Initially, Medicare Provider Utilization and Payment Data files, which contain 100% of claims for Part B non-institutional services and procedures to Medicare fee-for-service beneficiaries, were used to identify all HCPCS codes billed in 2012 and 2013 for which at least one claim was filed by a radiologist (defined herein as any physician listed in the CMS database as a diagnostic radiologist, interventional radiologist, or nuclear medicine physician). All identified codes having a Berenson-Eggers Types of Service (BETOS) code of I (for “imaging”) were selected for further evaluation. Among these, those not corresponding with imaging professional service (e.g., codes related to contrast administration or to radiation oncology services) were excluded. NITOS categories were assigned to all remaining codes by a panel of practicing radiologists familiar with physician service procedural coding.
Subsequently, all HCPCS codes billed by a radiologist in 5% sample Standard Analytic Files from 2004 through 2011 were identified, and NITOS categorization was assigned in a similar manner.
Next, all HCPCS codes identified as BETOS I in Physician/Supplier Procedure Summary (PSPS) Master Files (which contain 100% of billed Medicare Part B fee-for-service claims) from 2003 through 2014, that were not already NITOS-mapped were identified. These were categorized by the panel in a similar manner.
And, finally, because CMS first added BETOS identifiers to the PSPS files in 2003, a non-BETOS methodology was utilized to map HCPCS codes to NITOS categories from 1994 through 2002. A final search was performed for all HCPCS codes billed by radiologists from 1994 to 2002 within PSPS Master Files within the 70000 series of HCPCS. These generally correspond to imaging services. This last group of codes was similarly NITOS categorized by the expert panel.
The final resulting file provides NITOS mapping for all HCPCS codes corresponding with all Medicare fee-for-service imaging professional services billed by all radiologists from 1994 through 2018.