Abstract:
The increased dependency on information systems in all spheres of our lives has led to the exponential growth of electronic data. Therefore, more and more digital information is becoming available and readily accessible for decision making. However, the constantly growing data is difficult to use due to availability of too much information. In computing, this has popularly become known as information overload problem. In recent years, electronic health records have gain wide acceptance in many healthcare delivery settings which is changing the clinical health records documentation in medical practice. Physicians therefore routinely collect and document care of their patients using electronic health records and thus generating digital health records of patients during one or more clinical encounters which can be leveraged to improve healthcare delivery and clinical research. However, the abundant availability of detailed patient information has led to the problem of information overload which makes it difficult to look for information within the health record. It’s therefore difficult for users to find and utilize information to improve patient care. In clinical practice, many decisions require a doctor to read and summarize one or many clinical documents which are stored longitudinally, containing various medical assessments or observations per patient. But, there is always a limited time to read long and multiple documents and make decisions during care episodes. Users are therefore faced with an overwhelming amount of information about their patients which requires much time and effort to browse and review and make decisions under time pressure. Despite these deficiencies, visualization of textual documents has proposed as efficient technique for summarizing clinical documents so that it can be easily understood and consumed.
In this paper we review various text visualization techniques, their applications in summarizing medical documents. Further, we identify research gaps and give an overview of our work in progress to address the aforementioned problem and the gaps. Lastly, this work follows a generalized conclusion on health data visualization and references.
