and macronutrient profiles accordingly.
Free energy and Macronutrient Analysis for Onsite and Offsite Nutrient Consumption
Overall, 2 research staff were trained and designated to analyze each project card detail for energy (kcal) and macronutrient content. Items were analyzed using a combination of the USDA Food Composition Database and the dining facility's nutritional database. In the latter, bill of fare items were previously analyzed and recorded by the DC'due south RD using either product nutritional labels (when available) or the USDA Food Limerick Database. Each menu item was analyzed for serving size, calories, grams of protein, fat, and saccharide content per serving and scaled to 100 g.
The RD determined a standard serving size for each menu item (eg, 1 cup cooked oats, 1 cup vegetables, half loving cup beans, 4 oz lean poly peptide, or three-fourth cup grain). Participants were not restricted to the standard serving sizes and were free to request more or less food portions to run across their individual dietary needs. All deviations from standard portions were recorded by the research staff for each participant.
The primary chef coordinated study meals according to the report bill of fare preference. Each meal was prepared in a commercial kitchen on the UC Davis campus by trained food service personnel following a stringent hazard analysis critical command points (HACCP) protocol. All food was delivered to the designated research study area of the facility and received past a squad of research staff for onsite portioning and serving to written report participants. The study leads inspected each delivered menu item for accuracy, noting any deviations equally needed.
Report participants arrived at the dining facility during scheduled breakfast, tiffin, and dinner mealtimes. On arrival, they were greeted by research staff, and the repast was paid for at the door using preloaded meal cards. Each forenoon at breakfast, enquiry staff collected daily information from participants, including newspaper records of offsite foods consumed in the previous 24 hours and details of wristband utilize (charging, removals, and reported problems). A cursory daily in-person interview was conducted each morning to collect details on exercise, whatsoever perceived stress, h2o intake, defecation, and continuous glucose monitoring (CGM) skin contact in the previous 24 hours. At each repast, participants could request either the standard meal offering or certain menu items in more or fewer portions according to preference. The participants' meal choices were recorded on paper meal slips that were delivered to research staff responsible for food portioning, plating, and weighing.
All project staff were trained by the RD in advisable nutrient treatment and prophylactic, nutrient weighing, and meal recording duties. Earlier each meal, a team of research staff was briefed on how to portion and serve each menu detail. Individual menu items were weighed and recorded (0.0 g) using calibrated food scales, portioned using standardized tools, and served at each onsite meal. Each dish with multiple food components was deconstructed into individual items and was weighed and recorded individually (ie, burgers were deconstructed to individually weigh patty, bun, cheese, ketchup, mustard, and tomato). Staff causeless diverse roles to ensure optimal meal-time efficiency (ie, menu collector, food weigher, and data recorder). Later on recording the weight of each food item and time of meal (00:00), the plate was served to the appropriate participant. Participants were encouraged to consume all nutrient served at each meal, but this was not mandatory. The plate waste from each participant was deconstructed by ingredient and individually weighed at the stop of each repast period.
Later each participant finished eating, the research staff weighed and recorded each individual item left on the plate. The gram weight of each nutrient item consumed was quantified and entered into an electronic database. Free energy and macronutrient profiles of each menu item were obtained from the dining facility'south recipe, the food characterization, or the USDA Food Composition Database and calculated co-ordinate to the gram weight consumed.
Consuming foods outside of the study facility was discouraged just not prohibited to minimize the changes made to the participants' usual habits and metabolism. If food was consumed outside of the dining facility, participants were instructed to follow a specified procedure of self-reporting, including simply consuming packaged foods, weighing and recording the weight of each individual food detail, and providing the nutrient label from the package. To minimize the miscalculation of food intake of offsite foods, participants were provided with various packaged foods of known nutritional content (protein bars, jerky sticks, ramen noodles, fruit leather, and chocolate bars). They were asked to consume these foods; if this was not possible, they were required to photograph the nutrient and record the food item, brand, time of consumption, and food weight (thou) using a calibrated food scale and recording in a food diary. Offsite food diaries were collected daily from participants.
