ABSTRACT
A number of avoidable blindness could have been averted if efficient eye diagnostic system were available and accessible to persons suffering from various eye diseases especially in remote and inaccessible communities where there is no eye doctor or where the eye doctors are reluctant to go. The existing eye diagnostic model uses Bayesian Neural Network which is less efficient in making decisions involving small databases that are mostly prevalent in eye care medicine. It neither uses visual acuity of patient as validating tool for patient’s symptoms nor allows the patient to provide the level of confidence associated with his/her symptoms. In addition, the existing system does not integrate diagnosis and visual acuity measurement into a single platform. Development of a model that can surmount the setbacks prevalent in the existing system and at the same time provide reliable eye disease diagnosis, is sine qua non for prevention of avoidable blindness especially in developing countries. This dissertation is aimed at developing a Certainty Factor Reasoning Model for eye disease diagnosis. Rule Based Expert System Methodology and Object-Oriented Analysis and Design Methodology were used for knowledge and data representation respectively. Certainty Factor Method was used for reasoning under uncertainty. Visual Prolog 7.5 was used for the design and implementation of the new system while MYSQL 1.0.9 was used for the design and implementation of the database. Disease diagnosis was inferred by analyzing the visual acuity, the symptoms and certainty factor provided by the patient with the facts, rules and also the certainty factor provided by the domain expert. Test of accuracy for the diagnosis made by the new system using Cohen’s kappa tools revealed k-value of 0.827 while Test of accuracy and test/retest reliability for the visual acuity measurement using Bland-Altman tools revealed p-values of 0.042 and 0.00004 respectively. The stress/overload test was successfully carried out on 154 operations. These performance indicators imply that the new system is accurate, reliable and repeatable. This study has reinforced the belief that certainty factor model could be a useful reasoning tool for making decisions especially in uncertain situations and in small databases. It is believed that the model could help in eradication of avoidable blindness as well as provide training tools to eye doctors, clinical students and can also serve as information portal to the patients. Further research into development of web-enabled version of this model to extend its reach is recommended.
CHAPTER ONE
INTRODUCTION
- Background of the Study
Despite the advances already made in diagnosing eye diseases, accessing quality eye care services is still a challenge in developing countries. The chronic absence of eye care services in the rural communities in the developing countries and the difficulties encountered by the patients in consulting eye doctors have pushed some patients to seek self-help with its attendant consequences (Merck, 2006). Such self- help often culminates to self-medication. This may include use of un-recommended eye drugs, use of traditional harmful medications and use of other concoctions such as urine, engine oil, honey, sugar, holy water and breast milk. Also, some blinding eye conditions such as glaucoma and central retinal vein occlusion do not leave trace of early warning signs to the patient before they hit. Furthermore, Kurniawan et al.(2014),noted that in reality, poverty stops people in the developing world from consulting an eye doctor regularly. Thus, many patients do not get appropriate treatment for their eye disease until it is too late. Hence, a simple eye disease which could have been treated with appropriate eye glasses or recommended drugs may often propagate to blindness. If there is but a standard place or health kiosk where the patients could get reasonable eye-check attention, the current level of global blindness could have been abated to acceptable level (Serge & Tricia, 2012).Unfortunately, the rural communities are characterized by challenging conditions such as difficult terrain/topography, lack of good road network, lack of adequate electricity supply, lack of good communication facilities, lack of social infra-structure (pipe-born water, accommodation, standard market and leisure centers) (Ntsoane & Oduntan, 2010). These factors have made living in and visiting the rural communities less attractive to the eye care personnel. Consequently, setting up standard eye clinic in these areas is not considered a cost-effective investment. The result is that the distribution of eye clinics and doctors in most developing
countries is skewed towards the urban cities, thereby leaving the eye care services in the rural communities chronically under-served! (Abdull, et al., 2009). It is therefore evident from the aforementioned that the problem is from the health system in practice today in which patients (are forced by circumstances to) consult their doctors only when they are sick. Development of an automated eye diagnostic system which could detect eye diseases in the absence of the eye doctors, could help a lot in provision of un-interrupted and quality eye care services in remote and inaccessible communities.
