A data-driven governance framework for co-creation healthcare quality in public hospitals

In low- and middle-income countries (LMICs), unstructured patient-generated data remain underutilized in public healthcare systems, where they are often treated as isolated technical inputs rather than integrated into governance processes. This study develops a data-driven governance framework to support data-informed governance of healthcare quality in public hospitals, positioning patient experience (PX) as patient-generated data from which governance-relevant operational signals can be derived. An exploratory mixed-methods design was employed, combining a conceptual governance framework with empirical analysis of 450 publicly available patient reviews collected from Google Maps (Q4 2024 - Q3 2025) in a public hospital in Viet Nam. Topic modelling was conducted using BERTopic (PhoBERT embeddings, UMAP, HDBSCAN), and aspect-based sentiment analysis (ABSA) was performed using a prompt-based large language model. Extracted aspects were mapped to SERVQUAL dimensions and interpreted through the Donabedian framework using a rule-based dictionary approach. The findings indicate how PX-derived signals can be translated into structured, interpretable, and decision-relevant insights, representing a potential form of governance-oriented public health intelligence aligned with service quality dimensions and system-level quality domains. This study contributes to public healthcare systems research by redefining patient-generated data as governance-relevant inputs and reframing data-driven approaches. The findings are exploratory and context-specific but suggest a structured pathway for leveraging PX analytics to support continuous monitoring and priority-setting in healthcare governance.