Participants unable to report to the dining facility for a scheduled meal time received alternative options and selected a prepackaged repast from a convenience market managed by the dining facility. Nutritional information from the item was extracted from the nutrition characterization and recorded. A squad of staff recorded and analyzed the nutritional values of all offsite foods consumed by the participants. Information from the food intake data of i participant, as measured by the report reference method, is presented in Table 1.
Tabular array ane
Daily nutrient intake tape of 1 study participant.
| Time of meal (00:00) | Menu item | Corporeality consumed (yard) | Energy intake (kcal) | Source of nutrition information |
| nine:34 AM | Scrambled eggs | 53 | 69 | Product labela |
| 9:34 AM | Cooked oatmeal | 189 | 105 | Product labela |
| 9:34 AM | Blueberries | 69 | 39 | USDAb Food Composition Database |
| nine:34 AM | Bacon | 15 | 56 | USDA Nutrient Composition Database |
| 9:34 AM | Milk one% | 0 | 0 | Product label |
| 9:34 AM | Coffee, fresh brewed | 246 | 0 | Northward/Ac |
| 9:34 AM | Granulated sugar | 10 | 23 | USDA Nutrient Composition Database |
| two:11 PM | Bun | 90 | 218 | Product labela |
| 2:11 PM | Beefiness, ground, cooked | 125 | 156 | Product labela |
| 2:11 PM | Sauce | 40 | 53 | Product labela |
| 2:11PM | Mixed greens | 57 | 16 | USDA Food Composition Database |
| 2:eleven PM | Artichoke hearts, canned | 21 | 6 | USDA Nutrient Composition Database |
| 2:11 PM | Ruby tomatoes | 44 | 12 | USDA Food Limerick Database |
| 2:11 PM | Cucumbers, sliced | 51 | 5 | USDA Food Composition Database |
| 2:11 PM | Carrots, shredded | 25 | x | Product label |
| 2:eleven PM | Olive oil | 12 | 96 | Production label |
| ii:11 PM | Balsamic vinegar | 19 | 17 | Product label |
| 6:46 PM | Craven tamales | 304 | 669 | Production labela |
| 6:46 PM | Vegetables, roasted | 250 | 143 | Product labela |
| six:46 PM | Rice, cooked | 61 | 105 | Product characterizationa |
| 6:46 PM | Milk ane% | 0 | 0 | Product characterization |
| 11 AM | Energy bar | 52 | 210 | Product label |
| 1 PM | Energy bar | 48 | 190 | Product label |
| 8 PM | Dehydrated soup | 64 | 290 | Product label |
| 10 PM | Energy bar | 68 | 290 | Product label |
| North/A | N/A | 1913d | 2753d | Northward/A |
Quality Assurance
Before this written report, the PI conducted a serial of small pilot trials over one year to inform this report design and information collection procedures using the wristband technology. During these airplane pilot trials, it was observed that the form factor of the engineering was the main bulwark to collecting consistent, uninterrupted data during the postprandial digestion menstruum that lasts several hours beyond each repast. Practically, whatsoever betoken interruption during the meal or in the hours following it would result in loss of information and underestimation of calorie intake by the applied science. Unfortunately, indicate intermission occurred oft and for a variety of reasons in this study; for case, periodic loss of contact with the skin was likely depending on the user's wrist size and shape. In addition, the wristband required an 60 minutes each day to obtain a full charge; any loss of charge would disable data collection accordingly. Several strategies were used to mitigate these challenges with the course factor. Participants were instructed to charge the wristband fully before any meal to avoid missing food intake and its subsequent digestion (ie, charging band in the morning before consuming food for the day). It was adequate to charge the wristband at whatever signal during the day as long as no food had been consumed for 3 hours prior. On arriving at the first meal of the 24-hour interval, the research staff visually confirmed that the wristband was positioned on each participant, such that the sensor was in consummate and constant contact. Inquiry staff used a tertiary-party site (Dietitian'south Cabinet) to admission participants' deidentified data up to the minute from which the frequency of contact interruptions could be assessed. Those who had meaning interruptions were targeted for individual solutions to better sensor contact with the wrist, for example, tightening the wristband to reach optimal sensor positioning.
