University of Nevada - Reno
NurseAssist: A Deterministic and Privacy-Preserving Pipeline for Automated Clinical Assessment
Abstract
dc:description.abstractNursing education is rapidly modernizing, yet delivering efficient, accurate, and objective grading remains a critical challenge due to staffing limitations and high student-to-instructor ratios. Furthermore, automated graders suffer from non-determinism (i.e., variability in outputs), limited interpretability, and privacy concerns. This work introduces NurseAssist, an assistive system that automates the grading of student nurses performing triage on mock patients in an educational setting. The system implements an acoustic processing layer that converts audio input into a role-mapped, diarized, medical transcript. This information is passed to an inference layer, which processes the transcript and returns an easy-to-interpret graded assessment report. Additionally, a deterministic framework is applied to the system to ensure reproducible inference, system logging, and auditability through a reproducibility verifier tool. The system integrates powerful Large Language Model (LLM) capabilities with robust algorithmic system design to ensure accurate and efficient grading. A comprehensive assessment of the system is also provided, highlighting the broader potential of deterministic, reproducible pipelines for applications in education and beyond.
Degree
thesis:*- Level thesis:degree_level
- Master’s Degree
- Department thesis:degree_department
- Computer Science
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Molina, Joel Daniel
- Advisor dc:contributor.advisor
-
- Tavakkoli, Alireza
- Committee members dc:contributor.committeemember
-
- Hu, Rui
- Richard, Shannon
Subjects
dc:subject × 3Rights
- Language dc:language.iso
- en_US, English
Identifiers
dc:identifier.*- Repository record dc:identifier.uri
- https://scholarwolf.unr.edu/handle/11714/11883