Database Design

The database is a series of tables which contain the data from each data collection document; a few tables which consolidate information about the patient and the tumor; look-up tables which provide the required code set values as dictated by NAACCR, SEER, ICD-O-3, etc.; and an extract table which has a field for each field in the NAACCR file layout for pulling data together for cancer cases to be reported.

The database management system (DBMS) used is Microsoft SQL server. Our electronic health records are also in Microsoft SQL databases. The business intelligence tool utilized is Microsoft’s SQL Server Integration Services (SSIS), which allows us to develop packages that obtain the desired data. The tasks include creating the initial patient records and updating the tumor tables to  record specific information, such as the date of diagnosis, using business rules developed with the oncology group. After quality assurance activities are completed the data is then migrated on cases ready to be reported into an extract table. Elements are re-coded with the appropriate coding based on records stored in terms tables.

All the data elements on the data collection forms in the EHR are brought into the Cancer Registry database, even those not required in cancer registry reporting. This data is used for quality assurance, administrative purposes, practice management and potentially for future research purposes.

When new data collection forms (for instance, pathology tumor checklist templates or TNM forms) are created or revised, all the possible listbox responses or coded values that will be included in the state cancer case reporting are confirmed or entered into terms table with the appropriate coding for the extracts. We also manually code any free-text entries in “other” fields, like those for histologic type.

The data collection forms have descriptive information which can be used to populate the ten text fields in the data extract tables. We also extract text from imaging studies relevant to the cancer into the text fields. We’ve worked with the Colorado Central Cancer Registry to continually improve the population of these text fields.

Click to view larger image. Click to view larger image.

Main Tables in the NJH Cancer Registry Database

Tables Used to Extract Data from the Main Registry Tables Above