Reference only; verify against the target FastGPT release and deployment documentation.
What this category of data looks like
This category serves medical insurance handling institutions, designated hospital medical insurance offices, designated pharmacies, and medical insurance information platform operators. Daily processed documents include medical insurance settlement lists, expense details, medical record homepages, prescriptions, rejection lists, and settlement reconciliation statements. Data sources are the national medical insurance information platform, hospital HIS and medical insurance interfaces, medical insurance catalog and coding libraries, and cross-provincial medical treatment settlement platforms. Update frequency is daily. Document fields cover diagnosis and treatment item codes, pharmaceutical and consumable codes, expense amounts, insured person information, medical institution information, and must strictly comply with standardized medical insurance coding rules.
What constraints do these characteristics impose on the citation source and traceability link
Data sources are scattered and updated daily. Traceability must be linked to the latest medical insurance catalog and interface data to avoid coding matching failures. Documents for single or batch settlements have high standardization but many fields. Precise recall of the original sources of corresponding codes and expense details is required to prevent redundant or missing traceability results. Traceability for rejection lists and appeal reconciliation must be linked to historical settlement records. The traceability scope must be limited to business data from the corresponding period, while avoiding interference from irrelevant data. Errors directly affect medical insurance fund security and insured person rights. The accuracy of traceability results must strictly match business rules, and audit deviations caused by incorrect citations must not occur.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | Matches the length of single-line details in medical insurance settlement documents to avoid truncation of critical coding and expense information |
| Recall Result Limit | Top 5 entries | Controls the number of documents returned per traceability to avoid information overload during case file audit |
responseData | Structured summary | Adapts to the standardized field extraction requirements of medical insurance audits, facilitating subsequent rule verification |
detail | Enabled | Fully displays source identifiers and field information of cited documents to meet traceability verification requirements |
| Timeout Threshold | Calibrated based on actual testing | Adapts to response delays of medical insurance interfaces to avoid traceability failures caused by timeouts |
| Knowledge Base Scope Restriction | Only call the dedicated medical insurance settlement audit knowledge base | Avoid introducing non-medical insurance data to interfere with audit results |
Three Common Mistakes
- Interface prompt "No specified knowledge base found" or empty search results. Cause: The global variable datasetid set in the workflow is not correctly bound to the knowledge base retrieval node, or there is a break in the variable transfer path.
- Only the first query triggers knowledge base citations in a multi-round audit process, with no retrieval results in subsequent steps. Cause: Workflow context transfer rules are not configured, causing subsequent nodes to lose previously associated knowledge base parameters.
- Cited document fragments display red error markers. Cause: The cited document has been updated or deleted, or the cited content length exceeds the
quoteMaxTokenlimit, resulting in truncation failure.
How to Confirm Proper Configuration
- Submit a simulated medical insurance settlement audit query, check whether the retrieval results only include medical insurance settlement-related documents.
- View the displayed content of the citation source, confirm that key identifying information such as document source and update time is included.
- Adjust the
quoteMaxTokenparameter, test whether settlement documents of different lengths can be fully displayed without triggering truncation errors. - Trigger a multi-round audit process, verify whether subsequent steps can normally call the knowledge base for traceability.
The values provided are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.