A systematic review of predictive modeling for bronchiolitis. Int J Med Inform. 2014;83:691-714. [PMID: 25106933 DOI: 10.1016/j.ijmedinf.2014.07.005]

Clinical Features of Patients With Bronchiolitis Prior to the Initiation of Noninvasive Respiratory Support

Background: Effective management of infants hospitalized with bronchiolitis depends on clinician assessment of disease severity. Although environmental and demographic risk factors help identify severe cases, there is limited research on specific clinical and physiological characteristics associated with respiratory deterioration. This study aimed to identify physiologic variables and clinical parameters associated with respiratory deterioration in hospitalized infants with bronchiolitis. Methods: A single-center retrospective cohort study included previously healthy infants <2 years of age hospitalized for bronchiolitis. The primary outcome measure, deterioration, was defined as respiratory distress requiring noninvasive (including high-flow nasal cannula) or invasive respiratory support within 48 h of admission. A multivariable logistic regression analysis with preselected factors was used to assess the odds of deterioration. Variables included sex, age, affect and behavior, nasopharyngeal suctioning, number, location of retractions, SpO2/FIO2 (S/F ratio), breathing frequency, pulse rate, and respiratory severity score. A secondary analysis assessed retraction locations. Results: Of the 584 eligible patients, 154 (26%) experienced a deterioration event and required noninvasive or invasive respiratory support. Respiratory score (odds ratio [OR] 1.9 [95% CI 1.5-2.4]), total number of retractions (OR: 2.5 [95% CI 1.6-3.8]), S/F ratio (OR: 1.0 [95% CI 0.99-0.998), pulse rate (OR: 1.0 [95% CI 1.0-1.1]), nasopharyngeal suctioning (OR: 5.5 [95% CI 2.6-11.7]), and positive affect and behavior descriptors (OR: 0.3 [95% CI 0.1-0.7]) were associated with deterioration. Age, sex, negative affect and behavior descriptors, and breathing frequency were not statistically significant. Conclusions: These variables may be used to design predictive algorithms that alert clinicians of impending respiratory deterioration in infants with bronchiolitis.

RSV Severe Infection Risk Stratification in a French 5-Year Birth Cohort Using Machine-learning

BACKGROUND

Respiratory syncytial virus (RSV) poses a substantial threat to infants, often leading to challenges in hospital capacity. With recent pharmaceutical developments to be used during the prenatal and perinatal periods aimed at decreasing the RSV burden, there is a pressing need to identify infants at risk of severe disease. We aimed to stratify the risk of developing a clinically severe RSV infection in infants under 1 year of age.

METHODS

This retrospective observational study was conducted at the Hospices Civils de Lyon, France, involving infants born between 2014 and 2018. This study focused on infants hospitalized with severe and very severe acute lower respiratory tract infections associated with RSV (SARI-WI group). Data collection included perinatal information and clinical data, with machine-learning algorithms used to discriminate SARI-WI cases from nonhospitalized infants.

RESULTS

Of 42,069 infants, 555 developed SARI-WI. Infants born in November were very likely (>80%) predicted SARI-WI. Infants born in October were very likely predicted SARI-WI except for births at term by vaginal delivery and without siblings. Infants were very unlikely (<10%) predicted SARI-WI when all the following conditions were met: born in other months, at term, by vaginal delivery and without siblings. Other infants were possibly (10-30%) or probably (30-80%) predicted SARI-WI.

CONCLUSIONS

Although RSV preventive measures are vital for all infants, and specific recommendations exist for patients with high-risk comorbidities, in situations where prioritization becomes necessary, infants born just before or within the early weeks of the epidemic should be considered as a risk group.

Personalized Infant Risk Prediction for Severe Respiratory Syncytial Virus Lower Respiratory Tract Infection Requiring Intensive Care Unit Admission

Background

Currently, there are no available tools to identify infants at the highest risk of significant morbidity and mortality from respiratory syncytial virus (RSV) lower respiratory tract infection (LRTI) who would benefit most from RSV prevention products. The objective was to develop and internally validate a personalized risk prediction tool for use among all newborns that uses readily available birth/postnatal data to predict RSV LRTI requiring intensive care unit (ICU) admission.

Methods

We conducted a population-based birth cohort study of infants born from 1995 to 2007, insured by the Tennessee Medicaid Program, and who did not receive RSV immunoprophylaxis during the first year of life. The primary outcome was severe RSV LRTI requiring ICU admission during the first year of life. We built a multivariable logistic regression model including demographic and clinical variables available at or shortly after birth to predict the primary outcome.

Results

In a population-based sample of 429 365 infants, 713 (0.2%) had severe RSV LRTI requiring ICU admission. The median age of admission was 66 days (interquartile range, 37-120). Our tool, including 19 variables, demonstrated good predictive accuracy (area under the curve, 0.78; 95% confidence interval, 0.77-0.80) and identified infants who did not qualify for palivizumab, based on American Academy of Pediatrics guidelines, but had higher predicted risk levels than infants who qualified (27% of noneligible infants with >0.16% predicted probabilities [lower quartile for eligible infants]).

Conclusions

We developed a personalized tool that identified infants at increased risk for severe RSV LRTI requiring ICU admission, expected to benefit most from immunoprophylaxis.

Clinical practice guidelines: management of severe bronchiolitis in infants under 12 months old admitted to a pediatric critical care unit

PURPOSE

We present guidelines for the management of infants under 12 months of age with severe bronchiolitis with the aim of creating a series of pragmatic recommendations for a patient subgroup that is poorly individualized in national and international guidelines.

