Abstract:
The best way to prevent cervical cancer is through early detection of pre-cancerous lesions by regular screening to detect the presence of the Human Papilloma Virus (HPV), the cancer causing virus in the cervix of a woman. Although some diagnostic expert systems have been designed to detect the presence of pre-cancerous lesions in women, most are designed to work with Pap smear, a common method of screening for cervical cancer in more developed countries as against Visual Inspection with Acetic acid (VIA) that is more common in developing countries like Nigeria which is the major reason for this study. The system called Cervipert was developed using Adobe Dreamweaver as the programming environment, HTML, JavaScript, CSS as the front end while PHP and MySQL were used for the backend and Apache was used as the server. The inference engine which is the brain of Cervipert was implemented in PHP. The system was built to detect the presence of HPV in the cervix, recommend advice and or treatment for patients based on clinical symptoms inputted by user after which VIA test input have been done on patients. It was developed using the Dokas approach, a standard approach for developing expert systems and the ISO 13407 recommendations for developing user centered systems. The system was developed based on medical expert information collected through interviews and review of several literatures found online. Black box testing was done on Cervipert to ensure the system was doing what it was meant to do and manual debugging was done on the codes to remove excess and redundant codes. The system was tested by health workers using 20 patients that have previously been screened manually and it was found to be 90 % accurate. The developed software can be used as a tool to support screening programs for cervical cancer by health institutions and Non-governmental Organizations (NGO) and thereby reducing the high prevalence of this disease.
