Visualization of Medical Data Using Web Technologies

Abstract

Over the years, the medical expenses caused by inpatients have been increasing all around the world. One reason for this is increase in treatment costs. Some new medicines and methods may be cheaper, while some can cost significantly more. Unless hospital funding is increased accordingly, ways to save money must be found. One way to reduce the costs and free hospital resources is to discharge patient as soon as possible. Problem in this is how to do it without risking the patient’s health and possibly life by discharging her too early. After patients are discharged, they also are almost invisible to hospital systems, making it hard to monitor their recovery. Traditionally monitoring the recovery means patient visiting the hospital for checkup, or possibly counselling via phone. Checkups, however, can put strain on the patient especially if she lives further away from hospital, while through phone counselling it is hard to get full grasp on the situation. To solve these problems, patient remote monitoring solutions have been developed. Remote monitoring allows the patient to stay at home while a doctor or a nurse monitors her vital signs and other data at the hospital. In this thesis, a patient remote monitoring application for demonstration purposes was implemented using modern web technologies. The application was written mostly in JavaScript, using AngularJS framework, cascading style sheets and HTML5’s canvas element and server-sent events. Bootstrap CSS and JavaScript framework was also used to some extent. A generic Internet of Things cloud was used to store data and to retrieve it. Additionally, the data sent to the cloud is relayed to the application using server-sent events. For the evaluation, mocked sensor data was sent to the cloud back-end. The results were mostly positive. In extreme situation AngularJS begun to slow down and depending on the platform and setup, the CPU usage of the canvas rendering using the custom visualization library was considered too high. The end-to-end latency was mostly good, though because of occasional latency spikes, it cannot be used in critical situations where latency must be consistent. Overall, the application and the end-to-end system worked well. The work can be considered successful. Even though the application has some performance issues, some of them are very unlikely to occur in real usage. Therefore, it can be said that the application performed well and achieved its goal. It provides a good basis for further development and optimizations and proves that web technologies can be used in medical domain, possibly even in hospital wards.