<p><strong>OBJECTIVE: </strong>This study compares seven machine learning models developed to predict childhood obesity from age > 2 to ≤ 7 years using Electronic Healthcare Record (EHR) data up to age 2 years.</p>
<p><strong>MATERIALS AND METHODS: </strong>EHR data from of 860,510 patients with 11,194,579 healthcare encounters were obtained from the Children's Hospital of Philadelphia. After applying stringent quality control to remove implausible growth values and including only individuals with all recommended wellness visits by age 7 years, 27,203 (50.78 % male) patients remained for model development. Seven machine learning models were developed to predict obesity incidence as defined by the Centers for Disease Control and Prevention (age/sex adjusted BMI>95th percentile). Model performance was evaluated by multiple standard classifier metrics and the differences among seven models were compared using the Cochran's Q test and post-hoc pairwise testing.</p>
<p><strong>RESULTS: </strong>XGBoost yielded 0.81 (0.001) AUC, which outperformed all other models. It also achieved statistically significant better performance than all other models on standard classifier metrics (sensitivity fixed at 80 %): precision 30.90 % (0.22 %), F1-socre 44.60 % (0.26 %), accuracy 66.14 % (0.41 %), and specificity 63.27 % (0.41 %).</p>
<p><strong>DISCUSSION AND CONCLUSION: </strong>Early childhood obesity prediction models were developed from the largest cohort reported to date. Relative to prior research, our models generalize to include males and females in a single model and extend the time frame for obesity incidence prediction to 7 years of age. The presented machine learning model development workflow can be adapted to various EHR-based studies and may be valuable for developing other clinical prediction models.</p>
<p>Fine particulate matter (PM) and ozone (O) air pollutants are known risk factors for asthma exacerbation. We studied the association of these air pollutants with pediatric asthma exacerbation in the Philadelphia metropolitan region, and evaluated potential effect modification by children's characteristics (e.g., race/ethnicity, atopic conditions) and environmental factors (e.g., neighborhood tree canopy, meteorological factors, aeroallergens). We conducted a time-stratified case-crossover study of 54,632 pediatric (age ≤18 years) asthma exacerbation cases occurring from 2011-2014, identified through electronic health records (EHR) of the Children's Hospital of Philadelphia (CHOP) health system. We applied conditional logistic regression to estimate associations between air pollution and asthma exacerbation, using daily census-tract level pollutant concentrations estimated from the EPA Fused Air Quality Surface Using Downscaling (FAQSD) files. The associations were estimated within warm (Apr - Sep) and cold (Oct - Mar) months for unlagged exposure and for cumulative effects up to 5 days after exposure, with adjustment for temperature, relative humidity, and holidays. We found small increases in odds of asthma exacerbation with higher pollutant concentrations, with positive associations (OR, comparing concentrations of 75 to 25 percentile) observed for PM during both warm (1.03, 95% CI: 0.98 - 1.08) and cold months (1.05, 95% CI: 1.02 - 1.07), and for O during cold months (1.08, 95% CI: 1.02 - 1.14). The exposure-response relationship with PM during the cold months was essentially linear, whereas thresholds of effect were observed for the other associations at low-medium pollutant concentrations. Results were robust to multi-pollutant modeling and adjustment for additional covariates. We found no effect modification by most children's characteristics, while effect sizes were higher on days with detected tree and grass pollens during warm months. Our results suggest that even small decreases in pollutant concentrations could potentially reduce risk of childhood asthma exacerbation - an important finding, given the high burden of childhood asthma and known disparities in asthma control.</p>
<p>Pollen from trees, grasses, and weeds can trigger asthma exacerbation in sensitized individuals. However, there are gaps in knowledge about the effects, such as the relative risks from different plant taxa and threshold levels of effect. We aimed to describe the local association between pollen and asthma exacerbation among children in the City of Philadelphia, and to evaluate whether effects are modified by children's characteristics and clinical factors (e.g., child's age, race/ethnicity, comorbidities). We conducted a time-stratified case-crossover study of pediatric (age <18 years) asthma exacerbation, with cases identified through electronic health records (EHR) of the Children's Hospital of Philadelphia (CHOP) health system from March through October in the years 2011-2016. Daily pollen counts were obtained from the local National Allergy Bureau certified pollen counter. We applied conditional logistic regression to estimate odds ratios (OR) and 95% confidence intervals (CI) for the association between the pollen level (vs. none detected) and odds of asthma exacerbation, adjusting for temperature, relative humidity, and holidays. We estimated same-day exposure effects, as well as effects from exposure lagged by up to 5 days. There were 35,040 asthma exacerbation events during the study period, with the majority occurring among black, non-Hispanic children (81.8%) and boys (60.4%). We found increased odds of asthma exacerbation among Philadelphia children in association with tree pollen, both for total tree pollen and most individual tree types. Increased odds from total tree pollen were observed at the lowest levels studied (≤5 grains/m, unlagged, OR = 1.06, 95% CI: 1.02, 1.10), and