Abstract
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 article, 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 United States. 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 and 20 years using the data from 1 to 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.
Supplemental Material
Available for Download
Supplementary material
- [1] . 2015. Prediction of fat-free mass in children. Clinical Pharmacokinetics 54, 11 (2015), 1169–1178.Google Scholar
- [2] . 2016. Obesity and obstructive sleep apnea in children. Archives de Pediatrie: Organe Officiel de la Societe Francaise de Pediatrie 24 (2016), S34–S38.Google Scholar
- [3] . 2014. Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473 (2014).Google Scholar
- [4] . 2009. Clinical prediction models-a practical approach to development, validation and updating. Development 18, 500 (2009), 53–99.Google Scholar
- [5] . 2017. Deep Learning, vol. 1. MIT press Massachusetts, USA.Google Scholar
- [6] . 2018. Interventions to prevent global childhood overweight and obesity: A systematic review. Lancet Diabetes & Endocrinology 6, 4 (2018), 332–346.Google Scholar
- [7] . 2016. RETAIN: An interpretable predictive model for healthcare using reverse time attention mechanism. In Proceedings of the 30th International Conference on Neural Information Processing Systems (NIPS’16). Curran Associates Inc., Red Hook, NY, 3512–3520.Google Scholar
- [8] . 2016. Doctor AI: Predicting clinical events via recurrent neural networks. In Machine Learning for Healthcare Conference. PMLR, 301–318.Google Scholar
- [9] . 2016. Multi-layer representation learning for medical concepts. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 1495–1504.Google ScholarDigital Library
- [10] . 2016. Medical concept representation learning from electronic health records and its application on heart failure prediction. arXiv preprint arXiv:1602.03686 (2016).Google Scholar
- [11] . 2017. Using recurrent neural network models for early detection of heart failure onset. Journal of the American Medical Informatics Association 24, 2 (2017), 361–370.Google ScholarCross Ref
- [12] . 2020. Machine learning models to predict childhood and adolescent obesity: A review. Nutrients 12, 8 (2020), 2466.Google ScholarCross Ref
- [13] . 1998. Health consequences of obesity in youth: Childhood predictors of adult disease. Pediatrics 101, Supplement 2 (1998), 518–525.Google Scholar
- [14] . 2012. Prediction of childhood obesity by infancy weight gain: An individual-level meta-analysis. Paediatric and Perinatal Epidemiology 26, 1 (2012), 19–26.Google Scholar
- [15] . 2006. Parents’ perceptions of their child’s weight and health. Pediatrics 117, 3 (2006), 681–690.Google Scholar
- [16] . 2018. An efficient multi-layer ensemble framework with BPSOGSA-based feature selection for credit scoring data analysis. Arabian Journal for Science and Engineering 43, 12 (2018), 6909–6928.Google Scholar
- [17] . 2016. Understanding the causes of obesity in children with trisomy 21: Hyperphagia vs physical inactivity. Journal of Intellectual Disability Research 60, 9 (2016), 856–864.Google Scholar
- [18] . 2021. Childhood obesity facts. https://www.cdc.gov/obesity/data/childhood.html.Google Scholar
- [19] . 2015. Prediction of adolescent and adult adiposity outcomes from early life anthropometrics. Obesity 23, 1 (2015), 162–169.Google Scholar
- [20] . 2020. Time-series Imputation and Prediction with Bi-Directional Generative Adversarial Networks. arxiv:2009.08900 [cs.LG]Google Scholar
- [21] . 2021. Concurrent imputation and prediction on EHR data using bi-directional GANs: Bi-GANs for EHR imputation and prediction. In Proceedings of the 12th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics (BCB’21). Association for Computing Machinery, New York, NY, Article
7 , 9 pages. Google ScholarDigital Library - [22] . 2019. Obesity prediction with EHR data: A deep learning approach with interpretable elements. arXiv preprint arXiv:1912.02655 (2019).Google Scholar
- [23] . 2019. Predicting childhood obesity using electronic health records and publicly available data. PloS One 14, 4 (2019), e0215571.Google Scholar
- [24] . 1988. Obesity in achondroplasia. American Journal of Medical Genetics 31, 3 (1988), 597–602.Google Scholar
- [25] . 1997. Long short-term memory. Neural Computation 9, 8 (1997), 1735–1780.Google ScholarDigital Library
- [26] . 2019. Development and validation of a prediction model for fat mass in children and adolescents: Meta-analysis using individual participant data. BMJ 366 (2019).Google Scholar
- [27] . 2020. Carepre: An intelligent clinical decision assistance system. ACM Transactions on Computing for Healthcare 1, 1 Article 6 (mar 2020), 20 pages. Google ScholarDigital Library
- [28] . 2013. Severe obesity in children and adolescents: Identification, associated health risks, and treatment approaches: A scientific statement from the American Heart Association. Circulation 128, 15 (2013), 1689–1712.Google Scholar
