Improving child health through Big Data and data science

Highly comparative time series analysis of oxygen saturation and heart rate to predict respiratory outcomes in extremely preterm infants

Objective.Highly comparative time series analysis (HCTSA) is a novel approach involving massive feature extraction using publicly available code from many disciplines. The Prematurity-Related Ventilatory Control (Pre-Vent) observational multicenter prospective study collected bedside monitor data from>700extremely preterm infants to identify physiologic features that predict respiratory outcomes.Approach. We calculated a subset of 33 HCTSA features on>7 M 10 min windows of oxygen saturation (SPO2) and heart rate (HR) from the Pre-Vent cohort to quantify predictive performance. This subset included representatives previously identified using unsupervised clustering on>3500HCTSA algorithms. We hypothesized that the best HCTSA algorithms would compare favorably to optimal PreVent physiologic predictor IH90_DPE (duration per event of intermittent hypoxemia events below 90%).Main Results.The top HCTSA features were from a cluster of algorithms associated with the autocorrelation of SPO2 time series and identified low frequency patterns of desaturation as high risk. These features had comparable performance to and were highly correlated with IH90_DPE but perhaps measure the physiologic status of an infant in a more robust way that warrants further investigation. The top HR HCTSA features were symbolic transformation measures that had previously been identified as strong predictors of neonatal mortality. HR metrics were only important predictors at early days of life which was likely due to the larger proportion of infants whose outcome was death by any cause. A simple HCTSA model using 3 top features outperformed IH90_DPE at day of life 7 (.778 versus .729) but was essentially equivalent at day of life 28 (.849 versus .850).Significance. These results validated the utility of a representative HCTSA approach but also provides additional evidence supporting IH90_DPE as an optimal predictor of respiratory outcomes.

The Effect of Human Growth Hormone Treatment on the Development of Slipped Capital Femoral Epiphysis: A Cohort Analysis With 6 Years of Follow-up

BACKGROUND

Slipped capital femoral epiphysis (SCFE) is a common hip disorder in adolescents that can result in substantial complications, impacting the quality of life. Human Growth Hormone (HGH) administration may elevate the risk of SCFE, though the relationship remains unclear. Clarifying this association could enable better monitoring and earlier diagnosis of SCFE in patients receiving HGH. The aim of the study is to investigate the association between HGH administration and the incidence of SCFE.

METHODS

This retrospective cohort study utilized data from the TriNetX research database from January 2003 to December 2022. The study included 2 cohorts: an HGH cohort including 36,791 patients aged below 18 years receiving HGH therapy and a control group consisting of patients who did not receive HGH therapy. A 1:1 propensity score matching technique was employed to ensure comparability between the HGH and no-HGH cohorts. The primary outcome measure was the development of SCFE identified by International Classification of Diseases codes. For comparative analysis, both risk ratios (RR) and hazard ratios were computed to evaluate the association between HGH therapy and the development of SCFE.

RESULTS

The HGH cohort had an increased risk of SCFE compared with the no-HGH cohort (RR: 3.5, 95% CI: 2.073, 5.909, P <0.001) and had an increased hazard of developing SCFE (hazard ratio: 2.627, 95% CI: 1.555, 4.437, P <0.001). Patients with higher exposure to HGH (defined as >10 prescriptions) had an RR of 1.914 (95% CI: 1.160, 3.159, P =0.010) when compared with their counterparts with ≤10 prescriptions.

CONCLUSIONS

In the largest study to date, HGH administration was associated with an elevated risk of SCFE in children in a dose-dependent manner.

LEVEL OF EVIDENCE

Level III-therapeutic retrospective cohort study.

The shifting trends in the epidemiology and risk factors of non-accidental fractures in children

BACKGROUND

Fractures are a common presentation of non-accidental trauma (NAT) in the pediatric population. However, the presentation could be subtle, and a high degree of suspicion is needed not to miss NAT.

OBJECTIVE

To analyze a comprehensive database, providing insights into the epidemiology of fractures associated with NAT.

