Purpose: Measuring infants’ (0-2 years) physical activity is a growing area of research globally. Accelerometers have been widely used to measure older children’s and adults’ physical activity. An increasing number of studies have used accelerometers as a way to measure infant physical activity, which has resulted in the application of a variety of methods. To date, no existing study has focused specifically on synthesizing the various accelerometer methodologies in detail (e.g., type of accelerometer, wear location and time, data processing, etc.) used to measure daytime movement outcomes of interest among infants. Therefore, the purpose of this scoping review is to synthesize the published literature on accelerometer methodology to measure daytime physical activity among infants.

Methods: A systematic search of five online databases using carefully selected key terms was conducted to compile relevant literature. The results of the online database searches were screened for inclusion in the scoping review. In total, 105 articles met the inclusion criteria of using accelerometers to measure infants’ physical activity. The methodologies used in included studies were categorized by age groups, <1 month, 1-6 months, >6-12 months, >12-18 months, >18-24 months, and longitudinal (i.e., multiple measurements taken across multiple age groups).

Results/findings: Accelerometry methodologies (e.g., wear location and time, number of devices, device initialization) and study design qualities (e.g., outcome of interest and location of data collection) varied widely between and within the various age groups. Accelerometer brand or type of device demonstrated greatest variation across included studies. However, ActiGraph devices to measure physical activity within free-living environments were the most common.

Conclusions: This review provides evidence of the need for researchers to ensure the methodology used is reported in detail in order to help develop methodology that can accurately assess infant daytime movement.

Purpose: Mobile devices (e.g., tablets, smartphones) have been rapidly integrated into the lives of children and have impacted how children engage with digital media. The portability of these devices allows for sporadic, on-demand interaction, thus reducing the already tenuous accuracy of self-report estimates of mobile device use. Passive sensing applications objectively monitor time spent on a given device for greater accuracy but are unable to identify who is using the device. Behavioral biometric authentication, using embedded mobile device sensors to authenticate users, could address this significant limitation. This proof-of-concept study examined the preliminary accuracy of machine learning models trained on iPad sensor data to identify the unique child user of an iPad device.
Methods: Data were collected on an opportunistic sample of nine participants (8.2 ± 1.8 years, 5 female) as part of two larger studies during the sedentary portion of their semi-structured physical activity protocols. SensorLog app was downloaded onto study iPads and collected data from the accelerometer, gyroscope, and magnetometer sensors while the participant interacted with the iPad. Five machine learning models, Logistic Regression (LR), Support Vector Machine, Neural Net (NN), k-Nearest Neighbors (k-NN), and Random Forest (RF) were trained using 57 features extracted from the sensor output to perform multi-class classification. A train-test split of 80%-20% was used for model fitting. Model performance was evaluated using F1 score, accuracy, precision, and recall.
Results/findings: Model performance was highly satisfactory, with F1 scores ranging from 0.75 to 0.94 and precision ranging from 0.77 to 0.94. RF and k-NN had the highest performance across metrics, with F1 scores of 0.94 for both models. The highest performing features of the RF model were maximum yaw, maximum roll, and mean yaw, which are derived from rotation about the Y-axis (yaw) and the Z-axis (roll).
Conclusions: This study highlights the potential of using embedded mobile device sensors to continuously identify the user of a device in the context of children’s screen time measurement. Future research should explore the performance of this technology in larger samples of children across both standardized study conditions, as well as in free-living environments.

Purpose: Development of culturally-appropriate population-level interventions are critical to improve health equity. Yet, such research is often stymied by the lack of measurement tools that are specific, appropriate, and valid for diverse populations. Few valid home food environment assessment tools exist and none have been validated with large, immigrant or low-literacy populations. In 2008, our team developed and validated a Home Food Inventory (HFI) to assess the healthfulness and obesity risk of home food environments using a checklist format. The HFI showed good reliability and construct validity and it has been used extensively in the field. Yet, the original HFI was developed for English-speaking populations, is quite lengthy and is paper-based. We will present our process for developing an accessible home food environment assessment toolkit for use in English- and Spanish-speaking populations and describe our pilot findings. Methods: The HFI-core was developed via secondary analysis of HFI and other data from four large studies. English and Spanish HFI-core versions were created and contextualized with advice from a local and multicultural community advisory board. Then we conducted “think aloud” interviews for cognitive testing (n=11) and pilot testing (n=40) with participants to assess usability, satisfaction, and criterion validity. The toolkit will include a revised paper version (HFI-core) and multi-media electronic tools (eHFI) targeting foods known to impact diet-related health. Results: The number of food items in the HFI-core was reduced by 24% from the original HFI (from 175 to 134 items). “Think aloud” interviews informed use of language and instructions. Development of the Spanish HFI-core required careful consideration of local name variations of foods. Pilot findings showed high satisfaction (93% of participants were satisfied/very satisfied with the HFI-core). Average completion time was 17 minutes. Criterion validity varied across food groups and slightly by language. The HFI-core was revised and retested until acceptable criterion validity levels were reached (correlations>.60) in both languages. Conclusions: The HFI-core demonstrated high participant satisfaction and reduced response burden. This preliminary work sets the stage to advance diet-related health research at a time when our world is becoming more globally-integrated, yet significant health inequities exist.