Several efforts were made by researchers in the past in attempt to developing a self-automated system that could diagnose eye diseases in the absence of eye doctors. The existing models used various expert system methods, with each having one setback or the other such that the diagnostic output are still far outcry from the clinical expectations. Most of these models relied on production rule (rule base) as method of knowledge representation. Using production rule alone has its challenges in representing complex data structure for all possible evidences and their corresponding hypothesis which are often encountered by the domain experts in real life. To get around this challenge, object-oriented data structure had been tried of recent to drive rule base systems. Eye disease diagnosis requires input of symptoms by the patients, but most patients are not very certain about the full characteristics of their symptoms and thereby leaving behind the possibility of uncertainties (Bhupendra, 2016). These uncertainties which are intrinsic in most medical diagnostic systems were managed by the existing systems using mostly Bayesian network, fuzzy logic and Dempster-Shafer methods with sparing use of certainty factor method. Using Bayesian network, fuzzy logic and Dempster-Shafer methods do not give the patient the opportunity to weigh his/her symptoms thereby leaving out uncertainty management which is an intrinsic aspect of expert systems. It is therefore pertinent that development of a diagnostic system that could consider or manage the uncertainties emanating from the patient could improve the diagnostic quality.
Another essential input parameter in the diagnosis of eye conditions is the visual acuity status of the patient (Jesse et al, 2010). The visual acuity expresses the condition of the eye at each point in time. Hence each eye condition/disease has associated visual acuity status. The existing system had applied various methods of visual acuity measurement ranging from Snellen chart analogue to interactive projection system. Unfortunately, the existing system could not integrate the symptom input with the visual acuity status of the patient, thereby exposing diagnostic output to unreliable discrepancies as there is no way the patient‘s claims are verified.
From the foregoing, it is evident that a reliable eye diagnostic system shall have the ability to consider the visual acuity of the patient, the symptoms and the degree of uncertainties associated with both the symptoms and the diseases. Therefore, the proposed model shall fulfil above expectations by measuring the visual acuity as well as collecting symptoms from the patients with their associated certainty factors, analyze these input values together with the rules supplied by the domain experts and then proffer diagnosis. This could be achieved by designing an expert system model that could use certainty factor as a reasoning tool and rule base for knowledge representation.
1.2 Statement of the Problem
Each eye disease condition affects the functional quality of vision and invariably affects the visual acuity of the patient. Therefore, consideration of the level of visual acuity of a patient and matching this with the symptoms presented by the patients are inseparable determinants in diagnosis of eye diseases,(American_Optometric_Association, 2016). This implies that anefficient eye diagnostic model shouldconsider the visual acuity of the patient(Gary & Jeniffer, 2016). Unfortunately, existing diagnostic models rely mainly on the symptoms, signs and at times laboratory results as supplied by the patients without considering the visual acuity of such patients.The existing automated visual acuity
systemcould only offer diagnosis or measure visual acuity but not both despite the inseparability of the duo in eye diagnosis (Jacobs etal., 2014). Furthermore,grading of symptoms by patients with respect to degrees of uncertainties varies from one patient to another. Hence, if a model was to rely solely on the claims made by the patients, such a model should also allow the expression of the degree of certainty with which the patient has provided each symptom. Most existing models rely on Bayesian network as a method of reasoning but Bayesian model does not evaluate the level of belief or disbelief a patient has but relies on prior probability values supplied by the doctor during knowledge acquisition and thereafter datamines the causal relationship between the disease and symptoms. Existing models also depend on large databases with assumption of complete independencies between coexisting variables (symptoms), an assumption which cannot hold in all cases and which has also affected the range of eye diseases diagnosed, which currently stands at twelve. Therefore, a gap in technology is evident in the existing models and may have created credibility issuesamong the target users (Kragt, 2009) and (Zhang , 2006) as most of the models still wander at experimental level. The existing eye diagnostic system therefore is not designed for small knowledge bases as it uses Bayesian network, does not measure or use visual acuity as a way of verifying or validating patient‘s claims, does not allow the patient to provide the level of confidence or belief in the information he or she has supplied and does not integrate both diagnostic and visual acuity measurement into a single platform as expected in modern times (Shaun, 2017).
