- Genetic Data Enhances Understanding of Asthma Development
- Machine Learning Tool Predicts Hospitalizations for Asthma Patients
They created the CHILDhood Asthma Risk Tool (CHART) to identify the asthma indications in patients who are 3 years old, such as the timing and number of wheeze or cough episodes, and use of inhaled corticosteroids, to identify children with asthma or ongoing symptoms at 5 years old.
Researchers used data from a total of 2,511 children for the study, 2,354 (93.7 percent) of whom had outcome data that was available at 5 years old.
When applied in the CHILD Study, among those who were 3 years old, CHART outperformed physician assessments as well as the mAPI at predicting wheeze, asthma diagnosis, and healthcare use.
When externally validated in the Raine Study, CHART had a similar predictive performance regarding persistent wheezing among children who were 5 years old. In the CAPPS study, the tool displayed similar performance in predicting wheezing at 7 years old.
Based on this data, researchers concluded that the CHART tool displayed the ability to identify asthma risk among children as young as 3 years old.
"CHART could be easily incorporated as a routine screening tool in primary care to identify children who need monitoring, timely symptom control, and introduction of preventive therapies," they wrote.
There have been several recent efforts focused on applying predictive analytics to enhance care practices.
For example, in September, Mayo Clinic created an artificial intelligence model that used patient data from the start of labor, the most recent clinical assessment, and cumulative labor progress from admission to determine labor risk predictions, leading to improved clinical decision-making.
In October, Penn State researchers announced that they plan to use a grant from the National Science Foundation to create artificial intelligence and machine-learning algorithms that will analyze longitudinal data to perform health risk predictions.
Further, a study published in June described clinical risk prediction models for sepsis, delirium, and acute kidney injury. According to researchers, these models achieved high performance when implemented into live clinical workflows.