METHODS

Twenty-five French-speaking experts, all members of the Groupe Francophone de Réanimation et Urgence Pédiatriques (French-speaking group of paediatric intensive and emergency care; GFRUP) (Algeria, Belgium, Canada, France, Switzerland), collaborated from 2021 to 2022 through teleconferences and face-to-face meetings. The guidelines cover five areas: (1) criteria for admission to a pediatric critical care unit, (2) environment and monitoring, (3) feeding and hydration, (4) ventilatory support and (5) adjuvant therapies. The questions were written in the Patient-Intervention-Comparison-Outcome (PICO) format. An extensive Anglophone and Francophone literature search indexed in the MEDLINE database via PubMed, Web of Science, Cochrane and Embase was performed using pre-established keywords. The texts were analyzed and classified according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology. When this method did not apply, an expert opinion was given. Each of these recommendations was voted on by all the experts according to the Delphi methodology.

RESULTS

This group proposes 40 recommendations. The GRADE methodology could be applied for 17 of them (3 strong, 14 conditional) and an expert opinion was given for the remaining 23. All received strong approval during the first round of voting.

CONCLUSION

These guidelines cover the different aspects in the management of severe bronchiolitis in infants admitted to pediatric critical care units. Compared to the different ways to manage patients with severe bronchiolitis described in the literature, our original work proposes an overall less invasive approach in terms of monitoring and treatment.

AI Models for Predicting Readmission of Pneumonia Patients within 30 Days after Discharge

A model with capability for precisely predicting readmission is a target being pursued worldwide. The objective of this study is to design predictive models using artificial intelligence methods and data retrieved from the National Health Insurance Research Database of Taiwan for identifying high-risk pneumonia patients with 30-day all-cause readmissions. An integrated genetic algorithm (GA) and support vector machine (SVM), namely IGS, were used to design predictive models optimized with three objective functions. In IGS, GA was used for selecting salient features and optimal SVM parameters, while SVM was used for constructing the models. For comparison, logistic regression (LR) and deep neural network (DNN) were also applied for model construction. The IGS model with AUC used as the objective function achieved an accuracy, sensitivity, specificity, and area under ROC curve (AUC) of 70.11%, 73.46%, 69.26%, and 0.7758, respectively, outperforming the models designed with LR (65.77%, 78.44%, 62.54%, and 0.7689, respectively) and DNN (61.50%, 79.34%, 56.95%, and 0.7547, respectively), as well as previously reported models constructed using thedata of electronic health records with an AUC of 0.71–0.74. It can be used for automatically detecting pneumonia patients with a risk of all-cause readmissions within 30 days after discharge so as to administer suitable interventions to reduce readmission and healthcare costs.

[Predictors of hospitalization plus airway support among infants with recurrent wheezing in the emergency department]

OBJECTIVE

Most patients with recurrent wheezing are infants under 2 years of age. Clinical prediction models of the risk of receiving airway support during the hospital stay in this population have been poorly studied in tropical countries. This study aimed to evaluate the clinical predictors of hospitalization plus airway support among infants with recurrent wheezing evaluated in the emergency department in Colombia.

METHODS

A retrospective cohort study was performed. This study included all infants with two or more wheezing episodes who were younger than two years old in two tertiary centers in Rionegro, Colombia, between January 2019 and December 2019. The primary outcome measure was hospitalization plus any airway support. A multivariable logistic regression model was used to identify factors independently associated with hospitalization plus any airway support.

RESULTS

A total of 85 infants were hospitalized plus any airway support, of whom 34(40%) were treated with high flow nasal canula, 2(2%) received non-invasive ventilation, 6(7%) were mechanically ventilated, and 43 (51%) received conventional oxygen therapy. The multivariable logistic regression model showed that predictors of hospitalization plus airway support included prematurity (OR=1.79, 95%CI: 1.04-3.10), poor feeding (OR=2.22, 95%CI: 1.25-3.94), nasal flaring and/or grunting (OR=4.27, 95%CI: 2.41-7.56), and previous wheezing episodes requiring hospitalization (OR=3.36, 95%CI: 1.86-7.08). The model has a high specificity (99.6%) with acceptable discrimination and an area under the curve of 0.70(95%CI: 0.60-0.74).

CONCLUSIONS

The present study shows that prematurity, poor feeding, nasal flaring and/or grunting, and more than one previous episode of wheezing requiring hospitalization are independent predictors of hospitalization plus airway support in a population of infants with recurrent wheezing in the emergency department. More evidence must be collected to examine the results in other tropical countries.

Risk factors for severe bronchiolitis in Colombia

High medical costs of treatment of severe bronchiolitis in infants impose a severe economic burden, especially in tropical middle-income countries. There is a critical need therefore to explore the risk factors concerned. In our retrospective cohort study, we included all infants younger than two years admitted in Rionegro, Colombia, owing to bronchiolitis. We used log-binomial regression and estimate prevalence ratios. Out of a total of 417 included, 300 (72.12%) had severe bronchiolitis, with respiratory syncytial virus and current exposure to cigarette smoking being independent predictors.

A Clinical Decision Support System (KNOWBED) to Integrate Scientific Knowledge at the Bedside: Development and Evaluation Study

BACKGROUND

The evidence-based medicine (EBM) paradigm requires the development of health care professionals' skills in the efficient search of evidence in the literature, and in the application of formal rules to evaluate this evidence. Incorporating this methodology into the decision-making routine of clinical practice will improve the patients' health care, increase patient safety, and optimize resources use.

OBJECTIVE

The aim of this study is to develop and evaluate a new tool (KNOWBED system) as a clinical decision support system to support scientific knowledge, enabling health care professionals to quickly carry out decision-making processes based on EBM during their routine clinical practice.