Continuous claret glucose was monitored as a strategy to measure and account for nonadherence to the study's dietary intake reporting protocols. The FreeStyle Libre (FSL) Pro System (Abbott Diabetes Care Inc) CGM arrangement includes a unit with a water-resistant sensor that attaches to the back of the user's upper arm. Within the unit is an Enlite sensor that consists of a wire containing glucose oxidase at the tip that is inserted subcutaneously with a defended inserting device. Glucose oxidase catalyzes a biochemical reaction in the presence of glucose and oxygen, which transfers electrons to a receiving molecule and creates a electric current that can exist measured and converted into a glucose concentration [27]. The FSL Pro Organisation collects upwards to fourteen days of glucose readings, with recordings every fifteen min. A single reader can exist used to activate glucose data recording and download reports from multiple devices simultaneously. One written report showed that the FSL's hateful absolute relative departure compared with measured capillary blood glucose levels was xiii.2% (95% CI 12.0% to 14.four%) [29].
CGM sensors were secured to the tricep or rotator gage region of participants' arms on day 1 of the study, in the morn before consuming food or beverages. During the 14-24-hour interval test menses, units would occasionally become discrete. The research staff downloaded information files from the participants' sensors every 2 days to minimize whatsoever data loss. Text file reports were exported through the LibreView software program (Abbott Diabetes Intendance Inc, 2018) and a secure cloud-based system. CGM data were analyzed to appraise the adherence of individual participants to reference dietary intake reporting protocols. Meaning glucose increases (>20 mg/dL per 30 min) occurring exterior of specified study mealtimes or not reported in nutrient intake diaries were flagged for farther examination.
Statistical Analysis
The Bland-Altman analysis was conducted to compare daily energy intake (kcal/twenty-four hour period) estimated by both the reference method and the wristband technology. Regression analyses were used to examine trends in the information and sample characteristics. Statistics were conducted in Microsoft 2008 (version 12.three.1) and Prism 8 2019 (version 8.3.1).
Results
This study developed a dietary intake reference method to evaluate a wearable sensor with the potential to generate objective and precise data on the dietary intake of adult individuals. The data accuracy and practical application of the current GoBe2 model was interrogated over 2 14-day test periods in an intended sample of 25 participants. Of the 35 participants who were originally enrolled in phase 1 of the study, 304 measurements (kcal/day) collected from 24 participants were retained from phase ii after information cleaning to remove missing or aberrant values.
Of the full cases, x.9% (33/304) were excluded because they lacked an accompanying set of complete CGM data for the 24-hour period. Of the remaining cases, 22.1% (60/271) had at least ane event per mean solar day of rapid blood glucose increase that was inconsistent with the recorded repast time. Of those, 68.3% (41/sixty) were attributed to reported bouts of exercise or other physical activity. Although CGM was used to mensurate nonadherence to dietary intake reporting protocols, these data were not incorporated in the nowadays dataset.
As depicted in Figure ii, a Bland-Altman analysis showed a mean bias of −105 (SD 660) kcal/day, with 95% limits of agreement (LoA) between −1400 and 1189. Pearson correlation coefficient between the 2 methods was r=−0.496 (95% CI −0.576 to −0.406; P<.001). Linear regression analysis on the Banal-Altman plot revealed a regression equation of Y=−0.3401X+1963 that was significant (P<.001). A multiple regression analysis was conducted with the participants' age, sex, and BMI classification every bit contained confounding variables, simply no pregnant effects were seen on the bias. Analysis of variance tests were conducted to appraise the furnishings of the participants' age, sex activity, and BMI nomenclature on bias, and the effects were not meaning (P=.15, .18, and .12, respectively).
Banal-Altman (mean departure) plot of estimated nutrient intakes (kilo/day) past the test and reference method (North=304). Solid lines represent upper-lower limits of agreement, and the dashed line represents bias.
Give-and-take
Negative bias in the Bland-Altman analysis indicated a full general underestimation of daily calorie intake past the wristband compared with the reference method. Despite a relatively small bias, the LoAs were wide, making the results of the comparison ambiguous. Regression analyses indicated a tendency for the wristband to systematically overestimate for lower calorie intake and underestimate for higher intake.