1.3 Aim and Objectives of the Study
The aim is to develop a certainty factor reasoning model for eye disease diagnosis The specific objectives are to:
- Present or formulate the certainty factor reasoning model for eye disease diagnosis
- Analyze and design the certainty factor reasoning based expert system for diagnosis of eye diseases.
- Develop the certainty factor reasoning based expert system for diagnosis of eye diseases.
- Evaluate the performance of the proposed system by comparing it with the existing systems.
1.4 Significance of the Study
This study is significant as follow:
- Doctor‘s time will be saved since only genuine and serious cases will consult doctors. The spare time could be invested in research, literary work and other clinical activities.
- The system when implemented will serve as both training and learning tools to the medical students, general practitioners and other paramedics
- Prompt and early detection of eye diseases by the system could help in reduction of global blindness.
- Reduction in global blindness will translate to having more able-bodied citizens who could contribute positively to the growth of the economy of the state.
- The patients could learn disease characteristics with the system and then be on watch-out for early signs and symptoms at first occurrence.
- The system could serve as ocular first aid advisory in cases of ocular emergencies.
- Since the system could be located closer to the rural dwellers, it will save the patients the cost and risk of travel to urban cities to seek eye care services
- It will encourage implementation of certainty factor as a reasoning model in other diagnostic and troubleshooting expert systems
1.5 Scope of the Study
The scope of this work is limited to the expert systems meant for eye disease diagnosis. It covers the diagnosis of eye diseases which could be detected through diagnostic questions and visual acuity assessment only. The expert system is limited to the genre which uses rule base method for knowledge representation, certainty factor method for reasoning under uncertainty and object-oriented analysis and design method for data management.
1.6 Limitations of the Study
- Few eye clinics useautomated diagnostic systems
Only few eye clinics within the study region could afford automated eye diagnostic system while the majority are still using manual approach. Therefore the number of eye clinics available for selection as study centers for the analysis of existing system was limited.
b) Paucity of data in study area
Accessing relevant literatures for models which combine both visual acuity measurement and eye diagnosis into a single platform is rare. Consequently, getting adequate literature and data for comparison with the developed model has set a limitation to the extent of this study.
c) Fear of job competition by the eye doctors
During the process of knowledge engineering in which the domain experts were approached for release of information on eye disease diagnostic procedures, some eye doctors were skeptical and suspicious of the study as according to the comments made by some of them, if the patients could be testedwithout seeing eye doctor, they may not visit their doctors as frequent as they used to. Therefore, fear of direct or indirect job competition had made the domain experts (eye doctors) become economical with information they provided.
d) Dynamic and variable knowledge base
Eye Medicine is a fast-expanding field in terms of knowledge and way of management of eye conditions. Practices which are considered safe today will tomorrow become obsolete and forbidden. Although the system‘s knowledge base was built dynamically bearing in mind the nature of this fluctuating information, nevertheless developing expert system under this scenario has made it almost impossible to add the most current information on signs, symptoms, risk factors and treatment plans of eye diseases used in this dissertation.
e) Paucity of resource materials in domain area
The system was designed to measure the visual acuity as well as diagnose eye conditions. The integration of two different eye examination and diagnostic protocols into one system is a rare invention. The systems already available are either visual acuity tester or diagnostics and not both. Therefore, getting enough resource materials during development and system analysis of this dissertation was both challenging and scarce.
