Navegando por Autor "Alves, Thomaz Henrique Martins"

Agora exibindo 1 - 1 de 1
  • Imagem de Miniatura
    Item
    Um guia de boas práticas para criação de uma base de dados pública para auxiliar no desenvolvimento de softwares voltados ao tratamento de transtorno de pânico usando swartwatches
    (2025-06-26) Duarte, Clísthenes Freire da Cruz; Cunha, Mônica Ximenes Carneiro da; http://lattes.cnpq.br/1775024859845111; Nunes Filho; Nunes Filho, Ricardo Rubens Gomes; http://lattes.cnpq.br/1760182180822152; Cruz, Jailton Cardoso da; http://lattes.cnpq.br/9366016044068759
    This study sought to create a Good Practices Guide for creating a public and reliable database to assist in the development of applications aimed at detecting and/or treating Panic Disorder (PD), using as a tool the most commonly used wearable device today: the smartwatch. The methodological procedure used in this research consisted of four stages: 1) Systematic Literature Review (SLR) with the aim of verifying the state of the art on the topic in order to understand the situations that hinder the development of applications focused on PD; 2) a documentary research to verify the legal and ethical implications related to the collection and storage of data from PD patients; 3) an exploratory research that consisted of interviews with three psychologists and three psychiatrists in order to identify the legal, ethical and methodological criteria for the condition of collecting biometric data from PD patients; 4) the elaboration of a good practices guide that considers the aspects listed in the previous steps. SLR revealed that applications aimed at predicting and detecting PD have already been developed in hospital research environments, using known Artificial Intelligence algorithms such as Random Forest, Decision Tree, Linear Discriminant Analysis, Adaptive Boosting (AdaBoost), Regularized Greedy Forest, among others, and manipulating variables such as variation in pressure, heart rate, hand movements, speech frequency, electrodermal activity, among others, obtained through measurements carried out on PD patients. The sensors used to measure these variables in question have versions available in most modern smartwatch models. The documentary research stage revealed, through the analysis of the LGPD, that there are legal implications regarding data manipulation. It was also detected, through the analysis of the Professional Code of Ethics for Psychology, that there are ethical implications to be followed when conducting the data collection. The interview stage validated the correlation of the variables indicated in the RSL stage with the TP and added that variables obtained in standard questionnaires, such as related diseases, use of medications, patient impressions, among others, are important for data collection, especially if the patients evaluated have other associated diseases. It was also defined that the ideal profile of a participant in data collection is a person with TP who is not easily influenced and does not have other types of psychological disorders. Finally, from the data collected, an artifact was generated, called a Good Practices Guide, with eight recommended steps for carrying out efficient data collection to be made available to developers. The conclusion highlights the advantages of following the guide produced for standardizing data collection and leaves as future work the activities of preparing the Consent Form and the Cooperation Form, in partnership with legal professionals, and also the preparation of a standard questionnaire with nonbiometric variables that also make up the final database.