In Denmark, it is estimated that up to 55,000 people suffer from epilepsy. According to the Danish Epilepsy Association, 70 percent of epileptics become seizure free due to correct medication. However, the remaining 30 percent never become fully seizure free. Many epileptics get unpredictable seizures without any kind of warning, which can lead to physical damage and hospitalisation. It is estimated that epilepsy costs the Danish society nearly DKK 6 billion yearly. Apart from the direct cost of having epilepsy, the collected costs in regard to epilepsy are mainly composed of indirect social economic costs, such as public social benefits. The internet of things and big data analysis is expected to grow rapidly within health care. Our thesis is an experiment of how these technologies can be utilized to potentially improve the lives of epileptics. We explore this by collecting data about epileptics by using both their smartphone and a wearable, after which we create models by using big data analysis techniques. This serves our overall purpose of detecting factors affecting epileptic seizures as well as potentially predicting seizures. In our thesis, we created a solution to collect and analyse data about epileptic seizures. The solution consists of an iOS application, a Fitbit Charge HR, and a corresponding back-end. With this solution, we managed an 89 percent response rate of our daily reports for epileptics over the course of 1.5 months. This enabled us to create prediction models, which scored up to 59 percent correctness rate in predicting seizures. We further created a regression model with a p value > 0.0001, which potentially enables the Danish Epilepsy Association to gain new knowledge of what factors influence seizures. For visual data exploration, we were able to classify two seizures as predictable. Lastly, we show how personal needs and ethics are important considerations, when designing and building solutions for epileptics. Our thesis thereby contributes to both improving the lives of epileptics as well as to predict epileptic seizures. In the process of working towards improving the life of epileptics, we work within the design science framework. Apart from our technical solution and considerations when designing a solution for epileptics, we create a theoretical contribution by showing how it is possible to replace the existing elements from the IS research field in the design science framework with the CRISP model, while still complying with the design science guidelines.
|Educations||MSc in Business Administration and Information Systems, (Graduate Programme) Final Thesis|
|Number of pages||145|