Disease management
Medication non-adherence in allergic diseases is common in clinical
practice and can negatively impact disease control. To address this
issue, researchers explored ML approaches for disease management and
medication adherence. One such approach involves using ML to provide
early warnings for loss of control in the Asthma Mobile Health Study
data of 5,875 patients, containing over 75,000 daily surveys on symptoms
and medicine use, medical history, demographics, location and EuroQol 5D
questionnaire. The supervised classifier obtained an AUC of 0.87, but
peak flow readings did not further enhance its performance. External or
prospective validation is strongly needed.
In addition to early warning systems, chatbots have also been proposed
to support disease management by providing personalized advice to
patients and tracking medication compliance. One example is KBot, an
early prototype of a chatbot for asthma that utilizes contextual
information (such as high pollen triggers) and NLP for dialogue
processing. AI can also leverage the capabilities of wearables and
mHealth technologies to monitor disease outside clinical contexts. A
recent study tested a prototype application for real-time counting of
coughs using a deep learning model on ambient sound recorded by mobile
phone . This yielded accurate and real-time cough count with a
specificity of 92% and a specificity of 98%. Another study applied ML
to analyze the sounds of asthma inhalers to predict adequate usage and
drug actuations. Recorded sound on mobile devices has also been proposed
to monitor lung function in asthmatics. While requiring further
validation, these techniques could be used to develop future telehealth
solutions including smartphone-based applications, which have the
potential to aid decision-making and self-monitoring in asthma.
Fundamental research
AI can provide insights into disease classification, pathophysiology,
and the underlying biological mechanisms, by clustering large numbers of
data points into interpretable patterns.