The availability of new treatments for a disease depends on the success of clinical trials. In order for a clinical trial to be successful and approved, medical researchers must first recruit patients with a specific set of conditions in order to test the effectiveness of the proposed treatment. In the past, the accrual process was tedious and time-consuming. Since accruals rely heavily on the ability of physicians and their staff to be familiar with the protocol eligibility criteria, candidates tend to be missed. This can result and has resulted in unsuccessful trials.A recent project at the University of South Florida aimed to assist research physicians at H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida, with a screening process by utilizing a web-based expert system, Moffitt Expedited Accrual Network System (MEANS). This system allows physicians to determine the eligibility of a patient for several clinical trials simultaneously.We have implemented this web-based expert system at the H. Lee Moffitt Cancer Center & Research Gastroenterology (GI) Clinic. Based on our findings and staff feedback, the system has undergone many optimizations. We used data mining techniques to analyze the medical data of current gastrointestinal patients. The use of the Apriori algorithm allowed us to discover new rules (implications) in the patient data. All of the discovered implications were checked for medical validity by a physician, and those that were determined to be valid were entered into the expert system. Additional analysis of the data allowed us to streamline the system and decrease the number of mouse clicks required for screening. We also used a probability-based method to reorder the questions, which decreased the amount of data entry required to determine a patient's ineligibility.

The Preop medical expert system (Langton et. al, 1990) was originally developed using the expert system tool, Nexpert Object, on a VAX computer. Nexpert Object creates an expert system specification which is executed by an interpreter within Nexpert Object. The original implementation, however, has several limitations, including: 1. lack of physical portability 2. requires Nexpert Object to run 3. crude user interface. In order to overcome the first limitation, Preop is implemented on a PC DOS portable computer. This project is addresses the other two limitations. Creating a compiled version of Preop eliminates the need for the Nexpert Object interpreter, and implementing it as a Microsoft Windows application provides a better user interface. / Thesis / Master of Science (MSc)

Orientador: Prof. Dr. Peter M. E. Claessens / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Neurociência e Cognição, 2014. / Zumbido (tinnitus) é um distúrbio comum de audição, muitas vezes debilitante em graus variados. Dado que zumbido é uma condição multifacetada, com sintomas que frequentemente são psicológicos e subjetivos, e com muitas causas potenciais, a diagnose deste distúrbio não é trivial. Por exemplo, zumbido pode ser objetivo e mensurável ou subjetivo e produzido por fatores neurais que podem ser de localização mais periférica ou central. Este projeto de mestrado propõe o desenvolvimento de um sistema especialista médico para apoiar clínicos na indicação de tratamento para zumbido. Este estudo foca em três tipos de tratamento para zumbido, a saber, dieta, medicação e aparelho auditivo, como também nas combinações, para categorização supervisionada. Redes Bayesianas ingênuas (naive) foram utilizadas para relacionar uma diversidade de resultados de exames e elementos de anamnese a indicações de tratamento por clínicos. Como tratamentos não são mutualmente exclusivos, a categorização deve levar em conta casos multi-label, isto é, a possibilidade de indicações diferentes de tratamento simultâneas. Com o objetivo de mapear as probabilidades a posteriori das indicações diferentes de tratamento para classificação multi-label , a diferença entre as distribuições a posteriori foi usada como critério para resolver o problema multi-label. Esta estratégia foi avaliada e o desempenho comparada a uma estratégia mais simples de mapeamento single-label. Os resultados mostram que a acurácia da abordagem multi-label é melhor que o ajuste single-label. O sistema fornece assim um primeiro passo satisfatório do desenvolvimento de um sistema de apoio médico futuramente mais amplo, integrado e dinâmico. / Tinnitus is a common hearing disorder, often debilitating to varying degrees. Given that tinnitus is a multifaceted condition, with symptoms that are often psychological and subjective, and with many different possible causes, its diagnosis is not trivial. For example, tinnitus can be objective and measureable or subjective and produced by neural factors which can either be more peripheral or more centrally located. This Master¿s project proposes the development of a medical expert system to assist clinicians in the indication of treatment for tinnitus. This study focused on three types of treatment for tinnitus, namely, Diet, Medication and Hearing Aid, as well as on their combinations for supervised categorization. Naïve Bayes networks were used to relate a diversity of test results and elements of the anamnesis to treatment referrals by clinicians. Because treatments are not mutually exclusive, the categorization needs to take into account multi-labeling cases, that is, the possibility of several simultaneous treatment indications. In order to map the posterior probabilities of the different treatment indications to multi-labeling classification, the difference between posterior probabilities was used as a criterion to solve the multi-labeling problem. This strategy was evaluated and its performance compared to a simpler single-labeling mapping strategy. The result shows that the accuracy of the multi-labeling approach is higher than a single-labeling adjustment. The system thus provides a first satisfactory step in the development of a more encompassing, integrated and dynamic medical support system.