Abstract:
The Internet of Medical Things plays an important role in the healthcare domain for
real-time monitoring of patients with high reliability and accuracy. According to the
WHO in Pakistan, 30% to 40% of deaths are caused due to cardiac attacks which are ap proximately 200,000 deaths per year. A comprehensive literature study is conducted to
explore, analyze and compare existing system architectures for cardiac health monitor ing worldwide. Our preliminary survey shows that very few e-health architectures exist
in Pakistan; therefore to address this issue, we proposed an digital health monitoring
system that is able to detect the onset of various health anomalies in the patient’s vitals,
using advanced machine learning algorithms and data visualization using web portal.
Thus, reducing the burden on hospitals by introducing remote monitoring facilities to
patients as well as doctors. The fundamental purpose of the proposed research is to
incorporate cutting-edge machine learning classification algorithms to detect anomoly
in human vitals such as heart rate (HR), blood pressure (BP), blood oxygen saturation,
body temperature, respiration rate etc. in near real-time. In our preliminary research,
we evaluated the performance of multiple ML algorithms trained on the clinical data
set. Random-Forest achieved the highest accuracy on the test set (95%) among the eight
tested supervised classification algorithms. In order to provide a remote patient man agement and monitoring panel, we created an web portal to ensure the confidentiality
and security of patient data within the proposed system.
