Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/85031
Title: A clinical decision support system for medical prescription process
Authors: Ting, Siu-lun
Degree: Ph.D.
Issue Date: 2011
Abstract: Nowadays enormous amounts of new drugs are developed and launched onto the market. With the increased complexity of drug information and a degree of uncertainty surrounding it, medicine prescription has become a vexing process. This is particularly true from the general practitioners' perspective as they are primarily responsible for providing general health care to patients, and keeping abreast with the latest drug information for treatment of mutating diseases. Many researchers have proposed the development of a decision support system for addressing these issues in recent years. However, the existing approaches lack flexibility to deal with the complex and dynamic changes in drug information. Thus, providing support to the medical prescription process remains a major challenge to be adequately addressed. The main problems are the development of adaptive mechanisms of knowledge acquisition and knowledge modeling in the medical prescription process. This research proposes a new approach to addressing the above mentioned difficulties. Prescription practices differ from one physician to another. This leads to the existence of a large variety of methods, giving rise to considerable complexity in modeling the physician's decision logic. The proposed approach to modeling physicians' prescription logic takes into account both individual and collective wisdoms of a pool of physicians. It has been implemented in a Medical Prescription Decision Support System (MedicPDSS) that automatically generates drug suggestions after considering the medical information of a specific patient. The approach integrates computational intelligence and data mining techniques to model physicians' prescription behaviors, from which suggested options of safe medical prescriptions are generated. This system employs case-based reasoning to retrieve past prescription records that are related to treatment of the disease under consideration, and the association rules mining technique is then applied to integrate such results. The former step models the practices of individual physicians, and the latter one models their collective judgments. The suggested options of medical prescriptions are generated accordingly along with rankings of their appropriateness. Furthermore, the prescription selected by the physician is checked with data obtained by an automatic information retrieval engine to detect drug-drug interactions. A prototype system that applies this approach has been built and tested in a medical organization Humphrey and Partners Medical Services Limited (HPMS). With the implementation of the MedicPDSS in HPMS, a number of potential benefits are realized. The system produces significant improvement in a number of performance criteria, such as reducing the time for deciding on a prescription, minimizing prescription errors, optimizing the list of suggested drugs, and enhancing knowledge sharing of prescription practices. In addition, the physicians and nurses confirmed that MedicPDSS has helped them to improve quality of prescriptions. Even though the proposed system has been developed for a typical healthcare environment, its concept can be customized for other applications in healthcare environments that deal with specific diseases. It can also be used together with other computer-aided diagnostic systems and clinical decision support systems to further enhance the medical prescription process. Moreover, the system can be employed to support training for medical professionals.
Subjects: Drugs -- Prescribing.
Drugs -- Prescribing -- Data processing.
Drugs -- Prescribing -- Computer programs.
Hong Kong Polytechnic University -- Dissertations
Pages: xxi, 225 leaves : ill. ; 30 cm.
Appears in Collections:Thesis

Show full item record

Page views

50
Last Week
0
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.