Methods to identify and compare parenteral nutrition administered from hospital-compounded and premixed multichamber bags in a retrospective hospital claims database

Purpose: Use of parenteral nutrition (PN) is indicated for patients who are unable to meet their needs enterally. PN may be administered via custom‐compounded mix or commercially available ready‐to‐use multichamber bags (MCB), but little is known about potential differences in clinical outcomes between these delivery systems. This study was undertaken to assess the feasibility of comparing custom‐compounded and MCB PN in a large hospital claims database. Methods: Hospital claims data from the Premier Perspective Comparative Hospital Database (PCD) reported from 2005 through 2007 were analyzed. The authors searched the data for patients who received any PN products, including compounded PN and MCB PN. Coding algorithms for identifying patient characteristics, risk factors, and outcomes of interest were explored. Results: Using hospital billing claims, the authors identified patients in the database treated with premixed PN from multichamber bags (“MCB only,” n = 4699) and patients treated with custom‐compounded PN solution (“compounded PN,” n = 64,315). Methods of identifying PN administration groups, patient characteristics and risk factors, outcomes of interest, and data limitations are described. Conclusions: Exploratory analysis suggests that comparisons of PN administered via compounding and MCB are possible using the Premier data. The ability to control for many identifiable risk factors allows data to be presented for the use of PN and related outcomes in both a clinically sensible and relevant manner, albeit with some limitations.