Banca de QUALIFICAÇÃO: PAULO CESAR OLIVEIRA DANTAS DE ABRANTES

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : PAULO CESAR OLIVEIRA DANTAS DE ABRANTES
Date: 31/07/2026
TIME: 09:15
LOCAL: https://conferenciaweb.rnp.br/sala/andre-39
TITLE:

Healthbots for Conducting Clinical Screening and Remote Monitoring with Patient Mood Assessment

PAGES: 64
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUBÁREA: Telecomunicações
SUMMARY:

Assessing patient mood is important for managing chronic diseases, yet it is routinely omitted from remote monitoring because it is labor-intensive and difficult to scale. Conversational health agents, or healthbots, can sustain patient contact between visits, and recent advances in affective computing and large language models make it feasible to embed emotional assessment into everyday clinical interactions. However, few healthbots integrate mood assessment into structured clinical workflows, and the modeling and validation foundations required to do so reliably remain underdeveloped. This document presents a doctoral qualifying examination monograph, structured in the multi-paper format, which proposes the investigation of the design and implementation of a blended-care healthbot for clinical screening and remote patient monitoring while continuously assessing patients' emotional states through multimodal emotion recognition. The work is organized as three connected studies that progress from evidence synthesis to modeling foundations to system implementation. The first study is a scoping review, conducted under the Arksey and O'Malley framework and PRISMA-ScR guidelines, that maps AI-based healthbots that combine clinical screening and remote monitoring with mood assessment. It characterizes a developing but immature field and identifies recurring gaps in clinical integration and interoperability, temporal continuity, clinician involvement, transparency and privacy, and external clinical validation. The second study develops reproducible unimodal baseline models for emotion recognition from text, speech, and facial expression using public datasets, establishing transparent benchmarks for multimodal integration. The third study designs and implements a blended-care healthbot that fuses these unimodal models into a shared clinical vocabulary through weighted soft voting and embeds the resulting affective signal within a protocol-driven screening conversation, with risk stratification, prioritized clinician alerts, and bidirectional HL7 FHIR integration with electronic health records. Together, the three studies advance emotion-aware healthbots from a recognized capability toward a clinically actionable system, addressing the gaps identified in the literature. The concluding step of the doctoral project, presented here as the proposed continuation, is the prospective clinical validation of the system with patients. The implemented architecture was designed from the outset to support this validation.

COMMITTEE MEMBERS:
Presidente - Interno ao Programa - 2334927 - ANDRE KAZUO TAKAHATA
Membro Titular - Examinador(a) Interno ao Programa - 1761107 - RICARDO SUYAMA
Membro Titular - Examinador(a) Externo à Instituição - RODOLFO DE CARVALHO PACAGNELLA - UNICAMP
Membro Suplente - Examinador(a) Interno ao Programa - 1544392 - ALINE DE OLIVEIRA NEVES PANAZIO
Membro Suplente - Examinador(a) Externo à Instituição - RODRIGO BONACIN - CTI