- [29] . 2016. Interacting with predictions: Visual inspection of black-box machine learning models. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems. 5686–5697.Google Scholar
- [30] . 2020. Credit score classification using spiking extreme learning machine. Computational Intelligence 36, 2 (2020), 402–426.Google ScholarCross Ref
- [31] . 2013. Obesity is associated with sensorineural hearing loss in adolescents. The Laryngoscope 123, 12 (2013), 3178–3184.Google Scholar
- [32] . 2012. Identifying infants at risk of becoming obese: Can we and should we? Public Health 126, 2 (2012), 123–128.Google Scholar
- [33] . 2014. Deep learning for healthcare decision making with EMRs. In 2014 IEEE International Conference on Bioinformatics and Biomedicine (BIBM’14). IEEE, 556–559.Google Scholar
- [34] . 2016. Developing an algorithm to detect early childhood obesity in two tertiary pediatric medical centers. Applied Clinical Informatics 7, 3 (2016), 693–706.Google Scholar
- [35] . 2016. Directly modeling missing data in sequences with RNNs: Improved classification of clinical time series. In Machine Learning for Healthcare Conference. PMLR, 253–270.Google Scholar
- [36] . 2019. Statistical Analysis with Missing Data. Vol. 793. John Wiley & Sons.Google Scholar
- [37] . 2013. Childhood obesity risk evaluation based on perinatal factors and family sociodemographic characteristics: CORE index. European Journal of Pediatrics 172, 4 (2013), 551–555.Google Scholar
- [38] . 2016. 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.” European Journal of Pediatrics 175, 12 (2016), 1989–1996.Google Scholar
- [39] . 2019. LSTM model for prediction of heart failure in big data. Journal of Medical Systems 43, 5 (2019), 1–13.Google ScholarDigital Library
- [40] . 2019. Obesity hypoventilation syndrome. European Respiratory Review 28, 151 (2019).Google Scholar
- [41] . 2017. Deep diabetologist: Learning to prescribe hypoglycemic medications with recurrent neural networks. Studies in Health Technology and Informatics 245 (2017), 1277–1277.Google Scholar
- [42] . 2010. Obesity hypoventilation syndrome: A state-of-the-art review. Respiratory Care 55, 10 (2010), 1347–1365.Google Scholar
- [43] . 2005. Rapid growth in infancy and childhood and obesity in later life–A systematic review. Obesity Reviews 6, 2 (2005), 143–154.Google Scholar
- [44] . 2012. Estimation of newborn risk for child or adolescent obesity: Lessons from longitudinal birth cohorts. PloS One 7, 11 (2012), e49919.Google Scholar
- [45] . 2001. Data Table of BMI-for-age Charts. https://www.cdc.gov/growthcharts/html_charts/bmiagerev.htm.Google Scholar
- [46] . 2006. Rapid infancy weight gain and subsequent obesity: Systematic reviews and hopeful suggestions. Acta Paediatrica 95, 8 (2006), 904–908.Google Scholar
- [47] . 2020. Accelerated weight gain, prematurity, and the risk of childhood obesity: A meta-analysis and systematic review. PloS One 15, 5 (2020), e0232238.Google Scholar
- [48] . 2004. Identification, evaluation, and management of obesity in an academic primary care center. Pediatrics 114, 2 (2004), e154–e159.Google Scholar
- [49] . 2013. Early life risk factors of being overweight at 10 years of age: Results of the German birth cohorts GINIplus and LISAplus. European Journal of Clinical Nutrition 67, 8 (2013), 855–862.Google Scholar
- [50] . 2003. Comparing computer-interpretable guideline models: A case-study approach. Journal of the American Medical Informatics Association 10, 1 (2003), 52–68.Google ScholarCross Ref
- [51] . 2004. Body mass index charts: Useful yet underused. Journal of Pediatrics 144, 4 (2004), 455–460.Google Scholar
- [52] . 2016. Deepcare: A deep dynamic memory model for predictive medicine. In Pacific-Asia Conference on Knowledge Discovery and Data Mining. Springer, 30–41.Google ScholarDigital Library
- [53] . 2019. Multi-modal predictive models of diabetes progression. In Proceedings of the 10th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics. 253–258.Google ScholarDigital Library
- [54] . 2021. Predicting progression patterns of type 2 diabetes using multi-sensor measurements. Smart Health 21 (2021), 100206. Google ScholarCross Ref
- [55] . 2022. Predicting overweight and obesity in later life from childhood data: A review of predictive modeling approaches. Computational Sciences and Artificial Intelligence in Industry (2022), 203–220.Google Scholar
- [56] . 2016. Validation, optimal threshold determination, and clinical utility of the infant risk of overweight checklist for early prevention of child overweight. Childhood Obesity 12, 3 (2016), 202–209.Google Scholar
- [57] . 2006. Perceptions of children’s body sizes among mothers living on the Texas-Mexico border (La Frontera). Public Health Nursing 23, 6 (2006), 488–495.Google Scholar
- [58] . 2016. A risk score for childhood obesity in an urban latino cohort. Journal of Pediatrics 172 (2016), 29–34.Google Scholar
