Dataset Information

Discovering Pediatric Asthma Phenotypes on the Basis of Response to Controller Medication Using Machine Learning.

ABSTRACT:

Rationale

Pediatric asthma has variable underlying inflammation and symptom control. Approaches to addressing this heterogeneity, such as clustering methods to find phenotypes and predict outcomes, have been investigated. However, clustering based on the relationship between treatment and clinical outcome has not been performed, and machine learning approaches for long-term outcome prediction in pediatric asthma have not been studied in depth.

Objectives

Our objectives were to use our novel machine learning algorithm, predictor pursuit (PP), to discover pediatric asthma phenotypes on the basis of asthma control in response to controller medications, to predict longitudinal asthma control among children with asthma, and to identify features associated with asthma control within each discovered pediatric phenotype.

Methods

We applied PP to the Childhood Asthma Management Program study data (n = 1,019) to discover phenotypes on the basis of asthma control between assigned controller therapy groups (budesonide vs. nedocromil). We confirmed PP's ability to discover phenotypes using the Asthma Clinical Research Network/Childhood Asthma Research and Education network data. We next predicted children's asthma control over time and compared PP's performance with that of traditional prediction methods. Last, we identified clinical features most correlated with asthma control in the discovered phenotypes.

Results

Four phenotypes were discovered in both datasets: allergic not obese (A⁺/O^-), obese not allergic (A^-/O⁺), allergic and obese (A⁺/O⁺), and not allergic not obese (A^-/O^-). Of the children with well-controlled asthma in the Childhood Asthma Management Program dataset, we found more nonobese children treated with budesonide than with nedocromil (P = 0.015) and more obese children treated with nedocromil than with budesonide (P = 0.008). Within the obese group, more A⁺/O⁺ children's asthma was well controlled with nedocromil than with budesonide (P = 0.022) or with placebo (P = 0.011). The PP algorithm performed significantly better (P < 0.001) than traditional machine learning algorithms for both short- and long-term asthma control prediction. Asthma control and bronchodilator response were the features most predictive of short-term asthma control, regardless of type of controller medication or phenotype. Bronchodilator response and serum eosinophils were the most predictive features of asthma control, regardless of type of controller medication or phenotype.

Conclusions

Advanced statistical machine learning approaches can be powerful tools for discovery of phenotypes based on treatment response and can aid in asthma control prediction in complex medical conditions such as asthma.

SUBMITTER: Ross MK

PROVIDER: S-EPMC5822415 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Despite previous findings of therapeutic effects for heart rate variability biofeedback (HRVB) on asthma, it is not known whether HRVB can substitute either for controller or rescue medication, or whether it affects airway inflammation. Sixty-eight paid volunteer steroid naïve study participants with mild or moderate asthma were given 3 months of HRVB or a comparison condition consisting of EEG alpha biofeedback with relaxing music and relaxed paced breathing (EEG+), in a two-center trial. All participants received a month of intensive asthma education prior to randomization. Both treatment conditions produced similar significant improvements on the methacholine challenge test (MCT), asthma symptoms, and asthma quality of life (AQOL). MCT effects were of similar size to those of enhanced placebo procedures reported elsewhere, and were 65% of those of a course of a high-potency inhaled steroid budesonide given to a sub-group of participants following biofeedback training. Exhaled nitric oxide decreased significantly only in the HRVB group, 81% of the budesonide effect, but with no significant differences between groups. Participants reported becoming more relaxed during practice of both techniques. Administration of albuterol after biofeedback sessions produced a large improvement in pulmonary function test results, indicating that neither treatment normalized pulmonary function as a potent controller medication would have done. Impulse oscillometry showed increased upper airway (vocal cord) resistance during biofeedback periods in both groups. These data suggest that HRVB should not be considered an alternative to asthma controller medications (e.g., inhaled steroids), although both biofeedback conditions produced some beneficial effects, warranting further research, and suggesting potential complementary effects. Various hypotheses are presented to explain why HRVB effects on asthma appeared smaller in this study than in earlier studies. Clinical Trial Registration NCT02766374.