Grant Amount: US$ 9,000

Keywords: HEALTH, INFORMATION MANAGEMENT, ARTIFICIAL INTELLIGENCE, MALAYSIA

Geographic location: Malaysia

Objective

The objective of this project is to provide advanced, quality health care information and services through the use of information and communications technology (ICT) tools and Artificial Intelligence (AI).

Research context

The aim of this project is to provide quality health care information and services for the Malaysian people, especially those in rural communities. The project employs ICTs and AI to develop and implement a web-based Medical Information Repository (MIR) integrated with a computerized medical Decision Support System (DSS). The novelty of the project lies in the inclusion of an AI-based DSS into the MIR. The system is able to learn incrementally in real-time, non-stationary environments with minimum intervention. With this autonomous learning capability, medical practitioners are able to train and fine-tune the DSS, and to assume ownership of the system.

Target beneficiaries

The project serves as a prototype system to demonstrate the significance of information sharing through a web-based repository to synergize activities in medical practice, training, and research using ICT and AI methodologies. It is envisaged that medical practitioners and researchers will be able to access anonymous medical records with heuristic diagnostic rules from the MIR through the Internet. The information obtained helps clinicians apply the most effective curative and rehabilitative regimes to enhance quality of health care of patients, especially for those in poor and remote areas where infrastructure and medical expertise are scarce. The MIR and DSS can be used as a resource, which contains up-to-date health care procedures and information, for continuing education and training of clinicians and medical practitioners. In addition, the project can be integrated into the telemedicine flagship application under the Multimedia Super Corridor project spearheaded by the Malaysian government.

Outputs

The main output of the project is a web-based MIR and DSS software comprising:

• Anonymous medical records of patients useful for medical practitioners and researchers including physical symptoms, family history and bio-chemical test results;
• Heuristic prognostic and diagnostic rules useful for junior and inexperienced clinicians elicited from medical specialists and the DSS; and
• Disease statistics and facts useful for health care administrators and policy makers.

Research results and outcomes

The project involves collection of anonymous patient data from participating hospitals and development of an AI-based decision support tool to assist medical practitioners in making accurate and timely diagnostic decisions. It is anticipated that medical doctors and health care workers, especially those in remote areas where specialized medical knowledge (e.g. stroke diagnosis) is difficult to come by, will use the system for validation of patient information as well as consultation for medical decision-making.

To date, the algorithm of the DSS has been developed using AI methodologies. Rigorous lab-based testing and evaluation of the algorithms has been conducted. A database of more than 1,000 anonymous patient records on acute stroke diagnosis has been collected for analysis. Currently the process of data cleansing to remove "noise" and missing information is being conducted. The project is working with a consultant neurologist who is provides feedback and information on the important symptoms of a disease and how to interpret the relevant diagnoses. Students at the postgraduate level (Master’s and PhD) are being trained in the development of the DSS. A specific focus of this training relates to algorithm deployment and state-of-the-art methodologies in AI, particularly in the areas of artificial neural networks and fuzzy systems.

The greatest challenge that this project faced so far is in the establishment of the medical information repository. The project has insisted on using real (anonymous) patient records. However, the medical data collected was often incomplete, making the analysis difficult and time-consuming. Therefore the project was extended by six months to allow for a longer period of data collection and validation, which will ensure their proper analysis and interpretation. This process is important, as the AI-based decision will be "trained" to recognize disease symptoms autonomously by using the database. Thus, a good quality medical database is a crucial factor in determining the effectiveness of the resulting DSS.

After the project team conducts a systematic evaluation of the performance of the AI-based algorithm, medical doctors enlisted to further evaluate the system. Suggestions and feedback received will be taken into consideration to improve the effectiveness of the system and finally a web-based implementation of the DSS will be conducted.

Duration

Start Date: February 2005
End Date: July 2006
Total Duration: 18 Months

Contact information

Dr. LIM Chee Peng, Associate Professor
School of Electrical and Electronic Engineering, Universiti Sains Malaysia
Engineering Campus, 14300 Nibong Tebal, Penang, Malaysia
Telephone: +60 4 599 6033
Fax: +60 4 594 1023
Email: cplim@eng.usm.my

Website: http://www.usm.my