METHODS

Two components integrate the KNOWBED system: a web-based knowledge station and a mobile app. A use case (bronchiolitis pathology) was selected to validate the KNOWBED system in the context of the Paediatrics Unit of the Virgen Macarena University Hospital (Seville, Spain). The validation was covered in a 3-month pilot using 2 indicators: usability and efficacy.

RESULTS

The KNOWBED system has been designed, developed, and validated to support clinical decision making in mobility based on standards that have been incorporated into the routine clinical practice of health care professionals. Using this tool, health care professionals can consult existing scientific knowledge at the bedside, and access recommendations of clinical protocols established based on EBM. During the pilot project, 15 health care professionals participated and accessed the system for a total of 59 times.

CONCLUSIONS

The KNOWBED system is a useful and innovative tool for health care professionals. The usability surveys filled in by the system users highlight that it is easy to access the knowledge base. This paper also sets out some improvements to be made in the future.

Importance of respiratory syncytial virus as a predictor of hospital length of stay in bronchiolitis

IIntroduction : Bronchiolitis is the leading cause of hospitalization in children. Estimate potentially preventable variables that impact the length of hospital stay are a priority to reduce the costs associated with this disease. This study aims to identify clinical variables associated with length of hospital stay of bronchiolitis in children in a tropical middle-income country Methods: We conducted a retrospective cohort study in 417 infants with bronchiolitis in tertiary centers in Colombia. All medical records of all patients admitted to the emergency department were reviewed. To identify factors independently associated we use negative binomial regression model, to estimate incidence rate ratios (IRR) and adjust for potential confounding variables Results : The median of the length of hospital stay was 3.68 days, with a range of 0.74 days to 29 days, 138 (33.17%) of patients have a hospital stay of 5 or more days. After modeling and controlling for potential confounders age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, RSV isolation, and C-reactive protein were independent predictors of LOS Conclusions : Our results show that in infants with bronchiolitis, RSV isolation, age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, and C-reactive protein were independent predictors of LOS. As a potentially modifiable risk factor, efforts to reduce the probability of RSV infection can reduce the high medical cost associates with prolonged LOS in bronchiolitis.

Importance of respiratory syncytial virus as a predictor of hospital length of stay in bronchiolitis

Introduction: Bronchiolitis is the leading cause of hospitalization in children. Estimate potentially preventable variables that impact the length of hospital stay are a priority to reduce the costs associated with this disease. This study aims to identify clinical variables associated with length of hospital stay of bronchiolitis in children in a tropical middle-income country Methods: We conducted a retrospective cohort study in 417 infants with bronchiolitis in tertiary centers in Colombia. All medical records of all patients admitted to the emergency department were reviewed. To identify factors independently associated we use negative binomial regression model, to estimate incidence rate ratios (IRR) and adjust for potential confounding variables Results: The median of the length of hospital stay was 3.68 days, with a range of 0.74 days to 29 days, 138 (33.17%) of patients have a hospital stay of 5 or more days. After modeling and controlling for potential confounders age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, RSV isolation, and C-reactive protein were independent predictors of LOS Conclusions: Our results show that in infants with bronchiolitis, RSV isolation, age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, and C-reactive protein were independent predictors of LOS. As a potentially modifiable risk factor, efforts to reduce the probability of RSV infection can reduce the high medical cost associates with prolonged LOS in bronchiolitis.

Importance of respiratory syncytial virus as a predictor of hospital length of stay in bronchiolitis

Introduction : Bronchiolitis is the leading cause of hospitalization in children. Estimate potentially preventable variables that impact the length of hospital stay are a priority to reduce the costs associated with this disease. This study aims to identify clinical variables associated with length of hospital stay of bronchiolitis in children in a tropical middle-income country Methods: We conducted a retrospective cohort study in 417 infants with bronchiolitis in tertiary centers in Colombia. All medical records of all patients admitted through the emergency department were reviewed. To identify factors independently associated we use negative binomial regression model, to estimate incidence rate ratios (IRR) and adjust for potential confounding variables Results : The median of the length of hospital stay was 3.68 days, with a range of 0.74 days to 29 days, 138 (33.17%) of patients have a hospital stay of 5 or more days. After modeling and controlling for potential confounders age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, detection of RSV, and C-reactive protein were independent predictors of LOS Conclusions : Our results show that in infants with bronchiolitis, detection of RSV, age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, and C-reactive protein were independent predictors of LOS. As a potentially modifiable risk factor, efforts to reduce the probability of RSV infection can reduce the high medical cost associates with prolonged LOS in bronchiolitis.

Importance of respiratory syncytial virus as a predictor of hospital length of stay in bronchiolitis

Introduction : Bronchiolitis is the leading cause of hospitalization in children. Estimate potentially preventable variables that impact the length of hospital stay are a priority to reduce the costs associated with this disease. This study aims to identify clinical variables associated with length of hospital stay of bronchiolitis in children in a tropical middle-income country Methods: We conducted a retrospective cohort study in 417 infants with bronchiolitis in tertiary centers in Colombia. All medical records of all patients admitted to the emergency department were reviewed. To identify factors independently associated we use negative binomial regression model, to estimate incidence rate ratios (IRR) and adjust for potential confounding variables Results : The median of the length of hospital stay was 3.68 days, with a range of 0.74 days to 29 days, 138 (33.17%) of patients have a hospital stay of 5 or more days. After modeling and controlling for potential confounders age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, RSV isolation, and C-reactive protein were independent predictors of LOS Conclusions : Our results show that in infants with bronchiolitis, RSV isolation, age <6 months, comorbidities (CHD or neurological), BPD,  chest indrawing, and C-reactive protein were independent predictors of LOS. As a potentially modifiable risk factor, efforts to reduce the probability of RSV infection can reduce the high medical cost associates with prolonged LOS in bronchiolitis.

