|Objectives||The objectives of this session is to share the basic concept and detailed design and implementation of Intelligent Medical Platform that is a platform to support services on top of the incremental learning based on evolutionary knowledge bases from diverse sources by integrating legacy systems with multiple data formats, supported by clinical standard-inspired big data persistence. The system is fitted out with dialogue environment in the form of text, voice, and image to physicians and patients during treatment, diagnosis and health related recommendations, alerts, coaching, and education. It overcomes the existing systems limitations in terms of evolutionary knowledge base, dialoging ability of interaction, integration with legacy healthcare systems such as hospital information and management systems (HMIS), and handling multimodal data from diverse input sources|
|1. ||Silo of Intelligent Medical Platform Slides (PDF)|
The care of patients with different diseases (heart failure, kidney, eye, diabetes, ENT) is highly affected by miss-interpretations in decision making during the diagnosing, treatment, follow-up, and poor medication adherence. An intelligent clinical decision support system (CDSS) is needed to integrate with centralized EMR systems of the medical institutions to assist the physicians in decision making. The intelligent decisions are highly dependent on the evolutionary knowledge base of the CDSS. The existing clinical data, physician’s heuristics, experiences, and practices are considered as contributing resources to the knowledge base evolution. However, the knowledge base is different for each disease but sometimes interrelated into a unified knowledge base. Therefore, we proposed a concept of Silo to manage the knowledge bases of different diseases. Silo is a structure for storing bulk materials, therefore, the knowledge of different diseases in a bulk storage with unified format is called Silo. It provides the decision making services to assist the physicians in diagnosis, treatment, and follow-ups phases of patient care. The Silo construction process of all disease have five steps to construct the knowledge using requirements elicitation. The steps of Silo construction are knowledge acquisition, knowledge modeling, rules generation, implementation, and evaluation. The first three steps are knowledge base construction, while the fourth step is the development of decision making executable system. In the evaluation phase, we evaluate the system in the real environment of the hospitals.
|2. ||Interoperability of Medical System Slides (PDF)|
Wajahat Ali Khan
Over the last decade, rapid digitization in the field of healthcare information management, has compounded the problem of Heterogeneity, through the development of a plethora of medical software, devices, and standards. As a consequence, while the quantity of healthcare data has undergone rapid expansion the quality of care is still lacking. This has resulted in an overall increase in the cost of diagnostics, treatment and follow-up. In particular, medical data interoperability has proved to be a difficult obstacle, in moving towards an intelligent medical platform. Using a mediation based semantic reconciliation strategy, the Ubiquitous Health Platform (UHP) provides a solution to this problem. This platform, in turn, uses a multi-dimensional data container, in a semistructured data storage and processing engine; The Ubiquitous Health Profile (UHPr). With the ability to extract semantic value from a large volume of patient data, produced by a variety of data sources, at variable rates and conforming to different abstraction levels, the UHPr provides an intelligent, patientoriented, healthcare data management solution. Additionally, the UHP and UHPr is able to create and share, a comprehensive digital persona of patients, suffering from cardiovascular disease, using medical data from heterogeneous EMR, PHR (Medical IoT), and EHR sources.
|3. ||Medical Evidence Support System(Knowledge Button) Slides (PDF)|
There is an exponential growth in the medical literature, and medical practitioners are finding it difficult to obtain the most relevant information in their limited time span. Young physicians, particularly, are open to innovation, but they seek to minimize the costs either to themselves or to their patients. Without automation, processing of huge amount of literature is a costly and a challenging task for even the existing systems, leave aside manual efforts of the medical practitioners. Physicians, whether serving individual patients or populations, always have sought to base their decisions and actions on the best possible evidence. Based on evidence-adaptive clinical decision support systems, the researchers and developers need to customize the literature-based evidence for local conditions. Adhering to these needs and recommendations, the task of finding best possible evidence from the literature, customized to the local conditions, becomes a priority. We aim at developing a medical evidence support system for the automation of finding quality (best possible) evidences from a plethora of literature documents and rank them according the context (local condition). Our proposed system is called Knowledge Button and has three salient features: autonomous query construction, quality of articles recognition, and ranking & summarization of the articles to form a clinical evidence to be useful in the clinical practice.
|1. ||Overview of Intelligent Medical Platform Project Slides (PDF)|
This research materializes the convergence of the medical decision making with dialogue-based technologies. We propose Intelligent Medical Platform (IMP) to nurture medical coaching and recommendation services on top of the incremental learning methodology. The system is equipped with dialogue based environment that facilitates the users (physicians and patients) to interact with system in the form of free text, voice, and image. The services include recommendations, alerts, coaching, and education during treatment, diagnosis and health related self-management activities. It overcomes the existing systems limitations in terms of self-evolvable knowledge, ability to interact, integration with legacy healthcare systems, and handling multimodal data from diverse input sources. We demonstrated results of IMP’s prototype version with respect to knowledge acquisition, user interaction with system, and integration with legacy systems.
|2. ||Discussions on IMP projects and Solutions Slides (PDF)|