Our preliminary validation results indicate that although the ability of GoBe2 to make phenotypic discernments responsive to diet by noninvasive ways has wide-reaching utility in research and practise, notable feasibility challenges were observed for free-living study participants to reliably utilize the technology to achieve accurate and precise measurements. These challenges were largely attributed to limitations in the technology's class factor. In ascertainment, when positioned correctly on the arm and fully charged, the wristband'south calorie intake estimates generally appeared accurate and provided interesting visuals pertaining to the body's charge per unit of nutrient absorption. Nonetheless, to achieve precise detection and accurate interpretation of dietary intake, the unit's sensor required adequate skin contact be maintained at all times. Achieving this proved to be a considerable challenge for several reasons, including (1) battery life, as the unit of measurement required an hour of charging each day, which required removal of the device, preventing the detection of calories ingested inside several hours before removal; (2) the wristband'south bulky size, dimensions, and/or advent were challenges for some users to maintain condolement and position on the arm; and (iii) the user'due south ain wrist size and shape; for example, small or tapered wrists were probable to upshot in inconsistent sensor contact. As described previously, several strategies were included in the study blueprint to prevent data loss, such as targeting problematic cases early, checking in with participants, and monitoring sensor position daily. However, data loss from poor sensor contact was a significant barrier to the technology'south ability to reliably find calorie intake. Separate analyses, non reported in this study, further examine the technology'due south efficacy using data nerveless only during periods of protocol adherence apropos nutrient reporting and applied science use.
Establishing reliable adherence or compliance protocols is a widespread goal in measuring the dietary intake of human subjects [23]. Continuous glucose monitors were used to measure the participants' adherence to food intake recording protocols. Although CGM data practice not provide a direct measure of dietary intake, its measurement of the body's relative physiological response to nutrient intake tin serve equally a proxy to identify inconsistencies in reported intake data and blood glucose activity. Test of CGM data confirmed that although a few participants (n=ii) were probable nonadherent with dietary intake reporting protocols, aberrant increases in blood glucose levels could be attributed to multiple factors including exercise or other bouts of physical activeness. The authors concluded that circuitous outcomes on CGM measurement and the participants' adherence would be appropriately detailed in the context of measuring or impacting compliance in nutrition research. Some challenges to using the CGM devices to collect data over continuous 14 days were too related to form cistron limitations. The sensor included an adhesive material attached to the skin, but some devices became dislodged during the 14-mean solar day study menses (thirteen/72, 18% CGMs attached), causing complete or partial data loss. Of the 24 participants, two (8%) had repeated CGM sensor displacement, which was more likely to occur during concrete activity (biking, gym workout, and weight lifting) and/or excessive sweating. In these cases, a skin adhesive (Pare-Tac) was useful in reinforcing the CGM zipper. As the FSL Pro Organization did not include individual readers with each unit, participants were blinded to their personal glucose data. At the end of the study catamenia, data reports summarizing glucose patterns were generated and distributed to participants. Readouts included daily blood glucose averages (g/ dL) beyond each 24-hour catamenia, average glucose tendency lines across each 24-hour period, and likelihoods of hypoglycemia or hyperglycemia during specified windows. A total of seven days' worth of daily claret glucose trend lines were color coded and superimposed onto summary graphs. Participants were provided with full general guidance from the RD to interpret numerical data into a relevant and actionable context for health and diet.
By collaborating with the university facility, this study used existing food production operations, resources, and personnel to deport out an all-encompassing dietary observation report. Despite numerous strengths in the study design and utilization of a novel research environment, limitations were revealed during project implementation. For instance, in the nutrient facility where dishes were prepared for high throughput mass consumption, the verbal quantity of nutrients in each portion could not be consistently and routinely ensured using these methods alone. In addition, because that the project targeted students on a university campus, protocol adherence was less than anticipated, specially with regard to meal attendance. Of the 42 full study meals offered to each of the 24 participants during the second 14-twenty-four hour period testing menses (1008 full meals), 56% of the scheduled meals were attended (565 meals). To improve adherence in the future, stricter enforcement of meal attendance is recommended. Studies excluding offsite food consumption may help amend the accurateness of food intake reporting, with strategies in place to account for protocol adherence. Given that numerous factors were involved with intermittent information loss from the wristband technology, two weeks was defined equally the minimum period required to gather continuous data from 25 free-living participants for validation purposes. Longer study periods could affect adherence problems without stricter guidelines around participant meal omnipresence.