Inappropriate antibiotic prescribing for acute bronchiolitis in Colombia: a predictive model

Introduction

Acute bronchiolitis is the leading cause of hospitalization in the pediatric population. The inappropriate prescription of antibiotics in acute bronchiolitis is associated with bacterial resistance, higher costs, and risk of adverse effects in this population. The objective of this work is to develop a predictive model of inappropriate use of antibiotics in children with acute bronchiolitis in Colombia.

Methods

A retrospective cohort study was conducted in patients under 2 years of age with a diagnosis of acute bronchiolitis from two hospitals in Rionegro, Colombia. To identify factors independently associated with inappropriate use of antibiotics, we used logistic regression and estimated odds ratios (ORs). To assess discrimination, area under the curve (AUC) was estimated with a 95% confidence interval and plotted using AUC–ROC plots. To correct sampling bias of variance parameters and to evaluate the internal validity of the model, repeated curved validation “tenfold cross-validation” was used, comparing the area under the ROC curve obtained in the repetitions with that observed in the model

Results

A total of 415 patients were included. 142 patients (34.13%) had a prescription of some antibiotic during their hospital stay. In 92 patients (64.78%, 95% CI 56.3 to 72.6%) the prescription of antibiotics was classified as inappropriate. Age older than 1 year, chest retractions, temperature between 37.5 °C and 38.5 °C and leukocyte count between 10,000 and 15,000 million/mm³ were the predictive variables of inappropriate use of medications in this population.

Conclusion

The presence of fever between 37.5 °C and 38.5 °C, leukocytosis between 10,000 and 15,000 million/mm³, and age older than 1 year and presence of chest retractions, should alert the physician regarding the high risk of inappropriate prescription of antibiotics. Patients with acute bronchiolitis with a score on our scale greater than 2 should be carefully evaluated regarding the need for the use of antibiotics, if prescribed.

Viral lower respiratory tract infections-strict admission guidelines for young children can safely reduce admissions

Viral lower respiratory tract infection (VLRTI) is the most common cause of hospital admission among small children in high-income countries. Guidelines to identify children in need of admission are lacking in the literature. In December 2012, our hospital introduced strict guidelines for admission. This study aims to retrospectively evaluate the safety and efficacy of the guidelines. We performed a single-center retrospective administrative database search and medical record review. ICD-10 codes identified children < 24 months assessed at the emergency department for VLRTI for a 10-year period. To identify adverse events related to admission guidelines implementation, we reviewed patient records for all those discharged on primary contact followed by readmission within 14 days. During the study period, 3227 children younger than 24 months old were assessed in the ED for VLRTI. The proportion of severe adverse events among children who were discharged on their initial emergency department contact was low both before (0.3%) and after the intervention (0.5%) (p=1.0). Admission rates before vs. after the intervention were for previously healthy children > 90 days 65.3% vs. 53.3% (p<0.001); for healthy children ≤ 90 days 85% vs. 68% (p<0.001); and for high-risk comorbidities 74% vs. 71% (p=0.5).Conclusion: After implementation of admission guidelines for VLRTI, there were few adverse events and a significant reduction in admissions to the hospital from the emergency department. Our admission guidelines may be a safe and helpful tool in the assessment of children with VLRTI. What is Known: • Viral lower respiratory tract infection, including bronchiolitis, is the most common cause of hospitalization for young children in the developed world. Treatment is mainly supportive, and hospitalization should be limited to the cases in need of therapeutic intervention. • Many countries have guidelines for the management of the disease, but the decision on whom to admit for inpatient treatment is often subjective and may vary even between physicians in the same hospital. What is New: • Implementation of admission criteria for viral lower respiratory tract infection may reduce the rate of hospital admissions without increasing adverse events.

Models for estimating and projecting global, regional and national prevalence and disease burden of asthma: a systematic review

Background

Statistical models are increasingly being used to estimate and project the prevalence and burden of asthma. Given substantial variations in these estimates, there is a need to critically assess the properties of these models and assess their transparency and reproducibility. We aimed to critically appraise the strengths, limitations and reproducibility of existing models for estimating and projecting the global, regional and national prevalence and burden of asthma.

Methods

We undertook a systematic review, which involved searching Medline, Embase, World Health Organization Library and Information Services (WHOLIS) and Web of Science from 1980 to 2017 for modelling studies. Two reviewers independently assessed the eligibility of studies for inclusion and then assessed their strengths, limitations and reproducibility using pre-defined quality criteria. Data were descriptively and narratively synthesised.

Results

We identified 108 eligible studies, which employed a total of 51 models: 42 models were used to derive national level estimates, two models for regional estimates, four models for global and regional estimates and three models for global, regional and national estimates. Ten models were used to estimate the prevalence of asthma, 27 models estimated the burden of asthma – including, health care service utilisation, disability-adjusted life years, mortality and direct and indirect costs of asthma – and 14 models estimated both the prevalence and burden of asthma. Logistic and linear regression models were most widely used for national estimates. Different versions of the DisMod-MR- Bayesian meta-regression models and Cause Of Death Ensemble model (CODEm) were predominantly used for global, regional and national estimates. Most models suffered from a number of methodological limitations – in particular, poor reporting, insufficient quality and lack of reproducibility.