PARTICIPANTS AND SETTING

The TriNetX Research Network was utilized for this study, containing medical records from 55 healthcare organizations. TriNetX was queried for all visits in children under the age of 6 years from 2015 to 2022, resulting in a cohort of over 32 million.

METHODS

All accidental and non-accidental fractures were extracted and analyzed to determine the incidence, fracture location, and demographics of NAT. Statistical analysis was done on a combination of Python and Epipy.

RESULTS

Overall, 0.36 % of all pediatric patients had a diagnosis of NAT, and 4.93 % of fractures (34,038 out of 689,740 total fractures) were determined to be non-accidental. Skull and face fractures constituted 17.9 % of all NAT fractures, but rib/sternum fractures had an RR = 6.7 for NAT. Children with intellectual and developmental disability (IDD) or autism spectrum disorder (ASD) had a 9 times higher risk for non-accidental fractures. The number of non-accidental fractures significantly increased after 2019.

CONCLUSIONS

The study findings suggest that nearly 1 out of all 20 fractures in children under age 6 are caused by NAT, and that rib/sternum fractures are most predictive of an inflicted nature. The study also showed a significant increase in the incidence of NAT, during and after the pandemic.

Highly comparative time series analysis of oxygen saturation and heart rate to predict respiratory outcomes in extremely preterm infants

Objective

Highly comparative time series analysis (HCTSA) is a novel approach involving massive feature extraction using publicly available code from many disciplines. The Prematurity-Related Ventilatory Control (Pre-Vent) observational multicenter prospective study collected bedside monitor data from > 700 extremely preterm infants to identify physiologic features that predict respiratory outcomes. We calculated a subset of 33 HCTSA features on > 7M 10-minute windows of oxygen saturation (SPO2) and heart rate (HR) from the Pre-Vent cohort to quantify predictive performance. This subset included representatives previously identified using unsupervised clustering on > 3500 HCTSA algorithms. Performance of each feature was measured by individual area under the receiver operating curve (AUC) at various days of life and binary respiratory outcomes. These were compared to optimal PreVent physiologic predictor IH90 DPE, the duration per event of intermittent hypoxemia events with threshold of 90%.

Main Results

The top HCTSA features were from a cluster of algorithms associated with the autocorrelation of SPO2 time series and identified low frequency patterns of desaturation as high risk. These features had comparable performance to and were highly correlated with IH90_DPE but perhaps measure the physiologic status of an infant in a more robust way that warrants further investigation. The top HR HCTSA features were symbolic transformation measures that had previously been identified as strong predictors of neonatal mortality. HR metrics were only important predictors at early days of life which was likely due to the larger proportion of infants whose outcome was death by any cause. A simple HCTSA model using 3 top features outperformed IH90_DPE at day of life 7 (.778 versus .729) but was essentially equivalent at day of life 28 (.849 versus .850). These results validated the utility of a representative HCTSA approach but also provides additional evidence supporting IH90_DPE as an optimal predictor of respiratory outcomes.

Proceedings from the CIHLMU 2022 Symposium: "Availability of and Access to Quality Data in Health"

Data is an essential tool for valid and reliable healthcare management. Access to high-quality data is critical to ensuring the early identification of problems, the design of appropriate interventions, and the effective implementation and evaluation of health intervention outcomes. During the COVID-19 pandemic, the need for strong information systems and the value of producing high-quality data for timely response and tracking resources and progress have been very evident across countries. The availability of and access to high-quality data at all levels of the health systems of low and middle-income countries is a challenge, which is exacerbated by multiple parallels and poorly integrated data sources, a lack of data-sharing standards and policy frameworks, their weak enforcement, and inadequate skills among those handling data. Completeness, accuracy, integrity, validity, and timeliness are challenges to data availability and use. "Big Data" is a necessity and a challenge in the current complexities of health systems. In transitioning to digital systems with proper data standards and policy frameworks for privacy protection, data literacy, ownership, and data use at all levels of the health system, skill enhancement of the staff is critical. Adequate funding for strengthening routine information systems and periodic surveys and research, and reciprocal partnerships between high-income countries and low- and middle-income countries in data generation and use, should be prioritized by the low- and middle-income countries to foster evidence-based healthcare practices.