Collection of economic data using UB-04s: Is it worth the effort? Evidence from two clinical trials

Cost collection using UB-04 forms for economic evaluation is challenging, as UB-04 collection is time and effort intensive and compliance is imperfect. Alternative data sources could overcome those challenges. The objective of this study is to determine the usefulness of UB-04 data in estimating hospital costs compared to clinical case report form (CRF) data. Health care utilization costs were compared from financial information in UB-04s and from an assignment process using CRF data, from the WRAP-IT (23 infections) and the Micra IDE trials (61 adverse events and 108 implants). Charge-based costs were calculated by multiplying charges in UB-04s and hospital-specific Cost-to-Charge ratios from the Centers for Medicare and Medicaid Services cost reports. The cost assignment process used clinical information to find comparable encounters in real world data and assigned an average cost. Paired difference tests evaluated whether both methods yield similar results. The mean difference in total infection related costs between methods in the WRAP-IT trial was $152 +/-$22,565. In the Micra IDE trial, the mean difference in total adverse event related costs between methods was -$355 +/-$8,298 while the mean difference in total implant related costs between methods was $-3,488 +/-$13,859. Wilcoxon tests and generalized linear models could not reject the difference in costs between methods in the first two cases. Cost assignment methods achieve results similar to costs obtained through UB-04s, without the additional investment in time and effort. The use of UB-04 information for services that are not mature in a health care system may present unexpected challenges, necessitating a tradeoff with other methods of cost assignment.