Conclusions

Whilst global, regional and national estimates of asthma prevalence and burden continue to inform health policy and investment decisions on asthma, most models used to derive these estimates lack the required reproducibility. There is a need for better-constructed models for estimating and projecting the prevalence and disease burden of asthma and a related need for better reporting of models, and making data and code available to facilitate replication.

Machine learning-based prediction of acute severity in infants hospitalized for bronchiolitis: a multicenter prospective study

We aimed to develop machine learning models to accurately predict bronchiolitis severity, and to compare their predictive performance with a conventional scoring (reference) model. In a 17-center prospective study of infants (aged < 1 year) hospitalized for bronchiolitis, by using routinely-available pre-hospitalization data as predictors, we developed four machine learning models: Lasso regression, elastic net regression, random forest, and gradient boosted decision tree. We compared their predictive performance-e.g., area-under-the-curve (AUC), sensitivity, specificity, and net benefit (decision curves)-using a cross-validation method, with that of the reference model. The outcomes were positive pressure ventilation use and intensive treatment (admission to intensive care unit and/or positive pressure ventilation use). Of 1,016 infants, 5.4% underwent positive pressure ventilation and 16.0% had intensive treatment. For the positive pressure ventilation outcome, machine learning models outperformed reference model (e.g., AUC 0.88 [95% CI 0.84-0.93] in gradient boosted decision tree vs 0.62 [95% CI 0.53-0.70] in reference model), with higher sensitivity (0.89 [95% CI 0.80-0.96] vs. 0.62 [95% CI 0.49-0.75]) and specificity (0.77 [95% CI 0.75-0.80] vs. 0.57 [95% CI 0.54-0.60]). The machine learning models also achieved a greater net benefit over ranges of clinical thresholds. Machine learning models consistently demonstrated a superior ability to predict acute severity and achieved greater net benefit.

Predicting Appropriate Hospital Admission of Emergency Department Patients with Bronchiolitis: Secondary Analysis

Background

In children below the age of 2 years, bronchiolitis is the most common reason for hospitalization. Each year in the United States, bronchiolitis causes 287,000 emergency department visits, 32%-40% of which result in hospitalization. Due to a lack of evidence and objective criteria for managing bronchiolitis, clinicians often make emergency department disposition decisions on hospitalization or discharge to home subjectively, leading to large practice variation. Our recent study provided the first operational definition of appropriate hospital admission for emergency department patients with bronchiolitis and showed that 6.08% of emergency department disposition decisions for bronchiolitis were inappropriate. An accurate model for predicting appropriate hospital admission can guide emergency department disposition decisions for bronchiolitis and improve outcomes, but has not been developed thus far.

Objective

The objective of this study was to develop a reasonably accurate model for predicting appropriate hospital admission.

Methods

Using Intermountain Healthcare data from 2011-2014, we developed the first machine learning classification model to predict appropriate hospital admission for emergency department patients with bronchiolitis.

Results

Our model achieved an accuracy of 90.66% (3242/3576, 95% CI: 89.68-91.64), a sensitivity of 92.09% (1083/1176, 95% CI: 90.33-93.56), a specificity of 89.96% (2159/2400, 95% CI: 88.69-91.17), and an area under the receiver operating characteristic curve of 0.960 (95% CI: 0.954-0.966). We identified possible improvements to the model to guide future research on this topic.

Conclusions

Our model has good accuracy for predicting appropriate hospital admission for emergency department patients with bronchiolitis. With further improvement, our model could serve as a foundation for building decision-support tools to guide disposition decisions for children with bronchiolitis presenting to emergency departments.

International Registered Report Identifier (IRRID)

RR2-10.2196/resprot.5155

Design of a Clinical Decision Support System for Predicting Erectile Dysfunction in Men Using NHIRD Dataset

Erectile dysfunction (ED) affects millions of men worldwide. Men with ED generally complain failure to attain or maintain an adequate erection during sexual activity. The prevalence of ED is strongly correlated with age, affecting about 40% of men at age 40 and nearly 70% at age 70. A variety of chronic diseases, including diabetes, ischemic heart disease, congestive heart failure, hypertension, depression, chronic renal failure, obstructive sleep apnea, prostate disease, gout, and sleep disorder, were reported to be associated with ED. In this study, data retrieved from a subset of the National Health Insurance Research Database of Taiwan were used for designing the clinical decision support system (CDSS) for predicting ED incidences in men. The positive cases were male patients aged 20-65 who were diagnosed with ED between January 2000 and December 2010 confirmed by at least three outpatient visits or at least one inpatient visit, while the negative cases were randomly selected from the database without a history of ED and were frequency (1:1), age, and index year matched with the ED patients. Data of a total of 2832 ED patients and 2832 non-ED patients, each consisting of 41 features including index age, 10 comorbidities, and 30 other comorbidity-related variables, were retrieved for designing the predictive models. Integrated genetic algorithm and support vector machine was adopted to design the CDSSs with two experiments of independent training and testing (ITT) conducted to verify their effectiveness. In the 1st ITT experiment, data extracted from January 2000 till December 2005 (61.51%, 1742 positive cases and 1742 negative cases) were used for training and validating and the data retrieved from January 2006 till December 2010 were used for testing (38.49%), whereas in the 2nd ITT experiment, data in the training set (77.78%) were extracted from January 2000 till Deceber 2007 and those in the testing set (22.22%) were retrieved afterward. Tenfold cross validation and three different objective functions were adopted for obtaining the optimal models with best predictive performance in the training phase. The testing results show that the CDSSs achieved a predictive performance with accuracy, sensitivity, specificity, g-mean, and area under ROC curve of 74.72%-76.65%, 72.33%-83.76%, 69.54%-77.10%, 0.7468-0.7632, and 0.766-0.817, respectively. In conclusion, the CDSSs designed based on cost-sensitive objective functions as well as salient comorbidity-related features achieve satisfactory predictive performance for predicting ED incidences.

