Abstract
Background/Aims Cancer researchers often use variables that evolve over time due to changes in national standards, such as AJCC (American Joint Committee on Cancer) staging or Collaborative Staging versions. The evolution of these standards impacts the construction of composite variables such as Stage AJCC in the VDW, and Extent of Disease in the local Group Health data environment. Knowledge of how cancer data registry items evolve over time and how they map into composite variables will aid researchers and programmers design cancer research studies and maintain cancer data infrastructure.
Methods We are experimenting with ways to document cancer variables that effectively illustrate their evolution over time and the relationships among them. Currently these consist of time lines that are delimited by the item number from NAACCR (North American Association of Central Cancer Registries) and annotated with other information. Time lines of items that are related under the same concept are drawn adjacent to each other. When appropriate, the diagrams demonstrate how Group Health has populated VDW and local composite variables through each period of changing standards.
Results Early response from cancer researchers has been positive. The diagrams have been especially useful in assisting in the querying of cancer counter data via a complicated NAACCR ontology (on i2b2) and aiding discussions around how Group Health should build its VDW Tumor StageGen composite SEER Summary Staging variable.
Discussion The documentation of the composition of classes of cancer variables addresses a need in the cancer research community. Continuing to explore ways to illustrate the cancer variables is a low cost and efficient way of improving understanding around these topics for a broad audience and improving data use and quality.