This study validated participants' calorie intake equally recorded by the article of clothing device, in comparison with a reference diet. The deviations in and between methods could be explained by any combination of the post-obit factors: form factor limitations (skin contact/battery); the participants' nonadherence to dietary protocol (ie, consuming and failing to report ingested nutrient or drink); interindividual differences in measured intake versus actual nutrient absorption and metabolism; human error in calculating food intake using the USDA Database; potential deviations from the standardized recipe during the meal training procedure; disability of the USDA Database business relationship for nutrient variation depending on food ripeness, geographic location, and soil quality; inherent information loss considering of required 1-hour daily device charging periods; and inaccuracies pertaining to technology algorithm development. Time to come studies should contain these suggestions for improvement to further interrogate the potential of clothing devices to accurately capture caloric and macronutrient intake. Ongoing engineering adjustments are recommended to accurately approximate the energy and nutrient intakes of individuals consuming various diets.
Tools are urgently needed to obtain accurate and precise measurements of diet and nutrition. Enhancing knowledge about private phenotypes allows for more precise and predictive dietary guidance and intervention, and this has the potential to transform how people make informed diet and lifestyle choices. In today'south personalized marketplace, nosotros routinely use sophisticated technology to learn personalized step-by-footstep guidance that assures arrival at nearly any physical destination (eg, satellite navigation). In accordance with the natural diversity of humans as unique phenotypes, this concept could also exist applicable to the realm of food and nutrition. In other words, in that location is a demand for sensitive and specific devices to deliver step-by-step directions to any desired wellness destination. This requires the tools able to quantify health status and progress over important time scales and adjust trajectories according to biofeedback. Smartphones are the cornerstone of the customization and precision of modern life, incorporating precise personal information with global databases accessible through cloud storage and applying straightforward computational algorithms to guide decisions. This basic principle and its applications offer a sophisticated and diverse range of possibilities for enhancing our individual feel, whether through personalized navigation, physical activity tracking, tailoring fitness routines, and identifying a song or even a face. However, to date, the app market does not offer reliable solutions for automating the quantification of dietary intake that would significantly impact individualized quality of life decisions. Measurement and tracking devices provide practical utility for discerning phenotypic traits and defining progressive roadmaps to personalized wellness destinations. Automated food tracking devices could precisely inform diet and lifestyle choices appropriate to health status and guide individuals toward desired goals, including everything from diet planning to cardiometabolic functioning. Validation and effectiveness testing of candidate devices are essential steps to be taken for the employ of precision technologies to inform personalized diet and lifestyle guidance.
Conclusions
This study documented high variability in both the utility and accurateness of a wristband sensor to track nutritional intake (kcal/day). The researchers acknowledge that because dietary intake measurement of individuals has inherent challenges related to accuracy and variability, achieving precision of reference methods is a notable challenge. This written report highlights the need for innovative measurement tools that are precise, reliable, and efficacious to facilitate accurate personalized dietary measurement.
Acknowledgments
The authors would like to thank UC Davis DC for their collaboration, including Felipe Becerra, Marci Ofina, Leah Beck, RD, head chefs Roger Thompson, Cesar Cienfuegos, and staff at Cuarto and Segundo DC. The authors would also like to thank undergraduate research interns, including Melissa Vilas, Mengyang Lu, Xianyu Zhu, Melanie Hercules, Adaeze Ezeagwula, Jewel Esparza, Anna Liu, Fariba Osidary, Manvir Dhindsa, Alison Peng, Eileen Dihardja, Haley Adel, Brianna Bado, Macenzie Nielson, Taylor Janoe, Elena Chai, and Nealah Lee for their delivery to implementing research protocols, including dietary intake assessment and data entry. This research and publication were supported by unrestricted gifts to the Foods for Health Found from Healbe, LLC.
Abbreviations
| CGM | continuous glucose monitoring |
| DC | dining commons |
| FSL | FreeStyle Libre |
| LoA | limits of agreement |
| PI | principal investigator |
| RD | registered dietician |
| UC Davis | University of California, Davis |
| USDA | United States Department of Agriculture |
Footnotes
Conflicts of Interest: This inquiry and publication were supported by unrestricted gifts to the Foods for Health Plant from Healbe, LLC.