Point of care diaphragm ultrasound in infants with bronchiolitis: A prospective study

BACKGROUND

Bronchiolitis is the most common reason for hospitalization of children worldwide. Many scoring systems have been developed to quantify respiratory distress and predict outcome, but none of them have been validated. We hypothesized that the ultrasound evaluation of the diaphragm could quantify respiratory distress and therefore we correlated the ultrasound diaphragm parameters with outcome.

METHODS

Prospective study of infants with bronchiolitis (1-12 months) evaluated in a pediatric emergency department. Ultrasonography examinations of the diaphragm was performed (diaphragm excursion [DE], inspiratory excursion [IS], inspiratory/expiratory relationship [I/E], and thickness at end-expiration [TEE] and at end-inspiration [TEI]; thickening fraction [TF]).

RESULTS

We evaluated 61 infants, 50.8 % males. Mean TF was 47% (IQR 28.6-64.7), mean I/E 0.47 (± 0.15), mean DE 10.39 ± 4 mm. There was a linear correlation between TF and oxygen saturation at first evaluation (P = 0.006, r = 0.392). All children with lower values of TF required HFNC and one of them required CPAP. A higher IS was associated with the future need of respiratory support during admission (P = 0.007). IS correlated with the hours of oxygen delivery needed (P = 0.032, r = 0.422). TEI (t = 3.701, P = 0.002) was found to be main predictor of hours of oxygen delivery needed.

CONCLUSION

This study described ultrasound diaphragmatic values of previously healthy infants with bronchiolitis. DE, IS, and TEI correlated with outcome. If confirmed in larger studies, bedside ultrasound semiology of the diaphragm can be a new objective tool for the evaluation and outcome prediction of infants with bronchiolitis.

Design of a Clinical Decision Support System for Fracture Prediction Using Imbalanced Dataset

More than 1 billion people suffer from chronic respiratory diseases worldwide, accounting for more than 4 million deaths annually. Inhaled corticosteroid is a popular medication for treating chronic respiratory diseases. Its side effects include decreased bone mineral density and osteoporosis. The aims of this study are to investigate the association of inhaled corticosteroids and fracture and to design a clinical support system for fracture prediction. The data of patients aged 20 years and older, who had visited healthcare centers and been prescribed with inhaled corticosteroids within 2002-2010, were retrieved from the National Health Insurance Research Database (NHIRD). After excluding patients diagnosed with hip fracture or vertebrate fractures before using inhaled corticosteroid, a total of 11645 patients receiving inhaled corticosteroid therapy were included for this study. Among them, 1134 (9.7%) were diagnosed with hip fracture or vertebrate fracture. The statistical results showed that demographic information, chronic respiratory diseases and comorbidities, and corticosteroid-related variables (cumulative dose, mean exposed daily dose, follow-up duration, and exposed duration) were significantly different between fracture and nonfracture patients. The clinical decision support systems (CDSSs) were designed with integrated genetic algorithm (GA) and support vector machine (SVM) by training and validating the models with balanced training sets obtained by random and cluster-based undersampling methods and testing with the imbalanced NHIRD dataset. Two different objective functions were adopted for obtaining optimal models with best predictive performance. The predictive performance of the CDSSs exhibits a sensitivity of 69.84-77.00% and an AUC of 0.7495-0.7590. It was concluded that long-term use of inhaled corticosteroids may induce osteoporosis and exhibit higher incidence of hip or vertebrate fractures. The accumulated dose of ICS and OCS therapies should be continuously monitored, especially for patients with older age and women after menopause, to prevent from exceeding the maximum dosage.

Progressive sampling-based Bayesian optimization for efficient and automatic machine learning model selection

PURPOSE

Machine learning is broadly used for clinical data analysis. Before training a model, a machine learning algorithm must be selected. Also, the values of one or more model parameters termed hyper-parameters must be set. Selecting algorithms and hyper-parameter values requires advanced machine learning knowledge and many labor-intensive manual iterations. To lower the bar to machine learning, miscellaneous automatic selection methods for algorithms and/or hyper-parameter values have been proposed. Existing automatic selection methods are inefficient on large data sets. This poses a challenge for using machine learning in the clinical big data era.

METHODS

To address the challenge, this paper presents progressive sampling-based Bayesian optimization, an efficient and automatic selection method for both algorithms and hyper-parameter values.

RESULTS

We report an implementation of the method. We show that compared to a state of the art automatic selection method, our method can significantly reduce search time, classification error rate, and standard deviation of error rate due to randomization.

CONCLUSIONS

This is major progress towards enabling fast turnaround in identifying high-quality solutions required by many machine learning-based clinical data analysis tasks.

The modified respiratory index score (RIS) guides resource allocation in acute bronchiolitis

BACKGROUND AND OBJECTIVE

Bronchiolitis is a common disease in early childhood with increasing healthcare utilization. We aim to study how well a simple and improved respiratory score (the modified Respiratory Index Score [RIS]) would perform when predicting for a warranted admission.

METHODS

This is an observational prospective study, from June 2015 to December 2015 in a paediatric emergency department (ED) of a large tertiary hospital in Singapore. We included children aged less than 2 years old, presenting with typical symptoms and signs of bronchiolitis but excluded children with four or more previous wheezes, a gestation of <35 weeks, and known cardiopulmonary disease. We also performed a sensitivity analysis for children presenting with their first wheeze. We defined a warranted admission as a composite of: The need for airway intervention, intravenous hydration, and a hospital stay of 2 days or more.