exhibited a positive exposure-response pattern of effect; tree pollen levels above 1000 grains/m (unlagged) were associated with 64% increased odds of asthma exacerbation (95% CI: 1.45, 1.84). Grass pollen was associated with asthma exacerbation only at levels above the 99 percentile (52 grains/m), which occurred, on average, two days per year during the study period (with 2-day lag, OR = 1.38, 95% CI: 1.19, 1.60). There was an inverse association (reduced asthma exacerbation) with ragweed pollen that was consistent across analyses. Pollen from other weeds was associated with increased odds of asthma exacerbation, without a clear exposure-response pattern (2-day lag, significant increases ranging from 8% to 19%). Increased odds from tree pollen and weeds (other than ragweed) were higher among children with allergic rhinitis. While there are known benefits from urban vegetation for human health, there are risks as well. It is important to note, however, that pollen is released during a limited time frame each year, and advisories informed by local data can enable susceptible individuals to avoid outdoor exposure on high-risk days.</p>
<p>Extreme precipitation events may be an important environmental trigger for asthma exacerbations in children. We used a time stratified case-crossover design and data from a large electronic health record database at the Children's Hospital of Philadelphia (CHOP) to estimate associations of daily heavy precipitation (defined as > 95th percentile of the summertime distribution) with asthma exacerbation among children. We defined control days as those falling on the same day of the week within the same month and year as the case. We restricted our primary analyses to the summer months in years 2011-2016 and used conditional logistic regression models to estimate associations between heavy precipitation and acute asthma exacerbations in both outpatient (primary care, specialty care, and emergency department) and inpatient settings. We investigated numerous individual-level (e.g., age, sex, eczema diagnosis) and environmental measures (e.g., greenspace, particulate matter) as potential effect modifiers. The analysis include 13,483 asthma exacerbations in 10,434 children. Odds of asthma exacerbation were 11% higher on heavy precipitation vs. no precipitation days (95% CI: 1.02-1.21). There was little evidence of effect modification by most measures. These results suggest that heavy summertime precipitation events may contribute to asthma exacerbations. Further research using larger datasets from other health systems is needed to confirm these results, and to explore underlying mechanisms.</p>
<p><strong>OBJECTIVES: </strong>Our objectives were to (1) quantify the frequency of wheezing episodes and asthma diagnosis in young children in a large pediatric primary care network and (2) assess the variability in practice-level asthma diagnosis, accounting for common asthma risk factors and comorbidities. We hypothesized that significant variability in practice-level asthma diagnosis rates would remain after adjusting for associated predictors.</p>
<p><strong>METHODS: </strong>We generated a retrospective longitudinal birth cohort of children who visited one of 31 pediatric primary care practices within the first 6 months of life from 1/2005-12/2016. Children were observed for up to 8 years or until the end of the observation window. We used multivariable discrete time survival models to evaluate predictors of asthma diagnosis by 3-month age intervals. We compared unadjusted and adjusted proportions of children diagnosed with asthma by practice.</p>
<p><strong>RESULTS: </strong>Of the 161,502 children in the cohort, 34,578 children (21%) received at least one asthma diagnosis. In multivariable modeling, male gender, minority race/ethnicity, gestational age <34 weeks, allergic rhinitis, food allergy, and prior wheezing episodes were associated with asthma diagnosis. After adjusting for variation in these predictors across practices, the cumulative incidence of asthma diagnosis by practice by age 6 years ranged from 11-47% (interquartile range (IQR): 24-29%).</p>
<p><strong>CONCLUSIONS: </strong>Across pediatric primary care practices, adjusted incidence of asthma diagnosis by age 6 years ranged widely, though variation gauged by the IQR was more modest. Potential sources of practice-level variation, such as differing diagnosis thresholds and labeling of different wheezing phenotypes as "asthma", should be further investigated.</p>
<p><strong>BACKGROUND: </strong>PROMIS Pediatric patient-reported outcome measures were developed with children from the general population, and their content validity has not been established in children with chronic disease. This study was done to evaluate the content validity of the PROMIS Pediatric Pain Interference and Fatigue measures in children 8-17 years-old with Crohn's disease and the PROMIS Pediatric Fatigue, Sleep Disturbance, and Sleep-related Impairment measures for children 8-17 years-old with chronic kidney disease.</p>
<p><strong>METHODS: </strong>We conducted semi-structured interviews with individuals affected by Crohn's disease and chronic kidney disease. The interviews were done to elicit children's lived experiences of the PROMIS outcomes of interest. We used deductive content analysis to contrast the participants' reports of their symptoms and impacts on daily life with existing conceptual frameworks for the PROMIS measures, each of which was developed with input from children in the general population.</p>
