Forecasting Blood Supply in Chinese Major Cities by Fractional Grey Prediction Model and Linear Regression Model

Abstract

Background and Objectives: The aim of this study is to predict the quantity of blood donations in major Chinese cities from 2023 to 2026. Blood is a scarce and perishable resource, and its availability in a particular area depends on the amount of blood donated in that area. Therefore, it is essential for healthcare providers and policymakers to have accurate predictions of future blood donation quantities in order to plan for transfusion treatments and operations. This study aims to provide insights into the future blood supply and allocation in major Chinese cities. Methods: We used a combination of linear regression and fractional grey prediction models to predict blood donation quantities in major Chinese cities from 2023 to 2026. The models were developed based on historical data of blood donation quantity, resident population, and gross domestic production. We compared the predicted values from the models to cross-check the consistency of the predictions. When they were not consistent, the predictions based on the recent historical data were chosen. We cross-checked the predictions with historical data to assess their accuracy. Results: We found that the predicted future donation quantity in major Chinese cities from 2023 to 2026 would be around $20\pm3$U per 1000 persons, and the models suggested that these predictions were credible in at least half of the cities examined. We also observed a slightly larger differential order of the grey prediction model in the cities of northeast China compared to the rest of the cities. Furthermore, we found that the reliability of the predictions varied, with over half of the cities showing consistent predictions and others showing inconsistent or unreliable results. Conclusion: Our findings provide important insights for healthcare providers and policymakers to plan for blood supply and allocation in the future. However, the reliability of the predictions varied, indicating the need for updating historical data to improve the accuracy and reliability of prediction models. We also find that in northeast of China, the blood donation quantity might be bias from 20U per 1000 persons after 2026, according to the differential order of the fractional grey prediction model. Overall, our study highlights the importance of accurate predictions of blood donation quantities for effective healthcare planning and resource allocation.