Background: A combination of heart rate (HR) and accelerometry provides better estimates of energy expenditure (EE) than either metric alone. However, existing devices are not designed for children and may be distracting or uncomfortable. Furthermore, data from existing devices is often closed source, meaning that it cannot be truly validated for children. Therefore, we developed the PATCH (Platform for Accurate Tracking of Children’s Health), an open-source device designed to measure both HR and acceleration. The current study aimed to examine initial validity of the PATCH to measure children’s EE.
Methods: Participants for the current study were 18 children aged 3-8 (mean age 8.6yrs, 50% Female, 67% White). Children participated in a semi-structured protocol ranging in intensity from sedentary (e.g., using an iPad) to vigorous (e.g., running). The PATCH was affixed to the chest using a spun lace adhesive. PATCH HR was measured using photoplethysmography while acceleration was measured using a 3-axis accelerometer (±16 G). Indirect Calorimetry was used to measure oxygen uptake rate (VO2/min) normalized for body weight (kg). All data were aggregated at the minute level.
We used cross-sectional time series (CSTS) models to predict minute-by-minute EE from a combination of minute level HR and acceleration, participant characteristics (biological sex, age, weight, height) and significant interactions. Models included a random intercept to account for the nested nature of the data. Model fit was assessed using mean bias, mean absolute error (MAE), root mean squared error (RMSE), mean absolute percent error (MA%E) and variance explained (R2). We used Lin’s concordance correlation coefficient (CCC) to assess the absolute agreement between PATCH indirect calorimetry.
Results: Compared to indirect calorimetry, PATCH MAE for VO2 uptake rate was 1.54mL/min/kg (95% CI 0.1 to 5.0), RMSE was 2.61mL/min/kg, and MA%E was 11.1% (95% CI 0.8 to 31.6%). 95% limits of agreement for mean bias were -2.8mL/min/kg and 3.5mL/min/kg. The CSTS model predicted 86.2% of the variance in EE. Lin’s CCC was >0.9 indicating near perfect agreement.
Conclusions: Upon preliminary assessment, the PATCH device is a valid measure of children’s EE. Additional validation in free living conditions is needed to generalize beyond a laboratory setting

Purpose: Combining heart rate and accelerometry data offers more accurate energy expenditure estimates than using either measurement individually. Yet, existing dual-measurement devices aren't child-friendly, leading to low compliance due to their distracting and uncomfortable design. Additionally, the use of proprietary processing methods in these wearable devices prevents true validation. Consequently, our team has developed an open-source PATCH (Platform for Accurate Tracking of Children's Health) device tailored to track heart rate and movement in children aged 3-8. Our study evaluated various PATCH casing materials and body positions to determine the feasibility of deploying the PATCH in free-living studies with children.

Methods: We recruited a convenience sample of 60 children ages 3-8 years (mean age = 5.98 years, ±1.70, 47% female; 72% White). We tested six material and placement combinations, termed 'waves,' including polycarbonate, thermoplastic polyurethane (TPU), and medical-grade silicone elastomer on either the chest or lower back. Over two weeks, parents completed daily surveys to document the children’s use of the PATCH and to rate any issues such as pain, redness, itching, rashes, or tenderness on a scale from 0 (none) to 4 (severe). ANOVAs were used to examine differences by wave.

Results: Wave 3 (polycarbonate, lower back) had the highest wear (Mean = 12.40 days, SD = 1.78) and wave 5 (TPU, chest) had the lowest (Mean = 5.14 days, SD = 6.59). Waves 1 (polycarbonate, chest), 4 (polycarbonate, back) and 6 (silicone, chest) all had ≥10days of wear on average (>70% compliance). Wave 2 (silicone, chest) had an average of 9 days of wear. Participants in wave 5 wore the PATCH for significantly fewer days compared to other waves (f(5,53) = 3.51, p<.01) and reported significantly higher rates of rashes (f(5,54) = 4.27, p<.01). There were no other differences by wave for days of wear, itch, redness, pain, or tenderness (p>.05).

Conclusion: Placing an adhesive PATCH device on either the chest or lower back is a practical option for children. Considering the materials, TPU seems unsuitable, whereas polycarbonate or silicone casings could be viable choices.

Purpose: Using BMI z-scores (zBMI) to illustrate changes in adiposity in children with severe obesity may be less accurate due to the compression of z-scores corresponding to extremely high BMI values into a very narrow range (Freedman et al., 2017). Therefore, reporting the BMI percent of the 95th percentile (%BMIp95) has been recommended when evaluating BMI in children with moderate or severe obesity. The purpose of this research was to evaluate the various BMI metrics for changes in outcomes for a community-based Family Healthy Weight Program (FHWP).
Methods: 90 children between 6-12 years of age qualified to participate in a FHWP based on a BMI ≥95th percentile for age and gender. zBMI and %BMIp95 were calculated at baseline, after a 12-week intervention, and 6-month follow-up. Dual energy X-ray absorptiometry (DXA) was used to measure changes in body composition including fat mass (FM).
Results: 75 children (9.46 ± 1.74 years old) completed 12-week assessments and 47 completed 6-month assessments. 40% of children had severe obesity (%BMIp95: 140.5 ± 17.6). Change in FM at 12-weeks and 6-months had a strong correlation with changes in both zBMI and %BMIp95 for children with moderate obesity (Δ12wk: r=0.75 (p<0.05), r= 0.81(p<0.05); 6mo: r=0.82 (p<0.05), r=0.86 (p<0.05)). In addition, change in FM had moderate correlation with changes in both zBMI and %BMIp95 for children with severe obesity at 12-weeks (Δ12wk: r=0.52 (p<0.05), r= 0.61 (p<0.05)) and a strong correlation at 6-months (Δ6mo: r=0.71(p<0.05), r=0.87 (p<0.05)).
Conclusion: Although baseline correlations were weak between FM and zBMI (r=0.19, p>0.05) and strong between FM and %BMIp95 (r=0.73, p<0.05), strong correlations were also found with improved body composition changes for both change in zBMI and %BMIp95. Due to the nature of the community-based program, severe obesity among the highest range (%BMIp95>160) may be underrepresented in these data. Our results indicate that the use of zBMI or %BMIp95 may be options when evaluating change for child participants in a FHWP.