Purpose: The liver simulated allocation model (LSAM) is a discrete-event microsimulation of the liver allocation system used for policy evaluation and research. We investigated the calibration of a newly updated version of LSAM by donor service area (DSA).

Methods: We used the updated LSAM to simulate current liver allocation policy 7/1/2013-6/30/2016, then performed linear regression analysis to compare the simulation results with observed data from the same period by DSA. We show two approaches to grouping organs by DSA: all organs recovered within each DSA (regardless of recipient location, i.e., a procurement organization perspective), and all organs transplanted within each DSA (regardless of donor location, i.e., a transplant center perspective).

Results: Figure 1 shows example plots for transplant counts, and Table 1 presents slope and intercept terms for additional metrics. Perfect calibration between LSAM and reality would give a regression slope of 1 and a y-intercept of 0. We found small intercepts for all metrics tested and slopes of close to one for all except local transplant percentage.

Discussion: Liver transplant metrics can vary widely across DSAs; LSAM models some sources of this variation (geographic distribution of donors and candidates) but not others (institutional differences in offer acceptance). Local transplant percentage is not well calibrated, perhaps because it combines error from two estimated quantities (local transplant and total transplant counts). For other metrics of interest, LSAM produces well-calibrated liver transplant projections by DSA despite not modeling institutional differences in offer acceptance behavior.

CITATION INFORMATION: Pyke J., Weaver T., Schladt D., Salkowski N., Wey A., Lake J., Kim W., Gentry S., Israni A. Calibration of Updated LSAM Am J Transplant. 2017;17 (suppl 3).

« Back to 2018 American Transplant Congress