关键词: 献血者, 乙型肝炎表面抗原, 模型预测

Abstract: Objective To retrospectively analyze the trend and influencing factors of hepatitis B surface antigen (HBsAg) positivity rate among blood donors in Changsha from 2020 to 2023, and to construct an autoregressive integrated moving average (ARIMA) model based on monthly data for prediction. Methods A total of 629 721 blood donors in Changsha from 2020 to 2023 were included. Demographic characteristics of HBsAg-positive donors were analyzed using χ² tests and binary logistic regression. An ARIMA model was developed with EViews 12.0 to predict HBsAg positivity rates from January to March 2024, and its performance was evaluated. Results The overall HBsAg positivity rate among blood donors was 4.6 per 1 000 (2 874/629 721), showing an initial increase followed by a decrease. Lower HBsAg positivity rates were associated with donation year 2023 (OR= 0.578), birth year ≥2002 (OR=0.654), age 18-30 years (OR=0.529), non-Han ethnicity (OR=0.734), education level of bachelor's degree or higher (OR=0.821), repeat donations (OR=0.188), apheresis platelet donations (OR=0.787), and voluntary individual donations (OR=0.883) (all P<0.05). The ARIMA(2, 1, 0) model passed validation tests, with a mean absolute percentage error (MAPE) of 12.79% for predictions from January to March 2024, indicating moderate predictive accuracy. Conclusions The HBsAg positivity rate among voluntary blood donors in Changsha remained low from 2020 to 2023. Younger donors, non-Han ethnic groups, repeat donors, individuals with higher education levels, and apheresis platelet donors exhibited lower HBsAg positivity risks. Blood collection agencies should prioritize recruiting and retaining these populations to further reduce HBsAg positivity and enhance blood safety. The ARIMA(2, 1, 0) model demonstrated moderate predictive performance; future studies should incorporate additional variables to develop more precise prediction models.

Key words: Blood donors, HBsAg, Model prediction