National Institutes of Health Bethesda, Maryland, United States
Background/Case Studies: Plateletpheresis products expire quickly, therefore transfusion services rely on an accurate estimate of platelets coming into inventory. In a small hospital-based donor center, low collection days can lead to severe platelet shortages and delays in care. We evaluated a simple algorithm to predict which scheduled platelet donors were likely to have a successful donation, to anticipate and mitigate inventory shortages in a timely fashion.
Study
Design/Methods: We prospectively reviewed previous donation records for all platelet donors scheduled over a 6-month period in our center. Donors meeting the following criteria were defined as high risk for unsuccessful donation and subtracted from the total predicted collections for a given day: (1) > 3 deferrals out of a 6-donation history (2) history of prior deferrals due low hemoglobin (Hb), travel, aspirin use, or ongoing medical conditions (3) > 3 unsuccessful collections out of a 6-donation history (4) ≥ 2 no-shows or cancellations of a 3-donation history (excluding cancellations due to donor center closures) (5) recent whole blood donation (if pre-whole blood donation fingerstick Hb was ≤ 13.0) (6) first-time plateletpheresis donors (high risk of no-show, low Hb, or failing donor health questionnaire criteria). During community outbreaks of common communicable illnesses, we subtracted 1 to 2 per day from the overall estimate of successful collections. The prediction was determined to be accurate if the actual number of collections fell within ± 10% of predicted collections.
Results/Findings: From October 2024 thru March 2025, 1545 platelet donation appointments were scheduled for 523 donors. M:F 1:1, (0.4% did not identify as male or female). Donor age 19 – 30 years (2.9%), 31–50 years ( 16.6%), 51–60 years (23.7%), and 61 – 80 (10.1%) years. Overall, 69.9% of scheduled appointments resulted in successful donation, with 8% deferred, 4% technically unusable (TU), and 19% cancellations/no-shows. 97 first-time donors were scheduled, of which 37 (38%) successfully donated, 14 (14%) were deferred, 7 (7%) TU, and 38 (39%) cancellations or no-shows. First-time donors who were deferred or TU did not return within the period of study. The algorithm accurately predicted the number of successful platelet collections on 83% of days in the period of study. On days when the algorithm predicted < 50% successful collections, we recruited backup donors or purchased platelets in a timely fashion. At the end of the study period, we incorporated the algorithm into our routine platelet scheduling and inventory management.
Conclusions: Using simple criteria, we reliably predicted successful collections and hence the number of platelets going to inventory. This tool enhanced inventory management at our center by enabling better planning and resource allocation.
