Utility and applicability of the "Childhood Obesity Risk Evaluation" (CORE)-index in predicting obesity in childhood and adolescence in Greece from early life: the "National Action Plan for Public Health"

Obesity Prediction with EHR Data: A deep learning approach with interpretable elements

Childhood obesity is a major public health challenge. Early prediction and identification of the children at an elevated risk of developing childhood obesity may help in engaging earlier and more effective interventions to prevent and manage obesity. Most existing predictive tools for childhood obesity primarily rely on traditional regression-type methods using only a few hand-picked features and without exploiting longitudinal patterns of children's data. Deep learning methods allow the use of high-dimensional longitudinal datasets. In this paper, we present a deep learning model designed for predicting future obesity patterns from generally available items on children's medical history. To do this, we use a large unaugmented electronic health records dataset from a large pediatric health system in the US. We adopt a general LSTM network architecture and train our proposed model using both static and dynamic EHR data. To add interpretability, we have additionally included an attention layer to calculate the attention scores for the timestamps and rank features of each timestamp. Our model is used to predict obesity for ages between 3-20 years using the data from 1-3 years in advance. We compare the performance of our LSTM model with a series of existing studies in the literature and show it outperforms their performance in most age ranges.

Childhood overweight and obesity at the start of primary school: External validation of pregnancy and early-life prediction models

Tackling the childhood obesity epidemic can potentially be facilitated by risk-stratifying families at an early-stage to receive prevention interventions and extra support. Using data from the Born in Bradford (BiB) cohort, this analysis aimed to externally validate prediction models for childhood overweight and obesity developed as part of the Studying Lifecourse Obesity PrEdictors (SLOPE) study in Hampshire. BiB is a longitudinal multi-ethnic birth cohort study which recruited women at around 28 weeks gestation between 2007 and 2010 in Bradford. The outcome was body mass index (BMI) ≥91st centile for overweight/obesity at 4-5 years. Discrimination was assessed using the area under the receiver operating curve (AUC). Calibration was assessed for each tenth of predicted risk by calculating the ratio of predicted to observed risk and plotting observed proportions versus predicted probabilities. Data were available for 8003 children. The AUC on external validation was comparable to that on development at all stages (early pregnancy, birth, ~1 year and ~2 years). The AUC on external validation ranged between 0.64 (95% confidence interval (CI) 0.62 to 0.66) at early pregnancy and 0.82 (95% CI 0.81 to 0.84) at ~2 years compared to 0.66 (95% CI 0.65 to 0.67) and 0.83 (95% CI 0.82 to 0.84) on model development in SLOPE. Calibration was better in the later model stages (early life ~1 year and ~2 years). The SLOPE models developed for predicting childhood overweight and obesity risk performed well on external validation in a UK birth cohort with a different geographical location and ethnic composition.

European Childhood Obesity Risk Evaluation (CORE) index based on perinatal factors and maternal sociodemographic characteristics: the Feel4Diabetes-study

The aim of this study was to develop and examine the predictive accuracy of an index that estimates obesity risk in childhood based on perinatal factors and maternal sociodemographic characteristics. Analysis was conducted by using cross-sectional and retrospective data collected from a European cohort of 2775 schoolchildren and their families participating in the Feel4Diabetes-study. The cohort was randomly divided by using two-thirds of the sample for the development of the index and the remaining one third for assessing its predictive accuracy. Logistic regression analyses determined a prediction model for childhood obesity. The area under the receiver operating characteristic curve (AUC-ROC), sensitivity, specificity, and positive and negative predictive values (PPV, NPV) were calculated. Cut-off analysis was applied to identify the optimal value of the index score that predicts obesity with the highest possible sensitivity and specificity. Eight factors were found to be significantly associated with obesity and were included as components in the European "Childhood Obesity Risk Evaluation" (CORE) index: region of residence, maternal education, maternal pre-pregnancy weight status, gestational weight gain, maternal smoking during pregnancy, birth weight for gestational age, infant growth velocity, and exclusive breastfeeding during the first 6 months. Risk score ranged from 0 to 22 corresponding to a risk from 0.9 to 54.6%. The AUC-ROC was 0.725 with optimal cut-off ≥9 (sensitivity = 74.1%, specificity = 61.0%, PPV = 11.3%, NPV = 97.2%).Conclusion: The European CORE index can be used as a screening tool for the identification of infants at high-risk for becoming obese at 6-9 years. This tool could assist healthcare professionals in initiating preventive measures from the early life.Trial registration: The Feel4Diabetes-intervention is registered at https://clinicaltrials.gov/ ; number, CT02393872; date, March 20, 2015. What is Known: • As prevention of obesity should start early in life, there is a compelling rationale for the early identification of high-risk children to facilitate targeted intervention. What is New: • This study developed and assessed the predictive accuracy of an index for the Childhood Obesity Risk Evaluation (CORE), combining certain perinatal factors and maternal sociodemographic characteristics in a large European cohort. • The European CORE index can be used as a screening tool for identifying infants at high-risk for becoming obese at 6-9 years and assist health professionals in initiating early prevention strategies.

