Big data Diagnosis-Intervention Packet (DIP) aggregates "diagnoses" and "interventions" into logical categories using data from real-world hospital discharge record. However, problems concerning data sources, such as non-standardized and inconsistent writing of medical records, can lead to duplicate clustering and deviation from clinical treatment guideline and pathways, resulting in unclear case enrollment paths and making it difficult to distribute medical insurance funds. Using colorectal cancer as an example, through comprehensive consultation with clinical experts, this study identifies the problems encountered in practice with the Guangzhou 2020 version of the DIP Grouping and how it differs from actual clinical situations while also making optimization suggestions based on clinical treatment pathways. These suggestions were adopted by the Healthcare Security Administration of Guangzhou and implemented in the newly published DIP Grouping List for 2022. Through the sharing of Guangzhou's experience, this study aims to provide references for DIP grouping optimization in other regions and support the efficient and effective DIP payment method implementation.