Automatically selecting patients for clinical trials with justifications

• Autor(a) principal: Botelho, Tiago Miguel Moreira
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/155509
• topic: clinical trials; description logics ontologies; nonmonotonic rules; hybrid knowledge; explanations; justifications; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

Clinical trials are human research studies that are used to evaluate the effectiveness of a surgical, medical, or behavioral intervention. They have been widely used by researchers to determine whether a new treatment, such as a new medication, is safe and effective in humans. A clinical trial is frequently performed to determine whether a new treatment is more successful than the current treatment or has less harmful side effects. However, clinical trials have a high failure rate. One method applied is to find patients based on patient records. Unfortunately, this is a difficult process. This is because this process is typically performed manually, making it time-consuming and error-prone. Consequently, clinical trial deadlines are often missed, and studies do not move forward. Time can be a determining factor for success. Therefore, it would be advantageous to have automatic support in this process. Since it is also important to be able to validate whether the patients were selected correctly for the trial, avoiding eventual health problems, it would be important to have a mechanism to present justifications for the selected patients. In this dissertation, we present one possible solution to solve the problem of patient selection for clinical trials. We developed the necessary algorithms and created a simple and intuitive web application that features the selection of patients for clinical trials automatically. This was achieved by combining knowledge expressed in different formalisms. We integrated medical knowledge using ontologies, with criteria that were expressed using nonmonotonic rules. To address the validation procedure automatically, we developed a mechanism that generates the justifications for each selection together with the results of the patients who were selected. In the end, it is expected that a user can easily enter a set of trial criteria, and the application will generate the results of the selected patients and their respective justifications, based on the criteria inserted, medical information and a database of patient information.