Developing Mobile Knowledge BaseSystem prototype for selected chronicdisease (diabetes, hypertension, HIV)

Abstract

Nowadays, many different types of health problems do exist in the environment that we are livingin. Among those, some of them are live with patients throughout their lifetime. Diabetes,hypertensions, HIV, Cardiac, Cancer, etc. are some of them. These types of diseases needcontinuous healthcare services and lifelong physician consultation.Ethiopia is a one of the developing country in which the vast majority of the population experienceextremely limited access to medical care services, medical institutions and health extensionprograms especially chronic disease patients which needs continuous healthcare service andlifelong consultation.These days information technology is growing rapidly and is being apply in different areas. Thisresults in increase in productivity, quality of service and improved efficiency and effectiveness ofthe organization.Mobile phone knowledge base system is an intelligent software that deployed or accessed throughmobile devices, and it provide a knowledge base system services for clients with poor medicalfacilities. Like lack of experienced medical personnel and people having insufficient income to getproper medical attention and care.As a result, objective of the study was to investigate the applicability of rule-based reasoningapproach in the development of mobile knowledge-based system for selected chronic disease thatcan simulate the services of medical experts.To achieve this objective, knowledge is acquired using both structured and unstructured extensiveinterviews with domain experts, which are selected purposively from Tikur Anbessa (Black Lion)Specialized Hospital and Adama Hospital and Medical College. Additionally, knowledge isacquired from secondary sources by using document analysis method of knowledge elicitation.The knowledge acquired is modelled using decision tree structure and decision table that representssign and symptom, complication and procedures involved in disease diagnoses. Based on themodel, the prototype is developed with JESS which is a rule-based engine with scriptingenvironment and can be used to build Java software that has the capacity to “reason” usingknowledge that has been supplied in the form of declarative rules. JESS uses an enhanced versionof the Rete algorithm to process rules.The outcomes of this study are an Amharic web application and two Mobile applicationsprototypes in which the first is an android application to deploy on smart phones and the other is ajava application that can deployed under any non-smart phones with a java virtual machine. ThoseIVapplications enable to provide widespread knowledge base system services for clients through theirmobile phones or their computer.Prototype evaluation with domain experts shows 88.2% user acceptance, which is very good basedon per defined evaluation standards or measurement criteria. In addition, the performance of thesystem is evaluated by using predictive validation techniques with fifty-six test cases. The resultsof the validation test case indicate that the prototype is about 85.7% accurate.The prototype knowledge-based system needs further studies to expand its scope and to enhancethe performance of it by integrating with other knowledge representation techniques.