1.7 Definition of Terms
Amsler chart: A chart used to access and record the quality of the central vision.
Blindness: This is a condition in which an individual has no sight at all in the best eye or the sight he/she has could not enable him execute commonest tasks that are sight-demanding.
Color blindness: This is a condition in which an individual cannot recognize a particular color or group of colors.
Color Vision Test (CVT): This is one of the eye tests performed to evaluate how well an individual can recognize and differentiate colors.
Cone: The visual cell inside the eye that is responsible for resolution of image and acute vision.
Consultation: This is a session between a doctor and a patient with an intent of providing medical
services to the patient.
Contrast Sensitivity Test (CST): This is one of the eye tests performed to evaluate how well an individual can distinguish the foreground from the background and vice versa. A perfect contrasting pair is Black (foreground) on white (Background) or White (foreground) on Black (background). For example in the computer displays, the FontColor is declared as ForeColorwhile the background is declared asBackColor.
Cornea: This is the eye‘s transparent front coating.
Expert system: A computer program that applies artificial-intelligence methods to problem-solving
Glucometer: An Instrument used in measuring the blood sugar level.
Integrated system: This is a computer suite which incorporates many eye test logics into one platform.
Hyperopia: This is another name for long-sightedness
Keratometry: Instrument used to measure the curvature of the cornea.
Low vision: This is a situation in which even with the best surgical, optical and drug intervention, an individual‘s sight is still very poor that he/she cannot carry out common activities of daily living (such as watching TV, reading newspaper, cooking, signing bank cheque, crossing road, etc) without need for external assistance.
Long sightedness: This is a refractive error condition when an un-aided individual can see better at Far than near
Macular Function Test (MFT): This is one of eye tests performed to evaluate the quality of an individual‘s central vision.
Medical Assistant: This is a non-expert who does not specialize in domain area but is trained or enabled to perform some tasks which a genuine expert can do.
Model: A simplified version of a system which shows some or all the aspects of the internal workings of the system withthe intentionof analyzing and solving problems or making predictions.
No Light perception: This is a condition when an individual is completely blind
Ophthalmic/Ocular: That which pertains to the human eyes.
Ophthalmoscopy: Instrument used to assess the healthiness of the inner eye.
Optical: That which pertains to Light.
Optotype: This is the characters, symbols and images used in visual acuity test charts
Presbyopia: This is a condition when an individual gradually loses the ability to read near print as he or she grows old.
Refractive Error: The eye condition that can be corrected with eyeglasses.
Retina: The tissue in the eye that receives and transmits the image of an object of regard.
Retinoscopy: Instrument used to determine the refractive error of the eye.
Sphygomanometer: An instrument used to measure the systemic blood pressure.
System: In this dissertation, when written alone without further qualification, system means implementation of ―Certainty Factor Reasoning Model For Eye Disease Diagnosis‖ alias CFRMFEDD.
Vision/sight: A word used interchangeably in this dissertationto imply seeing
Visual Acuity Test (VAT): This is one of the basic eye tests performed to evaluate how sharp an individual can recognize and see objects.
Visual Cortex: The area in the brain responsible for the coordination and perception of visual stimulus.
Visual Field Test (VFT): This is one of eye tests performed to evaluate the extent of vision either in central or peripheral (side) region (measured in angular degrees).
WHO: World Health Organization.
This material content is developed to serve as a GUIDE for students to conduct academic research
DESIGN AND IMPLEMENTATION OF CERTAINTY FACTOR REASONING MODEL FOR EYE DISEASE DIAGNOSIS (CFRMFEDD)
A1Project Hub Support Team Are Always (24/7) Online To Help You With Your Project
Chat Us on WhatsApp » 09063590000
DO YOU NEED CLARIFICATION? CALL OUR HELP DESK:
09063590000 (Country Code: +234)
YOU CAN REACH OUR SUPPORT TEAM VIA MAIL: [email protected]
09063590000 (Country Code: +234)
