According to the researchers, a patient with asthma endures approximately 2190 hours of experiencing and treating or not treating their asthma symptoms. During 15-minute clinic visits, only a short amount of time is spent understanding and treating what is a complex disease, and only a fraction of the necessary data is captured in the electronic health record.
“Our patients and the pace of data growth are compelling us to incorporate insights from Big Data to inform care,” the researchers posit. “Predictive analytics, using machine learning and artificial intelligence has revolutionized many industries,” including the healthcare industry.
When used effectively, big data, in conjunction with electronic health record data, can transform the patient’s healthcare experience. This is especially important as healthcare continues to embrace both e-health and telehealth practices. The data resulting from these “thoughtful digital health innovations” can result in personalized asthma management, improve timeliness of care, and capture objective measures of treatment response.
According to the researchers, the use of machine learning algorithms and AI to predict asthma exacerbations and patterns of healthcare utilization are within both technical and clinical reach. The ability to predict who is likely to experience an asthma attack, as well as when that attack may occur, will ultimately optimize healthcare resources and personalize patient management.
The use of longitudinal birth cohort studies and multicenter collaborations like the Severe Asthma Research Program have given clinical investigators a broader understanding of the pathophysiology, natural history, phenotypes, seasonality, genetics, epigenetics, and biomarkers of the disease. Machine learning and data-driven methods have utilized this data, often in the form of large datasets, to cluster patients into genetic, molecular, and immune phenotypes. These clusters have led to work in the genomics and pharmacogenomics fields that should ultimately lead to high-fidelity exacerbation predictions and the advent of true precision medicine.
“This work,” the researchers noted, “if translated into clinical practice can potentially link genetic traits to phenotypes that can for example predict rapid response, or non-response to medications like albuterol and steroids, or identify an individuals’ risk for cortisol suppression.”
As with any innovation, though, challenges abound. One in particular is the siloed nature of the clinical and scientific insights about asthma that have come to light in recent years. Although data are now being generated and interpreted across various domains, researchers must still contend with a lack of data standards and disease definitions, data interoperability and sharing difficulties, and concerns about data quality and fidelity.
Machine learning and AI present their own challenges; namely, those who utilize these technologies must consider the issues of fairness, bias, privacy, and medical bioethics. Legal accountability and medical responsibility issues must also be considered as algorithms are adopted into routine practice.
“We must, as clinicians and researchers, constructively transform the concern and lack of understanding many clinicians have about digital health, [machine learning], and [artificial intelligence] into educated and critical engagement,” the researchers concluded. “Our job … is to use [machine learning and artificial intelligence] tools to understand and predict how asthma affects patients and help us make decisions at the patient and population levels to treat it better.”
Messinger AI, Luo G, Deterding RR. The doctor will see you now: How machine learning and artificial intelligence can extend our understanding and treatment of asthma [published online December 25, 2019]. J Allergy Clin Immunol. doi: 10.1016/j.jaci.2019.12.898
social experiment by Livio Acerbo #greengroundit #thisisnotapost #thisisart