References
i. Betts JA, Gonzalez JT. Personalised diet: what makes yous and so special? Nutr Bull. 2016 Nov xv;41(4):353–nine. doi: 10.1111/nbu.12238. doi: 10.1111/nbu.12238. [CrossRef] [CrossRef] [Google Scholar]
3. Celis-Morales C, Lara J, Mathers JC. Personalising nutritional guidance for more effective behaviour change. Proc Nutr Soc. 2015 May;74(2):130–8. doi: 10.1017/S0029665114001633. [PubMed] [CrossRef] [Google Scholar]
iv. O'Sullivan A, Henrick B, Dixon B, Barile D, Zivkovic A, Smilowitz J, Lemay D, Martin Westward, German JB, Schaefer SE. 21st century toolkit for optimizing population wellness through precision diet. Crit Rev Food Sci Nutr. 2018;58(17):3004–xv. doi: ten.1080/10408398.2017.1348335. http://europepmc.org/abstract/MED/28678528. [PMC complimentary article] [PubMed] [CrossRef] [Google Scholar]
vii. Piccolo B, Keim N, Fiehn O, Adams SH, van Loan MD, Newman JW. Habitual physical activity and plasma metabolomic patterns distinguish individuals with depression vs high weight loss during controlled energy brake. J Nutr. 2015 April;145(4):681–xc. doi: 10.3945/jn.114.201574. http://europepmc.org/abstract/MED/25833772. [PMC complimentary article] [PubMed] [CrossRef] [Google Scholar]
8. Schork NJ. Personalized medicine: fourth dimension for one-person trials. Nature. 2015 April thirty;520(7549):609–eleven. doi: ten.1038/520609a. [PubMed] [CrossRef] [Google Scholar]
9. Woolf SH, Purnell JQ. The skilful life: working together to promote opportunity and meliorate population health and well-being. J Am Med Assoc. 2016 April 26;315(xvi):1706–8. doi: 10.1001/jama.2016.4263. [PubMed] [CrossRef] [Google Scholar]
11. Glasgow RE, Kwan BM, Matlock DD. Realizing the full potential of precision health: the demand to include patient-reported health beliefs, mental health, social determinants, and patient preferences data. J Clin Transl Sci. 2018 Jun;2(3):183–v. doi: 10.1017/cts.2018.31. http://europepmc.org/abstract/MED/30370072. [PMC costless article] [PubMed] [CrossRef] [Google Scholar]
thirteen. Schaefer S, Ching C, Breen H, German language J. Wearing, thinking, and moving: testing the feasibility of fitness tracking with urban youth. Amer J Health Educ. 2016 Jan 8;47(1):viii–xvi. doi: x.1080/19325037.2015.1111174. doi: 10.1080/19325037.2015.1111174. [CrossRef] [CrossRef] [Google Scholar]
14. Vigneshvar Due south, Sudhakumari CC, Senthilkumaran B, Prakash H. Recent advances in biosensor technology for potential applications-an overview. Front Bioeng Biotechnol. 2016;four:11. doi: 10.3389/fbioe.2016.00011. doi: 10.3389/fbioe.2016.00011. [PMC free article] [PubMed] [CrossRef] [CrossRef] [Google Scholar]
15. Xie J, Wen D, Liang 50, Jia Y, Gao 50, Lei J. Evaluating the validity of current mainstream wearable devices in fettle tracking under various concrete activities: comparative written report. JMIR Mhealth Uhealth. 2018 Apr 12;half dozen(4):e94. doi: 10.2196/mhealth.9754. https://mhealth.jmir.org/2018/iv/e94/ [PMC costless article] [PubMed] [CrossRef] [Google Scholar]
sixteen. Dorn D, Gorzelitz J, Gangnon R, Bell D, Koltyn K, Cadmus-Bertram L. Automatic identification of physical activeness type and duration by article of clothing activity trackers: a validation study. JMIR Mhealth Uhealth. 2019 May 23;7(5):e13547. doi: ten.2196/13547. https://mhealth.jmir.org/2019/5/e13547/ [PMC complimentary commodity] [PubMed] [CrossRef] [Google Scholar]