Machine Learning Models to Predict Childhood and Adolescent Obesity: A Review

The prevalence of childhood and adolescence overweight an obesity is raising at an alarming rate in many countries. This poses a serious threat to the current and near-future health systems, given the association of these conditions with different comorbidities (cardiovascular diseases, type II diabetes, and metabolic syndrome) and even death. In order to design appropriate strategies for its prevention, as well as understand its origins, the development of predictive models for childhood/adolescent overweight/obesity and related outcomes is of extreme value. Obesity has a complex etiology, and in the case of childhood and adolescence obesity, this etiology includes also specific factors like (pre)-gestational ones; weaning; and the huge anthropometric, metabolic, and hormonal changes that during this period the body suffers. In this way, Machine Learning models are becoming extremely useful tools in this area, given their excellent predictive power; ability to model complex, nonlinear relationships between variables; and capacity to deal with high-dimensional data typical in this area. This is especially important given the recent appearance of large repositories of Electronic Health Records (EHR) that allow the development of models using datasets with many instances and predictor variables, from which Deep Learning variants can generate extremely accurate predictions. In the current work, the area of Machine Learning models to predict childhood and adolescent obesity and related outcomes is comprehensively and critically reviewed, including the latest ones using Deep Learning with EHR. These models are compared with the traditional statistical ones that used mainly logistic regression. The main features and applications appearing from these models are described, and the future opportunities are discussed.

Novel e-Health Applications for the Management of Cardiometabolic Risk Factors in Children and Adolescents in Greece

Obesity in childhood and adolescence represents a major health problem. Novel e-Health technologies have been developed in order to provide a comprehensive and personalized plan of action for the prevention and management of overweight and obesity in childhood and adolescence. We used information and communication technologies to develop a “National Registry for the Prevention and Management of Overweight and Obesity” in order to register online children and adolescents nationwide, and to guide pediatricians and general practitioners regarding the management of overweight or obese subjects. Furthermore, intelligent multi-level information systems and specialized artificial intelligence algorithms are being developed with a view to offering precision and personalized medical management to obese or overweight subjects. Moreover, the Big Data against Childhood Obesity platform records behavioral data objectively by using inertial sensors and Global Positioning System (GPS) and combines them with data of the environment, in order to assess the full contextual framework that is associated with increased body mass index (BMI). Finally, a computerized decision-support tool was developed to assist pediatric health care professionals in delivering personalized nutrition and lifestyle optimization advice to overweight or obese children and their families. These e-Health applications are expected to play an important role in the management of overweight and obesity in childhood and adolescence.

Assessment of the Effectiveness of a Computerised Decision-Support Tool for Health Professionals for the Prevention and Treatment of Childhood Obesity. Results from a Randomised Controlled Trial

