Director, Blood Bank Department of Pathology, Michigan Medicine at the University of Michigan Ann Arbor, Michigan, United States
Background/Case Studies: In 2019, a large, regional blood supplier experienced critically low component inventory levels > 80% of days. Similarly, a 2021 national survey found ~60% of Transfusion physicians reported recent platelet shortages. Our 1,107 bed, tertiary-care, level 1 trauma center lacks a blood donation center; Inventory is dependent on partnership with one national blood supplier. Our center experiences frequent critical platelet shortages [CPS, ≤ 10 units, 2.53 standard deviations, SD, below weekday mean], threatening transfusion-dependent patients on high-acuity solid organ transplant, hematology-oncology, obstetrics, and trauma services. Factors contributing to CPS were poorly understood. A strategic approach to CPS inventory management was lacking.
Study
Design/Methods: Retrospective ~hourly inventory was gathered from 2/2022-2/2024 (Tableau Professional v2023.1.5, Salesforce Inc., San Francisco, CA). Dates with < 10 measurements were excluded. CPS were correlated with potentially contributing factors, including Day -1 (D-1) and D0 midnight inventory and total daily utilization, approximated by daily peak-to-nadir inventory delta (Prism v10.2.2, GraphPad, La Jolla, CA). Overall mean peak-to-nadir deltas for weekdays v. weekends were approximated by midnight to 13:00 deltas. Results are reported as mean (SD) platelet units, unless otherwise noted.
Results/Findings: Weekday and weekend inventory was 59.5 (19.6) and 58.7 (18.6). Overall daily inventory peak and nadir were 78.9 (16.9) and 41.5 (17.0), with mean weekday peak inventory at midnight and mean nadir at 13:00, respectively. Clinically impactful inventory variation was not noted among weekdays (data not shown). Six v. two CPS occurred on weekdays v. weekends, respectively (Figure A). During each CPS, inventory reached ≤ 10 units between 08:00 and 16:00, increasing above 10 within four hours, excepting one event. Between 12/27-12/28/2022, inventory reached 10 units at 12:00 (12/27/2022), nadired at five units, and did not exceed 10 units until 12/28/2022 at 02:00. CPS correlated with D-1 and D0 midnight inventory > 1 SD below mean (n=5/8, 62.5% and n=6/8, 75.0%, respectively). CPS co occurred with D-1 and D0 utilization > 1 SD above mean in n=1/8 (12.5%) and n=3/8 (37.5%) of events. For each CPS, platelet supply remained below mean for the duration of the calendar day. A job aid was developed to prompt ad hoc platelet ordering to return inventory to mean. Since implementation, no further CPS have occurred. Conclusions: CPS, as expected, correlated with decreased inventory ( > 1 SD below mean) and at times co occurred with increased utilization. During CPS, baseline ad hoc platelet ordering alleviated CPS without returning inventory to mean. Following implementation of an inventory management tool prompting ordering to return inventory to mean, no further CPS have occurred.
.jpg "Jensyn Cone Sullivan, MD photo")