The validity of hospital administrative data in monitoring variations in breast cancer surgery.

To assess the validity of using hospital administrative data to measure variations in surgery for early-stage breast cancer, ICD-9-CM coded information was compared with corresponding tumor registry data for 1293 breast cancer patients undergoing lumpectomy or mastectomy at a tertiary referral center from January 1989 to October 1993. Relative to "gold standard" tumor registry data, the administrative data proved 83.4% sensitive and 80.4% specific in identifying women with localized disease who would be potential candidates for lumpectomy. The proportion of women with localized disease undergoing lumpectomy in groups defined by race and insurance status was nearly identical, whichever data were used. Administrative data, which is often readily and publicly available, may be useful in studying variations in breast cancer treatment in key demographic groups.