Big data modeling to predict platelet usage and minimize wastage in a tertiary care system

Area of Science:

• Health Care Management
• Biostatistics
• Hematology

Background:

• Maintaining an adequate blood product supply is critical for patient care.
• Platelet products face significant wastage due to unpredictable demand and short shelf life.
• Current inventory management struggles with the variability and cost of platelet products.

Purpose of the Study:

• To develop and validate a statistical model for forecasting platelet demand.
• To minimize platelet wastage due to expiration while preventing shortages.
• To optimize blood product collection and ordering strategies.

Main Methods:

• Investigated institutional platelet usage patterns over 29 months.
• Analyzed relationships between platelet usage and hospital-wide patient data.
• Employed a convex statistical formulation incorporating weekday/weekend patterns, complete blood count data, and hospital census.

Main Results:

• Platelet usage is significantly correlated with weekday/weekend patterns, abnormal CBC counts, and hospital census data.
• The developed mathematical model successfully minimized expiration rates from 10.5% to 3.2%.
• The model maintained a minimum of 10 platelet units daily, avoiding shortages.

Conclusions:

• A validated statistical model can effectively forecast platelet demand.
• Implementing this model can substantially reduce blood product wastage and associated costs.
• Potential national savings estimated at $80 million annually if widely adopted.