Assessing the Quality of SK&A's Office-Based Physician Database for Identifying Oncologists.

Health services research increasingly uses commercial databases that capture provider practice characteristics. Little is known about how these data sets compare along other dimensions with publicly available data. We assess the quality of one of the most commonly used commercial databases, SK&A's office-based physician database, for capturing oncologists who bill the Medicare fee-for-service program. Using 2017 data, we find that nearly 74% of the oncologists in Medicare claims can be found in the SK&A data. Weighted by patients, service volume, or spending, match rates increase to 77%, 96%, and 92%, respectively. Matched oncologists have a high concordance (above 95%) on subspecialty as well as contact information other than street address. Oncologists who appear only in Medicare tend to have low service volumes and spending relative to those who are matched while over half of oncologists who appear only in SK&A have a pediatric subspecialty.