- [59] . 2015. Childhood obesity: Causes and consequences. Journal of Family Medicine and Primary Care 4, 2 (2015), 187.Google ScholarCross Ref
- [60] . 2019. Obesity in achondroplasia patients: from evidence to medical monitoring. Orphanet Journal of Rare Diseases 14, 1 (2019), 1–9.Google Scholar
- [61] . 2013. Developing prediction equations and a mobile phone application to identify infants at risk of obesity. PLoS One 8, 8 (2013), e71183.Google Scholar
- [62] . 2017. Deep EHR: A survey of recent advances in deep learning techniques for electronic health record (EHR) analysis. IEEE Journal of Biomedical and Health Informatics 22, 5 (2017), 1589–1604.Google ScholarCross Ref
- [63] . 2018. Machine learning models in electronic health records can outperform conventional survival models for predicting patient mortality in coronary artery disease. PloS One 13, 8 (2018), e0202344.Google Scholar
- [64] . 2002. Infant weight gain and childhood overweight status in a multicenter, cohort study. Pediatrics 109, 2 (2002), 194–199.Google Scholar
- [65] . 2011. Predicting the risk of newborn children to become overweight later in childhood: The PIAMA birth cohort study. International Journal of Pediatric Obesity 6, 2 Part 2 (2011), e170–e178.Google Scholar
- [66] . 2021. PEDSnet: A Pedistric Learning Health System. https://pedsnet.org.Google Scholar
- [67] . 2021. PEDSnet Common Data Model. https://pedsnet.org/data/common-data-model.Google Scholar
- [68] . 2018. Hybrid credit scoring model using neighborhood rough set and multi-layer ensemble classification. Journal of Intelligent & Fuzzy Systems 34, 3 (2018), 1543–1549.Google ScholarDigital Library
- [69] . 2019. A novel hybrid credit scoring model based on ensemble feature selection and multilayer ensemble classification. Computational Intelligence 35, 2 (2019), 371–394.Google Scholar
- [70] . 2019. Obesity and dyslipidemia. Metabolism 92 (2019), 71–81.Google Scholar
- [71] . 2014. Reference intervals of complete blood count constituents are highly correlated to waist circumference: Should obese patients have their own “normal values”? American Journal of Hematology 89, 7 (2014), 671–677.Google Scholar
- [72] . 2013. Estimating overweight risk in childhood from predictors during infancy. Pediatrics 132, 2 (2013), e414–e421.Google Scholar
- [73] . 2017. Deepr: A convolutional net for medical records. (2017).Google Scholar
- [74] . 2008. Vesicoureteral reflux. Journal of the American Society of Nephrology 19, 5 (2008), 847–862.Google Scholar
- [75] . 2019. Multiple MACE risk prediction using multi-task recurrent neural network with attention. In 2019 IEEE International Conference on Healthcare Informatics (ICHI ’19). IEEE, 1–2.Google Scholar
- [76] . 2012. ADADELTA: An adaptive learning rate method. ArXiv abs/1212.5701 (2012).Google Scholar
- [77] . 2018. Patient2vec: A personalized interpretable deep representation of the longitudinal electronic health record. IEEE Access 6 (2018), 65333–65346.Google ScholarCross Ref
- [78] . 2017. Using machine learning to predict obesity in high school students. In 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM’17). IEEE, 2132–2138.Google Scholar
- [79] . 2018. Predicting childhood overweight and obesity using maternal and early life risk factors: A systematic review. Obesity Reviews 19, 3 (2018), 302–312.Google Scholar
Index Terms
- Obesity Prediction with EHR Data: A Deep Learning Approach with Interpretable Elements
Recommendations
Predicting opioid overdose risk of patients with opioid prescriptions using electronic health records based on temporal deep learning
Graphical abstractDisplay Omitted
Highlights- We train a model to predict opioid overdose in the future based on past electronic health records.
AbstractThe US is experiencing an opioid epidemic, and opioid overdose is causing more than 100 deaths per day. Early identification of patients at high risk of Opioid Overdose (OD) can help to make targeted preventative interventions. We aim ...
Applying interpretable deep learning models to identify chronic cough patients using EHR data
Highlights- Evaluated language-based data representation with different learning models to predict chronic cough patients using EHR data.
- The best approach can gain 0.952 and 0.930 sensitivity and specificity, respectively.
- Identify chronic ...
Abstract Background and ObjectiveChronic cough (CC) affects approximately 10% of adults. Many disease states are associated with chronic cough, such as asthma, upper airway cough syndrome, bronchitis, and gastroesophageal reflux disease. The lack of an ...
Deep learning prediction models based on EHR trajectories: A systematic review
Abstract Background: Electronic health records (EHRs) are generated at an ever-increasing rate. EHR trajectories, the temporal aspect of health records, facilitate predicting patients’ future health-related risks. It enables healthcare systems to ...
Graphical abstractDisplay Omitted
Comments