We examined the effectiveness of a computerised decision-support tool (DST), designed for paediatric healthcare professionals, as a means to tackle childhood obesity. A randomised controlled trial was conducted with 65 families of 6–12-year old overweight or obese children. Paediatricians, paediatric endocrinologists and a dietitian in two children’s hospitals implemented the intervention. The intervention group (IG) received personalised meal plans and lifestyle optimisation recommendations via the DST, while families in the control group (CG) received general recommendations. After three months of intervention, the IG had a significant change in dietary fibre and sucrose intake by 4.1 and −4.6 g/day, respectively. In addition, the IG significantly reduced consumption of sweets (i.e., chocolates and cakes) and salty snacks (i.e., potato chips) by −0.1 and −0.3 portions/day, respectively. Furthermore, the CG had a significant increase of body weight and waist circumference by 1.4 kg and 2.1 cm, respectively, while Body Mass Index (BMI) decreased only in the IG by −0.4 kg/m². However, the aforementioned findings did not differ significantly between study groups. In conclusion, these findings indicate the dynamics of the DST in supporting paediatric healthcare professionals to improve the effectiveness of care in modifying obesity-related behaviours. Further research is needed to confirm these findings.

Predicting childhood overweight and obesity using maternal and early life risk factors: a systematic review

BACKGROUND

Childhood obesity is a serious public health challenge, and identification of high-risk populations with early intervention to prevent its development is a priority. We aimed to systematically review prediction models for childhood overweight/obesity and critically assess the methodology of their development, validation and reporting.

METHODS

Medline and Embase were searched systematically for studies describing the development and/or validation of a prediction model/score for overweight and obesity between 1 to 13 years of age. Data were extracted using the Cochrane CHARMS checklist for Prognosis Methods.

RESULTS

Ten studies were identified that developed (one), developed and validated (seven) or externally validated an existing (two) prediction model. Six out of eight models were developed using automated variable selection methods. Two studies used multiple imputation to handle missing data. From all studies, 30,475 participants were included. Of 25 predictors, only seven were included in more than one model with maternal body mass index, birthweight and gender the most common.

CONCLUSION

Several prediction models exist, but most have not been externally validated or compared with existing models to improve predictive performance. Methodological limitations in model development and validation combined with non-standard reporting restrict the implementation of existing models for the prevention of childhood obesity.

Mid-Upper-Arm Circumference and Arm-to-Height Ratio to Identify Obesity in School-Age Children

BACKGROUND

In resource-poor settings, anthropometric parameters are evaluated as potential alternatives to the body mass index (BMI) for detecting overweight and obesity in children. To this end, the mid-upper-arm circumference (MUAC) and the mid-upper-arm circumference-to-height ratio (AHtR) were evaluated as proxies to BMI in Thai school-age children.

STUDY DESIGN

An observational, cross-sectional study was performed on school-aged children.

PARTICIPANTS

Children in grades 1 through 6 at all public elementary schools in the Ongkharak district, Nakhon Nayok, Thailand during May and June 2013 were included. This is a rural district with low per capita income.

METHODS

Weight, height, and MUAC were measured in school-age children and analyzed to identify optimal cut-off values for MUAC and AHtR for detection of overweight and obesity in comparison to BMI. Receiver operating characteristic (ROC) curve analysis determined the validity of MUAC and AHtR use.

RESULTS

Data from 3,618 children, aged 6.0-12.99 years, were analyzed. MUAC correlated with age and height (P < 0.001), but especially with body weight (r = 0.888 to 0.914) and BMI (r = 0.859 to 0.908) in both genders, while AHtR correlated with body weight and BMI (P < 0.001), but not with age. Cut-off values of MUAC for obesity diagnosis ranged from 18.9 to 25.5 cm for boys and from 19.8 to 25.4 cm for girls. Accuracy was excellent for both boys (AUC = 0.952-0.991) and girls (AUC = 0.917-0.990). Cut-off of MUAC for overweight diagnosis ranged from 17.2 to 22.4 cm for boys (AUC = 0.883-0.965) and from 18.0 to 23.2 cm for girls (AUC = 0.905-0.931). AHtR cut-off values for obesity and overweight diagnosis at 0.16 and 0.145, respectively, were determined with excellent diagnostic accuracy (AUC ranged from 0.920 to 0.975).

CONCLUSION

MUAC and AHtR were reliable tools to detect overweight and obesity in Thai school-age children. Cut-off points for MUAC were age and gender specific, while AHtR at 0.16 and 0.145 were the optimal values for both genders, independent of age. These anthropometric measurements showed excellent accuracy in predicting overweight and obesity with high specificity and sensitivity.