Policy on Release of Old Cases with Update Requests

Investigators who are currently using SEER-Medicare data have contacted NCI and IMS to inquire about obtaining updated files to include data for more current cases. If investigators request updated data, they should be aware that IMS will send them new data for all cases, not just the new cases. Several researchers have asked why they are receiving new data for all cases instead of data for the new cases only.

Many years ago, IMS tried releasing data only for the updated cases. There were major difficulties for IMS and for the researchers who were trying to use these files, as the differing versions of the data set did not reconcile. There are several reasons that different versions of the SEER-Medicare data do not reconcile. SEER and CMS are constantly updating their data bases. As a result, with each update of the linkage there are slight differences in which patients are in the SEER or the Medicare data. Some old cases will no longer be in the data; some new cases from earlier years will now be found. In addition, the layouts of the Medicare files change. This makes it difficult to append data from a recent version of the SEER-Medicare data to earlier versions of the dataset.

In order to minimize the production costs of SEER-Medicare files, IMS has created standardized data extraction programs that allow for the selection of cases based on cancer site(s) and years of diagnosis. From IMS' earlier experience, the additional programming required to reconcile cases between old and new linkages is time consuming and results in longer processing time. In the long run, trying to append the two data sets is more costly than producing the entire data set de novo.