<p><strong>RESULTS: </strong>On average, we elicited an average of 7 pain interference and 7 fatigue concepts from Crohn's disease participants (n = 37), while chronic kidney disease participants (n = 26) provided 9 concepts for fatigue, 4 for sleep disturbance, and 7 for sleep-related impairment. Concept saturation was achieved after 16-19 interviews across the four PROMIS measures. Children with these two chronic health conditions reported the same breadth and types of lived experiences as children from the development samples drawn from the general population.</p>
<p><strong>CONCLUSION: </strong>The study supports the content validity of several PROMIS Pediatric measures for children with Crohn's disease and chronic kidney disease. These findings provide evidence that PROMIS Pediatric measures, developed as universally relevant patient-reported outcomes, may be more broadly applicable to children with chronic disease.</p>
<p><strong>BACKGROUND: </strong>The rarity of pediatric glomerular disease makes it difficult to identify sufficient numbers of participants for clinical trials. This leaves limited data to guide improvements in care for these patients.</p>
<p><strong>METHODS: </strong>The authors developed and tested an electronic health record (EHR) algorithm to identify children with glomerular disease. We used EHR data from 231 patients with glomerular disorders at a single center to develop a computerized algorithm comprising diagnosis, kidney biopsy, and transplant procedure codes. The algorithm was tested using PEDSnet, a national network of eight children's hospitals with data on >6.5 million children. Patients with three or more nephrologist encounters (=55,560) not meeting the computable phenotype definition of glomerular disease were defined as nonglomerular cases. A reviewer blinded to case status used a standardized form to review random samples of cases (=800) and nonglomerular cases (=798).</p>
<p><strong>RESULTS: </strong>The final algorithm consisted of two or more diagnosis codes from a qualifying list or one diagnosis code and a pretransplant biopsy. Performance characteristics among the population with three or more nephrology encounters were sensitivity, 96% (95% CI, 94% to 97%); specificity, 93% (95% CI, 91% to 94%); positive predictive value (PPV), 89% (95% CI, 86% to 91%); negative predictive value, 97% (95% CI, 96% to 98%); and area under the receiver operating characteristics curve, 94% (95% CI, 93% to 95%). Requiring that the sum of nephrotic syndrome diagnosis codes exceed that of glomerulonephritis codes identified children with nephrotic syndrome or biopsy-based minimal change nephropathy, FSGS, or membranous nephropathy, with 94% sensitivity and 92% PPV. The algorithm identified 6657 children with glomerular disease across PEDSnet, ≥50% of whom were seen within 18 months.</p>
<p><strong>CONCLUSIONS: </strong>The authors developed an EHR-based algorithm and demonstrated that it had excellent classification accuracy across PEDSnet. This tool may enable faster identification of cohorts of pediatric patients with glomerular disease for observational or prospective studies.</p>
<p><strong>OBJECTIVE: </strong>To evaluate the reliability and validity of the PROMIS Pediatric Global Health scale, a 7-item measure of perceived physical, mental, and social health, in children with asthma.</p>
<p><strong>METHODS: </strong>From February 2014 to February 2015, convenience samples of children 8-17 years-old (n = 182) and parents of children 5-17 years-old (n = 328) visiting an emergency department for treatment of asthma were enrolled. The Asthma Control Test was used to characterize children as controlled versus not controlled, and the PROMIS Asthma Impact Scale was used to assess the effects of asthma symptoms on functional status. We conducted longitudinal analyses among 92 children and 218 parents at 3 weeks, and 74 children and 171 parents at 8 weeks after enrollment.</p>
<p><strong>RESULTS: </strong>The PGH-7 reliability across the three time points ranged from 0.66 to 0.81 for child-report and 0.76 to 0.82 for parent-proxy. In cross-sectional analyses, children with controlled asthma had PGH-7 scores 0.40-0.95 standard deviation units higher than those who were uncontrolled. The PGH-7 was responsive to changes in overall general health between time points, with moderate effect sizes (0.5-0.6 standard deviation units). In longitudinal analyses, PGH-7 scores were no different between those who stayed uncontrolled versus became controlled at 3 weeks of follow-up; however, by 8 weeks of follow-up, the differences between these groups was 0.7-0.8 standard deviation units, indicative of large effects.</p>
<p><strong>CONCLUSIONS: </strong>The PGH-7 is a reliable and valid patient-reported outcome for assessing general health among children with asthma. It is a useful complement to other asthma-specific outcome measures.</p>
<p>Publicly financed insurance programs are tasked with maintaining coverage for eligible children, but published measures to assess coverage have not been evaluated. Therefore, we sought to identify and categorize measures of health insurance continuity for children and adolescents. We conducted a systematic review of Medline and HealthStar databases, review of reference lists of eligible articles, and contact with experts. We categorized measures into 8 domains based on a conceptual framework. We identified 147 measures from 84 eligible articles. Most measures evaluated the following domains: always insured (41%), repeatedly uninsured (36%), and transition out of coverage (29%), while fewer assessed single gap in coverage, always uninsured, transition into coverage, change in coverage, and eligibility. Only 18% of measures assessed associations between continuity of coverage and child and adolescent health outcomes. These results suggest that a number of measures of continuity of coverage exist, but few measures have assessed impact on outcomes.</p>