According to the DMAIC method of Six Sigma, a quality risk control system for clinical trials at the large hospital should be built, and the main risk areas and governance paths need to be determined. According to the back-end data from a third-tier Grade A hospital in 2021–2025, both industry-sponsored trials (ISTs) and investigator-initiated trials (IITs) were chosen. A total of 937 problems has been identified so far, and among them, 715 will be the subjects of this study from 2023 to 2025. The 11 risk areas have been partitioned, and according to the Pareto analysis method, frequency-based sigma (σ) priority indicators and the DMAIC process model will be used for evaluation. The first seven kinds of risk were about 89.9 per cent. The first were source data and documentation (182 items, 25.5%, σ = 2.16), and next was informed consent and participant rights (102 items, 14.3%, σ = 2.57). IIT projects had a relatively high number of problems per project compared with IST projects in general; in 2024, it was 11.80 problems per project and dropped to 6.50 in 2025. During the pre-startup quality control, there were 52 defects, and about 98.1 per cent of them have been rectified. In short, there are two types of risks in clinical trials, and among them, “vital few” refers to a small number of aggregation points in some areas. Embedding DMAIC in all links of the clinical trial process and combining it with corrective and preventive actions (CAPA) and risk-based process management will build a quality risk control system based on a structured issue database, key risk indicator (KRI) early warning mechanism, proactive prevention and closed-loop rectification.