A holistic health monitoring framework supported by a novel diagnostic algorithm

Abstract

In the present era, technology has greatly influenced the field of medicine and healthcare. Recent innovations in wireless transmission and biosensor technology have driven the concept of potential convergences between healthcare and telecommunications. The emergent use of telemedicine technologies for remote monitoring of patients with chronic disease has enabled clinicians to manage patients remotely and in a pro-active manner with a large number of healthcare organizations and hospitals trying to implement various remote monitoring healthcare applications. Currently there are many applications available for research purposes as well as for commercial use from industry.

Many of the industrial or commercial applications, implemented by healthcare organizations, may include wireless sensors, biometric wristwatches, wireless ECG systems, mobile cardiac telemetry systems, blood pressure and glucose meters, etc. Such industrial applications aid the medical doctors in monitoring the daily activities and healthcare status of the patients considerably by utilizing the biosensors and networking technologies. Healthcare solutions and platforms vary by their purpose and features and there are no efficient classifications combining the common or distinct features among these solutions.

This thesis aims to improve the state of the art by introducing a comprehensive framework that has covered most of the required features for monitoring patients (remotely) clustered in a modular structure, which makes it flexible and scalable by adding or removing further modules or features without affecting the other modules or interrupting the platform’s core operations. This modular framework employs specific functions for each module to eliminate the redundancy of the tasks or the potential overlap between them and therefore reduces cost, time and effort. Having a framework such as that proposed here will allow researchers and developers to focus more on the knowledge intrinsic to the patient-relevant data being collected and analysed as opposed to technical developments and specific programming details.

It also provides a guideline on what are the key features in the most common healthcare solutions and the drawbacks of each category. It also characterizes the existing healthcare monitoring solutions into two major categories, which include research prototypes and industrial applications. These features incorporating non-intrusive, security-enabled, mobile-aware, integration support and context-aware features. Similarly, it is envisioned there is a new concept of collecting medical knowledge from external databases, such as social networks, and utilise such information to support diagnosing decisions through expert systems as well as learning techniques.

Overall, this thesis presents a systematic procedure to be used for diagnosing various kinds of diseases through developing an algorithm that incorporates mathematical expressions to determine a variable called an “Indicator” that searches a look-up table of predefined medical conditions to predict the most likely disease the patient may be suffering. By designing this algorithm and implementing a software simulator, it was also found that the proposed diagnosing algorithm is much faster and more efficient over conventional search methods in calculating the Indicator and diagnosing the medical condition, being on average between 10% and 48% faster than sequential search methods and when considering more than 40 medical conditions (diseases), reached a 92.5% level of accuracy assuming there was no intersection with vital sign values or the Indicator’s range.

Date of Award	Jul 2017
Original language	English
Awarding Institution	Coventry University
Supervisor	Saad Amin (Supervisor), Adel Serhani (Supervisor) & Mahmoud Al Ahmad (Supervisor)