Towards Reliable Clinical Data: A Collaborative Data Governance Architecture with Lifecycle Integration

Clinical data are critical for healthcare decision-making and research, but persistent concerns over accuracy, completeness, and consistency limit real-world reliability. Existing quality frameworks often emphasize post-hoc control and lack integration into clinical workflows. This study developed and evaluated a proactive, workflow-embedded, lifecycle-oriented governance architecture to address these limitations.

This study proposed the Comprehensive Clinical Data Governance Architecture (CCDGA), consisting of a Data Lifecycle Framework (DLF) to define control nodes, a Collaborative Data Governance Framework (CDGF) for role-based interventions, and a Data Governance Committee for organizational oversight. These components establish a closed-loop mechanism embedded upstream in data capture. A postpartum hemorrhage (PPH) case study was conducted in a tertiary maternal hospital, using interrupted time series (ITS) analysis of 64 monthly observations.

Completeness rate improved significantly after the intervention (β₂ = 30.849, P < 0.001), while the post-intervention trend was not significant (β₃ = -0.049, P = 0.579). Standardized data entry improved markedly (β₂ = 80.163, P < 0.001) and showed a positive trend (β₃ = 0.193, P = 0.045). Alignable record rate also improved (β₂ = 29.812, P < 0.001), but its post-intervention trend was not significant (β₃ = -0.141, P = 0.126).

CCDGA introduces a lifecycle-informed governance model that integrates collaborative interventions with organizational oversight. By embedding governance upstream in data generation rather than post-hoc correction, it shifts assurance from retrospective auditing to proactive, process-integrated improvement. In the PPH evaluation, completeness and alignable records stabilized at high levels and standardized entry improved steadily, supporting scalable and sustainable clinical data governance through KPI-based monitoring and escalation pathways.