RESULTS

Among 1,818 patients, the median age was 10.8 months (IQR 7.2-15.9). The median modified RIS score was 4.0 (IQR 3.0-5.0). A total of 19 (1.0%) children required respiratory support, 101 (5.6%) received intravenous hydration, and 571 (31.4%) required a hospital stay of 2 days or more. After adjusting for age and duration of illness, a modified RIS score of >4 predicted significantly for a warranted admission (adjusted Odds Ratio: 3.28, 95% confidence interval: 2.62-4.12). The association remained significant among children presenting with their first wheeze.

CONCLUSIONS

This simple respiratory tool predicts for the need for respiratory support, intravenous hydration, and a significant hospital stay of 2 days or more. Pediatr Pulmonol. 2017; 52:954-961. © 2017 Wiley Periodicals, Inc.

PredicT-ML: a tool for automating machine learning model building with big clinical data

Background

Predictive modeling is fundamental to transforming large clinical data sets, or “big clinical data,” into actionable knowledge for various healthcare applications. Machine learning is a major predictive modeling approach, but two barriers make its use in healthcare challenging. First, a machine learning tool user must choose an algorithm and assign one or more model parameters called hyper-parameters before model training. The algorithm and hyper-parameter values used typically impact model accuracy by over 40 %, but their selection requires many labor-intensive manual iterations that can be difficult even for computer scientists. Second, many clinical attributes are repeatedly recorded over time, requiring temporal aggregation before predictive modeling can be performed. Many labor-intensive manual iterations are required to identify a good pair of aggregation period and operator for each clinical attribute. Both barriers result in time and human resource bottlenecks, and preclude healthcare administrators and researchers from asking a series of what-if questions when probing opportunities to use predictive models to improve outcomes and reduce costs.

Methods

This paper describes our design of and vision for PredicT-ML (prediction tool using machine learning), a software system that aims to overcome these barriers and automate machine learning model building with big clinical data.

Results

The paper presents the detailed design of PredicT-ML.

Conclusions

PredicT-ML will open the use of big clinical data to thousands of healthcare administrators and researchers and increase the ability to advance clinical research and improve healthcare.

Incidence of Hospitalization for Respiratory Syncytial Virus Infection amongst Children in Ontario, Canada: A Population-Based Study Using Validated Health Administrative Data

Importance

RSV is a common illness among young children that causes significant morbidity and health care costs.

Objective

Routinely collected health administrative data can be used to track disease incidence, explore risk factors and conduct health services research. Due to potential for misclassification bias, the accuracy of data-elements should be validated prior to use. The objectives of this study were to validate an algorithm to accurately identify pediatric cases of hospitalized respiratory syncytial virus (RSV) from within Ontario’s health administrative data, estimate annual incidence of hospitalization due to RSV and report the prevalence of major risk factors within hospitalized patients.

Study Design and Setting

A retrospective chart review was performed to establish a reference-standard cohort of children from the Ottawa region admitted to the Children’s Hospital of Eastern Ontario (CHEO) for RSV-related disease in 2010 and 2011. Chart review data was linked to Ontario’s administrative data and used to evaluate the diagnostic accuracy of algorithms of RSV-related ICD-10 codes within provincial hospitalization and emergency department databases. Age- and sex-standardized incidence was calculated over time, with trends in incidence assessed using Poisson regression.

Results

From a total of 1411 admissions, chart review identified 327 children hospitalized for laboratory confirmed RSV-related disease. Following linkage to administrative data and restriction to first admissions, there were 289 RSV patients in the reference-standard cohort. The best algorithm, based on hospitalization data, resulted in sensitivity 97.9% (95%CI: 95.5–99.2%), specificity 99.6% (95%CI: 98.2–99.8%), PPV 96.9% (95%CI: 94.2–98.6%), NPV 99.4% (95%CI: 99.4–99.9%). Incidence of hospitalized RSV in Ontario from 2005–2012 was 10.2 per 1000 children under 1 year and 4.8 per 1000 children aged 1 to 3 years. During the surveillance period, there was no identifiable increasing or decreasing linear trend in the incidence of hospitalized RSV, hospital length of stay and PICU admission rates. Among the Ontario RSV cohort, 16.3% had one or more major risk factors, with a decreasing trend observed over time.

Conclusion

Children hospitalized for RSV-related disease can be accurately identified within population-based health administrative data. RSV is a major public health concern and incidence has not changed over time, suggesting a lack of progress in prevention.

Predicting Appropriate Admission of Bronchiolitis Patients in the Emergency Department: Rationale and Methods

BACKGROUND

In young children, bronchiolitis is the most common illness resulting in hospitalization. For children less than age 2, bronchiolitis incurs an annual total inpatient cost of $1.73 billion. Each year in the United States, 287,000 emergency department (ED) visits occur because of bronchiolitis, with a hospital admission rate of 32%-40%. Due to a lack of evidence and objective criteria for managing bronchiolitis, ED disposition decisions (hospital admission or discharge to home) are often made subjectively, resulting in significant practice variation. Studies reviewing admission need suggest that up to 29% of admissions from the ED are unnecessary. About 6% of ED discharges for bronchiolitis result in ED returns with admission. These inappropriate dispositions waste limited health care resources, increase patient and parental distress, expose patients to iatrogenic risks, and worsen outcomes. Existing clinical guidelines for bronchiolitis offer limited improvement in patient outcomes. Methodological shortcomings include that the guidelines provide no specific thresholds for ED decisions to admit or to discharge, have an insufficient level of detail, and do not account for differences in patient and illness characteristics including co-morbidities. Predictive models are frequently used to complement clinical guidelines, reduce practice variation, and improve clinicians' decision making. Used in real time, predictive models can present objective criteria supported by historical data for an individualized disease management plan and guide admission decisions. However, existing predictive models for ED patients with bronchiolitis have limitations, including low accuracy and the assumption that the actual ED disposition decision was appropriate. To date, no operational definition of appropriate admission exists. No model has been built based on appropriate admissions, which include both actual admissions that were necessary and actual ED discharges that were unsafe.

