Leveraging Machine Learning For Personalized Treatment Plans In Respiratory Disorders

Abstract

Background:The prospect for machine learning (ML) to be incorporated into healthcare systems is impressive and will help tailor treatment approaches to individuals, especially patients with asthma, chronic obstructive pulmonary disease (COPD), and interstitial lung disease (ILD). However, the real-world embedding and uptake of ML in clinical practice is still not well understood.

Objectives:The present study consists of the aim of deploying ML for the personalization of the treatment plan for respiratory diseases and measuring the satisfaction of healthcare providers, data science specialists, and patients. It also aims to determine important drivers of satisfaction and acceptance of ML-based solutions.

Methods:A structured, quantitative methodology was adopted using a questionnaire sent out to medical practitioners, data analytic professionals, and patients. Descriptive phrases and inferential statistical methods including correlation, regression, and testing of hypotheses were utilized to explain the collected data. Cronbach’s Alpha was used to test the internal consistency of the respective study instruments, while the Shapiro-Wilk test was used to test the normality of the data collected.

Results:Due to the result of the Shapiro-Wilk test, showed that the sample data of satisfaction with ML in respiratory care has not been normally distributed (p < 0.05). Survey items’ internal consistency estimate was measured by Cronbach’s Alpha which yielded a score of 0.561. Correlation analysis of data indicated low to moderate relationships of key variables among each other. It was observed in the case of the boxplot that the level of satisfaction varied among different professional groups, where healthcare professionals were less satisfied with ML interventions than data science professionals.

Conclusions:The significant potential of AI in personalizing the treatment of respiratory disorders is there, but the barriers to its adoption specifically among healthcare professionals who showed less satisfaction with AI interventions remain. Development of the measurement tool used in the study, in future studies, should help overcome the reliability while differences in the level of satisfaction highlight several aspects that affect the perception of ML in respiratory care. Education as well as communication between healthcare professionals and data scientists still has to improve to build trust and make ML optimally used in clinics