A Mobile-Based Fuzzy System for Diagnosing Syphilis (Sexually Transmitted Disease)

Published Online December 2014 in MECS DOI: 10.5815/ijitcs.2015.01.04

Africa has been hypothetically found to be synonymous to diseases due to the high rate at which the continent is affected by various diseases. There are various deadly diseases that are faced by people residing in Africa but some of these are under medical control. Syphilis (sexually transmitted disease) caused by Treponema pallidum [1] is one of these diseases. It is a disease that is prevalent in lesser and more developed countries [2]. Although Syphilis is mostly referred to as a sexually transmitted disease (STD) but, it could be congenitally transmitted or transmitted through blood transfusion [3]. There are various reasons for the prevalence of this disease among which are ignorance, environmental factors, lack of adequate competent medical personnel who are capable of putting the disease under medical control, and insufficient preventive measures against the spread of the concerned ailment. Out of the previously mentioned reasons for syphilis prevalence, it has been found that the lack of competent medical personnel who are capable of providing

Medicare is the major cause of its prevalence. The annual increase in the population of syphilis’s patients against the competent medical personnel is not balanced and in fact, calls for immediate aid.

The innovations in information technology (IT) and the need for improvement on the current state of Medicare due to the unbalanced ratio of patients to medical personnel have brought up a research interest like artificial intelligence in medicine (AIM). This area is concerned with the application of IT concepts in providing a human like Medicare to patients in various forms. These forms include the development of expert systems that could be used in providing a human expertlike diagnosis or therapy to patients or even used as support systems for growing medical personnel. Telemedicine is another area that looks into the use of telecommunication technologies in providing medical care.

However, this paper considers the development of a fuzzy system that is capable of providing approximate diagnostic services for patients who suffer from Syphilis using uncertain symptoms as inputs. This type of diagnosis is a human like diagnosis found in the field of medicine. The target for the hardware platform for executing the concerned system has been diligently chosen due to the expected coverage scope of the system since it is believed that patients of the concerned disease may not have direct access to personal computers (PCs), and other non mobile computing devices. Therefore, the fuzzy system executes on readily available mobile cell phones of patients that are infected with the concerned ailment. Also, the fuzzy system has been developed as a standalone system that does not provide distributed service. This has made the accessibility of the system available at absolutely no cost since there is no need for any internet or network connection before it is accessed.

• II.    Literature Review

This section provides a systematic theoretical framework for the study and it is categorically considered below:

• A.    Overview of Syphilis

Syphilis is a systemic disease from the outset and is caused by the spirochaete, Treponema pallidum (T. pallidum) [1]. The infection can be classified as congenital or acquired (through sex or blood transfusion) [1]. Syphilis is a disease that has four stages: primary, secondary, latent and tertiary. The primary stage has been considered mild stage in this study based on its degree of severity. This stage has a symptom of sore called chancre that appears around anus, genital parts or around any other area of contact. It is always painless and may be accompanied by swollen glands but later disappears by itself.

The second stage that has been tagged moderate stage begins approximately seven weeks [3] after the sore firstly appeared. Common symptoms associated with this stage are rashes appearing on the body (palms, legs, trunk, arms, legs and so on) and mucous membrane lesion. Other accompanying symptoms include muscle ache, throat sore, fever, headache, loss of appetite, fatigue, swollen glands and so on. These symptoms last for few weeks and disappear if not treated and later enter into the latent stage. The tagged severe stage (latent) is always void of symptoms and this finally leads to the tertiary stage (very severe stage). The tertiary stage leads to death and permanent damage of the body even after it might have been treated. Difficulties in coordinating muscle movements, paralysis, gradual blindness, numbness and dementia are some of the symptoms associated with the tertiary stage.

• B.    Overview of Medical Diagnostic Systems

The development of medical diagnostic systems right from the onset has been considering various approaches: production rules based, Bayesian network based, statistical model based, fuzzy logic based and artificial neural network based medical diagnostic systems.

The production rules (if then rules) based medical diagnostic systems consider diagnosis using certainties. Here, all rules are formed by combining antecedents using Boolean logic operators (AND, OR and NOT). The antecedents here are symptoms or medical tests that are carried out in the laboratory. Example of such rules is:

IF (patient’s body temperature is >37 degree Celsius AND patient complains of ache in the head) THEN (patient suffers headache)

MYCIN was one of the diagnostic systems that used this type of approach in making medical diagnostic. The diagnostic operation as mandated by the approach used demanded that all antecedents are precise and there was no room for uncertainties in its operation. This diagnosis based on certainties has made it service to be non human like. Other systems that use this system exist but discussion of these is not a focus in this paper.

Bayesian network based medical diagnostic systems operate using uncertainties and diagnosis making is always based on probability. This approach represents the probabilistic relationships existing between diseases and symptoms. Due to the mutual relationship existing between diseases and their symptoms showed by a patient, the model can mathematically compute the probability of the diseases that might be responsible for the symptoms. ORAD is an example of a medical diagnostic system that uses this probabilistic model [4]. Bayesian network based medical diagnostic systems are always prone to overestimation of uncommon but important features and underestimation of important common features [4] [5] due to its probabilistic reasoning nature.

Statistical based medical diagnostic model operates based on the frequencies of occurrence of some diseases in an area. A region might be facing some outbreaks of diseases say “A” and “D”. When a patient shows some symptoms that are associated with disease “A” and “B” in such an area, the patient might be diagnosed to be suffering from “A” since disease “A” is found occurring mostly in the region. This diagnostic service here is not always precise and mostly error prone.

The Fuzzy logic based medical diagnostic systems are effective with the handling of diagnostic service is approximately equal to the human like way of diagnosing. It handles diagnosis using uncertainties. Although it uses rules like we have in production rules based approach but, the rules here are fuzzy rules that are based on approximation and not on exactness. Reasoning here is fuzzy based and not a probabilistic one as contained in the Bayesian network systems. It is always most effective if this approach of diagnostic system development is combined with the artificial neural network approach of medical diagnostic system. The later approach is based on the natural working method of human neural network.

• C.    Related Work

The research interest of artificial intelligence in medicine (AIM) is not a few years old. It is therefore believed to have been some various forms of research works in the field that all attempted applying concepts in artificial intelligence (AI) into solving some problems in medicine. This paper will only review some literatures that are directly related to the topic under discourse.

A computer assisted diagnostic system for red eye was developed in [6] using visual basic programming language, Microsoft excel and some other tools. It was found to operate on the approach of production rules that is always based on certainties. This system was found to have assisted in delivering Medicare to patients who suffered red eye ailment in the remote parts where medical practitioners did not exist or were inadequate. This system due to its approach of operation could not provide a fuzzy based diagnostic that is always inherent in medicine.