OBJECTIVE

The goal of this study is to develop a predictive model to guide appropriate hospital admission for ED patients with bronchiolitis.

METHODS

This study will: (1) develop an operational definition of appropriate hospital admission for ED patients with bronchiolitis, (2) develop and test the accuracy of a new model to predict appropriate hospital admission for an ED patient with bronchiolitis, and (3) conduct simulations to estimate the impact of using the model on bronchiolitis outcomes.

RESULTS

We are currently extracting administrative and clinical data from the enterprise data warehouse of an integrated health care system. Our goal is to finish this study by the end of 2019.

CONCLUSIONS

This study will produce a new predictive model that can be operationalized to guide and improve disposition decisions for ED patients with bronchiolitis. Broad use of the model would reduce iatrogenic risk, patient and parental distress, health care use, and costs and improve outcomes for bronchiolitis patients.

A systematic review of predictive models for asthma development in children

BACKGROUND

Asthma is the most common pediatric chronic disease affecting 9.6 % of American children. Delay in asthma diagnosis is prevalent, resulting in suboptimal asthma management. To help avoid delay in asthma diagnosis and advance asthma prevention research, researchers have proposed various models to predict asthma development in children. This paper reviews these models.

METHODS

A systematic review was conducted through searching in PubMed, EMBASE, CINAHL, Scopus, the Cochrane Library, the ACM Digital Library, IEEE Xplore, and OpenGrey up to June 3, 2015. The literature on predictive models for asthma development in children was retrieved, with search results limited to human subjects and children (birth to 18 years). Two independent reviewers screened the literature, performed data extraction, and assessed article quality.

RESULTS

The literature search returned 13,101 references in total. After manual review, 32 of these references were determined to be relevant and are discussed in the paper. We identify several limitations of existing predictive models for asthma development in children, and provide preliminary thoughts on how to address these limitations.

CONCLUSIONS

Existing predictive models for asthma development in children have inadequate accuracy. Efforts to improve these models' performance are needed, but are limited by a lack of a gold standard for asthma development in children.

MLBCD: a machine learning tool for big clinical data

Background

Predictive modeling is fundamental for extracting value from large clinical data sets, or “big clinical data,” advancing clinical research, and improving healthcare. Machine learning is a powerful approach to predictive modeling. Two factors make machine learning challenging for healthcare researchers. First, before training a machine learning model, the values of one or more model parameters called hyper-parameters must typically be specified. Due to their inexperience with machine learning, it is hard for healthcare researchers to choose an appropriate algorithm and hyper-parameter values. Second, many clinical data are stored in a special format. These data must be iteratively transformed into the relational table format before conducting predictive modeling. This transformation is time-consuming and requires computing expertise.

Methods

This paper presents our vision for and design of MLBCD (Machine Learning for Big Clinical Data), a new software system aiming to address these challenges and facilitate building machine learning predictive models using big clinical data.

Results

The paper describes MLBCD’s design in detail.

Conclusions

By making machine learning accessible to healthcare researchers, MLBCD will open the use of big clinical data and increase the ability to foster biomedical discovery and improve care.

Biomarkers of respiratory syncytial virus (RSV) infection: specific neutrophil and cytokine levels provide increased accuracy in predicting disease severity

Despite fundamental advances in the research on respiratory syncytial virus (RSV) since its initial identification almost 60 years ago, recurring failures in developing vaccines and pharmacologic strategies effective in controlling the infection have allowed RSV to become a leading cause of global infant morbidity and mortality. Indeed, the burden of this infection on families and health care organizations worldwide continues to escalate and its financial costs are growing. Furthermore, strong epidemiologic evidence indicates that early-life lower respiratory tract infections caused by RSV lead to the development of recurrent wheezing and childhood asthma. While some progress has been made in the identification of reliable biomarkers for RSV bronchiolitis, a "one size fits all" biomarker capable of accurately and consistently predicting disease severity and post-acute outcomes has yet to be discovered. Therefore, it is of great importance on a global scale to identify useful biomarkers for this infection that will allow pediatricians to cost-effectively predict the clinical course of the disease, as well as monitor the efficacy of new therapeutic strategies.

From bronchiolitis guideline to practice: A critical care perspective

Acute viral bronchiolitis is a leading cause of admission to pediatric intensive care units, but research on the care of these critically ill infants has been limited. Pathology of viral bronchiolitis revealed respiratory obstruction due to intraluminal debris and edema of the airways and vasculature. This and clinical evidence suggest that airway clearance interventions such as hypertonic saline nebulizers and pulmonary toilet devices may be of benefit, particularly in situations of atelectasis associated with bronchiolitis. Research to distinguish an underlying asthma predisposition in wheezing infants with viral bronchiolitis may one day lead to guidance on when to trial bronchodilator therapy. Considering the paucity of critical care research in pediatric viral bronchiolitis, intensive care practitioners must substantially rely on individualization of therapies based on bedside clinical assessments. However, with the introduction of new diagnostic and respiratory technologies, our ability to support critically ill infants with acute viral bronchiolitis will continue to advance.