17. Nording ML, Yang J, Georgi K, Karbowski CH, German JB, Weiss RH, Hogg RJ, Trygg J, Hammock BD, Zivkovic AM. Individual variation in lipidomic profiles of salubrious subjects in response to omega-3 fat acids. PLoS One. 2013;8(10):e76575. doi: x.1371/journal.pone.0076575. http://dx.plos.org/x.1371/journal.pone.0076575. [PMC costless article] [PubMed] [CrossRef] [Google Scholar]
20. Zeevi D, Korem T, Zmora N, Israeli D, Rothschild D, Weinberger A, Ben-Yacov O, Lador D, Avnit-Sagi T, Lotan-Pompan M, Suez J, Mahdi JA, Matot Due east, Malka G, Kosower North, Rein M, Zilberman-Schapira G, Dohnalová L, Pevsner-Fischer Grand, Bikovsky R, Halpern Z, Elinav E, Segal Due east. Personalized nutrition past prediction of glycemic responses. Cell. 2015 Nov nineteen;163(v):1079–94. doi: 10.1016/j.cell.2015.11.001. https://linkinghub.elsevier.com/retrieve/pii/S0092-8674(fifteen)01481-6. [PubMed] [CrossRef] [Google Scholar]
21. Bakker M, van Erk MJ, Pellis L, Wopereis S, Rubingh C, Cnubben N, Kooistra T, van Ommen B, Hendriks H. An antiinflammatory dietary mix modulates inflammation and oxidative and metabolic stress in overweight men: a nutrigenomics approach. Am J Clin Nutr. 2010 Apr;91(4):1044–59. doi: 10.3945/ajcn.2009.28822. [PubMed] [CrossRef] [Google Scholar]
22. Mozaffarian D, Ludwig DS. Dietary guidelines in the 21st century-a time for food. J Am Med Assoc. 2010 Aug 11;304(6):681–2. doi: 10.1001/jama.2010.1116. [PubMed] [CrossRef] [Google Scholar]
23. Archer East, Marlow M, Lavie C. Controversy and debate: memory-based methods paper 1: the fatal flaws of food frequency questionnaires and other retentivity-based dietary assessment methods. J Clin Epidemiol. 2018 Dec;104:113–24. doi: ten.1016/j.jclinepi.2018.08.003. [PubMed] [CrossRef] [Google Scholar]
24. Martin CK, Correa JB, Han H, Allen 60 minutes, Rood JC, Champagne CM, Gunturk BK, Bray GA. Validity of the remote nutrient photography method (RFPM) for estimating energy and nutrient intake in near real-time. Obesity (Silver Spring) 2012 Apr;xx(4):891–9. doi: 10.1038/oby.2011.344. doi: 10.1038/oby.2011.344. [PMC free article] [PubMed] [CrossRef] [CrossRef] [Google Scholar]
25. Wrangham R. The curiously long absenteeism of cooking in evolutionary thought. Learn Behav. 2016 Jun;44(2):116–vii. doi: 10.3758/s13420-016-0223-four. [PubMed] [CrossRef] [Google Scholar]
27. Dungan Chiliad, Verm Northward. Monitoring Technologies – Continuous Glucose Monitoring, Mobile Technology, Biomarkers of Glycemic Command. Due south Dartmouth, MA: MDText; 2000. [Google Scholar]
28. Lohan T, Roche AF, Martorell R. Anthropometric Standardization Reference Manual. Champaign, Illinois: Human being Kinetics; 1988. [Google Scholar]
29. Ólafsdóttir AF, Attvall S, Sandgren U, Dahlqvist S, Pivodic A, Skrtic S, Theodorsson East, Lind M. A clinical trial of the accurateness and handling experience of the flash glucose monitor freestyle libre in adults with type 1 diabetes. Diabetes Technol Ther. 2017 Mar;19(3):164–72. doi: 10.1089/dia.2016.0392. http://europepmc.org/abstruse/MED/28263665. [PMC free article] [PubMed] [CrossRef] [Google Scholar]
Articles from JMIR mHealth and uHealth are provided here courtesy of JMIR Publications Inc.
How To Synch Healbe Gobe2 With Phone,
Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7407252/
Posted by: vanhorneworidence.blogspot.com
0 Response to "How To Synch Healbe Gobe2 With Phone"
Post a Comment