Skip to article content

The mission of the Harvey L. Neiman Health Policy Institute® is to establish foundational evidence for health policy and radiology practice that promotes the effective and efficient use of health care resources and improves patient care.

ICE-T Inpatient Cost Evaluation Tool


The ICE-T contains data on total hospital reimbursement, total imaging reimbursements, and the imaging share of reimbursements across inpatient diagnosis related groups (DRG) for Medicare claims data.



The ICE-T uses data derived from the 5% Medicare Research Identifiable Files (RIF) from 2008-2013. The 5% RIF files contain detailed claims information for all Part A and Part B claims associated with a 5% national sample of the Medicare population. The specific data files used in the analysis are as follows:

The ICE-T Medicare Part A estimates for each episode were calculated using the 5% RIF Inpatient Base Claims and Inpatient Revenue Center files. These files contain the final action fee-for-service claims data submitted by inpatient hospital providers for the reimbursement of facility costs.

The ICE-T Medicare Part B estimates for each episode were calculated using the 5% RIF Carrier Claims and Carrier Lines files. These files contain the final action fee-for-service claims data submitted by providers for reimbursement separate from the inpatient and outpatient claims files.

Note: Outpatient, hospice, and skilled nursing facility claims associated with inpatient episodes have not been included in ICE-T episode costs. Practices negotiating bundled payments should rely on the imaging cost data rather than on the overall bundle share for bundles that include substantive post-acute care in the episode.



Inpatient episodes were constructed by matching Part B carrier claims to each of the Part A inpatient claims identified in the 5% inpatient files. First, the carrier claims were restricted to claims performed in the inpatient place of service. The resulting claims were then matched to the inpatient claims using the beneficiary ID and date of service listed in the carrier claims and the beneficiary ID, admission date, and discharge date listed in the inpatient claims.

All Part B carrier claims listed as being performed in the inpatient place of service from the date of admission until 90 days after the patient was discharged were considered part of the inpatient episode of care.

Note: Practices using different episode endpoints or including services not performed in the inpatient place of service will need to adjust the ICE-T benchmarks appropriately.



Costs were calculated using the total allowed charges reported on all of the claims associated with the inpatient episodes.

Part A imaging costs were calculated as the sum of allowed charges associated with the following revenue center codes indicated in the inpatient revenue center files: 0255, 0320, 0321, 0322, 0323, 0324, 0329, 0350, 0351, 0352, 0359, 0400, 0401, 0402, 0403, 0404, 0409, 0610, 0611, 0612, 0614, 0615, 0616, 0618, 0619, 0972, 0340, 0341, 0342, and 0349.

Part B imaging costs were calculated as the sum of all inpatient carrier claim allowed charges where the Berenson-Eggers Type of Service Code (BETOS) reported on the claim began with an “I”. The allowed charges were included regardless of whether the claim was filed as a professional component (PC), technical component (TC), or global claim. However, because the claims were restricted to inpatient claims there are relatively few technical and global claims within the data.



These estimates are aggregate data across the 5% national sample and will not apply to all care environments. For example, the type of practice (e.g. academic, for-profit, etc.), underlying patient populations (case mix), and/or existing geographic variation in care can cause providers to experience substantial variation from the reported aggregate data.

The ICE-T does not report data at the state level. Although Medicare data exists at the state level, there are insufficient sample sizes in the 5% RIF files for specific inpatient DRGs to ensure adequate power for reliable state estimates.

The ICE-T data is an average of all DRG claims for each year. The values generated by the ICE-T app should be compared against individual practice data, as individual practice samples may be very different from the national average.


Return to Interpretation Home