The Cancer Registry community collaborated on a survival algorithm that uses day information in the survival calculation. One goal of this activity was to harmonize as much as possible with efforts such as CONCORD. It may also be possible to evaluate other methods that require more continuous survival data.
Since day is a confidential variable to some U.S. registries, the process was established to work in two ways. The first way uses SAS code that can be downloaded from this site. A registry can run the SAS code locally and only submit survival in months. The other way requests complete Date of Diagnosis (including day) and Date of Last Contact (including day) in the appropriate location in the NAACCR record layout. In either case, the same data (survival in months) will be available for survival analysis. The second option, with complete day information, allows for more analytic flexibility, but choose what works best for your registry.
Regardless of the option chosen, please do not recode any missing date fields for Date of Diagnosis or Date of Last Contact. One purpose of the algorithm is to standardize specification of missing month and day fields.
The information you will need includes:
- Descriptions of the eight survival variables, three of which are recoded date fields (where values are assigned for missing date components).
- Three New Data Elements for Survival with Active Followup (PDF, 67 KB)
- Three New Data Elements for Survival - Presumed Alive (PDF, 74 KB)
- Survival Date of Diagnosis Recode (PDF, 79 KB)
- Vital Status Recode - This is the vital status as of the study cutoff date. The formula is, vital status recode = vital status. If date of last contact > study cutoff date, vital status recode = alive.
- CalculateSurvivalTimeInMonths.sas (SAS, 42 KB)
- Calculation of Survival Time Fields (PDF, 534 KB)
For additional information:
If you have a question about the process, like which option should I choose, please contact your standard setter.
If you have a technical question about the survival calculation or anything else, please email email@example.com