Towards Optimizing Hospital Resource Allocation through Process Mining: A Data-Driven Approach H/F

Le poste proposé

0.

Scientific Context and Problem Statement

In the context of increased pressure on hospital resources and high demands for quality care, healthcare facilities face considerable challenges in optimizing the management and allocation of their human resources. The complexity of hospital processes, characterized by significant variability in activities and patient flow, makes the allocation of healthcare staff (doctors, nurses, nursing assistants) particularly challenging. Additionally, the growing digitization generates a substantial volume of data daily within Hospital Information Systems (HIS), yet this data often remains underutilized for process and resource optimization. In this context, Process Mining emerges as a promising approach to derive actionable insights from digital traces of hospital processes, offering new perspectives for analyzing and improving healthcare management practices.

Recent research highlights the growing interest in applying Process Mining within the healthcare sector, demonstrating its potential to analyze and optimize hospital processes. However, its specific application to human resource optimization remains an expanding field of research, especially in the French context, where regulatory and organizational constraints pose particular challenges. Optimizing human resource allocation in hospitals is crucial for both care quality and operational efficiency, yet leveraging available data for this objective continues to face several obstacles.

Thus, the central question of this research is: how can Process Mining, applied to data derived from open health data sources (PMSI, SAE, ATIH), be used to optimize human resource allocation in healthcare facilities? This issue requires leveraging advanced data science methods, including data mining techniques and artificial intelligence algorithms, to extract and transform these data into structured event logs that can be processed by Process Mining algorithms.

The methodology to be followed consists of two main phases. The first phase involves preprocessing, cleaning, and standardizing open data to produce high-quality event logs suitable for Process Mining. The second phase aims to apply Process Mining algorithms to discover process patterns, analyze their compliance, and identify optimization opportunities for hospital resource allocation. To validate this data-driven approach, a concrete case study on hospital human resource allocation will be conducted. This case study will allow for experimenting with and evaluating the developed algorithms in a real-world context, demonstrating the potential of Process Mining for hospital resource optimization. The overall approach aims to develop a reproducible methodology for analyzing and improving healthcare human resource allocation processes.

Work Plan

The completion of this work highlights several aspects to be developed during this internship:

1. Conduct a literature review to identify applications of Process Mining in healthcare, then gather and structure the necessary health data, ensuring compliance with confidentiality and quality requirements.
2. Develop preprocessing algorithms using data mining and artificial intelligence techniques to transform open data into event logs that can be utilized by Process Mining.
3. Design a methodology for analyzing hospital processes based on Process Mining.
4. Apply Process Mining to model human resource allocation processes, documenting key steps, organizational constraints, and observed bottlenecks.
5. Analyze inefficiency factors and specific constraints related to human resource allocation using visualization tools to identify areas for improvement.
6. Formulate recommendations to optimize human resource allocation and propose directions for future optimization algorithm development.
7. Prepare a final report and presentation materials to disseminate the results and guide future research on hospital process optimization.

Context

Laboratory Presentation

CESI LINEACT (UR 7527), the Laboratory for Digital Innovation for Business and Learning to Serve Regional Competitiveness, anticipates and supports technological transformations in sectors and services related to industry and construction. CESI’s longstanding close relationship with companies is a key element of our research activities, leading us to focus on applied research that is closely aligned with and conducted in partnership with businesses. A human-centered approach, coupled with technology, as well as a strong territorial network and links with training programs, has enabled us to develop cross-disciplinary research that places human needs and usage at the center, addressing technological challenges through these lenses.

Our research is organized into two interdisciplinary scientific teams and two application domains:

1. Team 1: « Learning and Innovating » focuses on Cognitive Sciences, Social Sciences, Management Sciences, Training, and Innovation Sciences. The main scientific objectives are to understand the effects of the environment, particularly situations enhanced by technical objects (platforms, prototyping workshops, immersive systems, etc.), on learning, creativity, and innovation processes.
2. Team 2: « Engineering and Digital Tools » focuses primarily on Digital Sciences and Engineering. Key scientific objectives include modeling, simulation, optimization, and data analysis of cyber-physical systems. Research also addresses decision-support tools and studies human-system interactions, particularly through digital twins integrated with virtual or augmented environments.

These two teams develop and integrate their research within the two application domains of the Industry of the Future and the City of the Future, supported by research platforms, notably the Rouen platform dedicated to the Factory of the Future and the Nanterre platforms dedicated to the Factory and Building of the Future.