摘要
目的建立大连市年度临床供血量数据预测模型,为制定献血者招募计划和血液成分制备计划提供科学依据。方法应用SPSS 22.0中灰色预测模型GM(1,1),对2009—2018年大连市红细胞、新鲜冰冻血浆、血小板和冷沉淀年度临床供血量进行统计累加,产生一定规律的序列,得到原始序列的估计值,建立微分方程即预测数据模型;回代预测2009—2018年各血液成分的临床用量,进行拟合精度检验。结果红细胞临床供血量预测模型为x^(1)(t+1)=2823845.09e0.034t−2742150.09,2009—2018年红细胞供血量的标准差Sx=14650.03,误差数列e(t)的标准差Se=2343.55,后验差比值C=0.15997,所有|e(t)−e¯|均<0.6745Sx=9881.45,小误差概率P=1.00。血小板临床用供血量预测模型为x^(1)(t+1)=58875.37e0.1t—54210.37,2009—2018年血小板供血量的标准差Sx=3064.69,误差数列e(t)的标准差Se=444.06,后验差比值C=0.14490,所有|e(t)−e¯|均<0.6745,Sx=2067.13,即小误差概率P=1.00。新鲜冰冻血浆临床供血量预测模型为x^(1)(t+1)=3535926.55e0.273t—3279035.55,2009—2018年新鲜冷冻血浆供血量的标准差Sx=2675310.31,误差数列e(t)的标准差Se=888183.60,后验差比值C=0.33199,其中1个|e(t)−e¯|≥0.6745Sx=1804496.80,P=0.89。冷沉淀临床用供血量预测模型为x^(1)(t+1)=272229.72e0.09t—256499.72,2009—2018年冷沉淀供血量的标准差Sx=11926.45,误差数列e(t)的标准差Se=1694.53,后验差比值C=0.14208,所有|e(t)−e¯|均<0.6745Sx=8044.39,即小误差概率P=1.00。结论采用灰色预测模型GM(1,1)建立预测大连市年度各种血液成分临床供血量预测模型,拟合效果十分理想,为大连市开展无偿献血招募和血液成分制备提供科学依据。
Objective To establish the prediction model of annual clinical blood supply-demand in Dalian city in order to provide scientific basis for organizing blood donor recruitment and making blood component preparation plan.Methods The grey prediction model GM(1,1)in SPSS 22.0 was used to statistically accumulate the annual clinical blood supply of red blood cells(RBC),fresh frozen plasma(FFP),platelets and cryoprecipitate in Dalian from 2009 to 2018,generated a certain regular sequence,obtained the estimated value of the original sequence,and then established the differential equation,so called the prediction data model.The model was used to predict the clinical dosage of each blood component from 2009 to 2018 and analyzed with the fitting accuracy test.Results The prediction model of RBC clinical dosage was x^(1)(t+1)=2823845.09e0.034t-2742150.09.The standard deviation of clinical RBC blood supply was 14650.03,the standard deviation of error sequence e(t)was 2343.55,the posterior error ratio was 0.15997,all the values of|e(t)−e¯|were less than 0.6745,Sx was equal to 9881.45,and the small error probability was equal to 1 from 2009 to 2018.The prediction model of platelet clinical dosage was x^(1)(t+1)=58875.37e0.1t—54210.37.The standard deviation of clinical platelets dosage was 3064.69,the standard deviation of error sequence e(t)was 444.06,the posterior error ratio was 0.14490,all the values of|e(t)−e¯|<0.6745,Sx was equal to 2067.13,and the small error probability was equal to 1 from 2009 to 2018.The standard deviation of clinical FFP dosage was 2675310.31,the standard deviation of error sequence e(t)was 888183.60,the posterior error ratio was 0.33199.One of the|e(t)−e¯|values was greater than 0.6745,Sx was equal to 1804496.80,and the small error probability was equal to 0.89 from 2009 to 2018.The prediction model of FFP clinical dosage was x^(1)(t+1)=3535926.55e0.273t—3279035.55.The prediction model of clinical cryoprecipitate dosage was x^(1)(t+1)=272229.72e0.09t—256499.72.The standard deviation of clinical cryoprecipitate dosage was 11926.45,the standard deviation of error sequence e(t)was 1694.53,the posterior error ratio was 0.14208,all the values of|e(t)−e¯|were less than 0.6745,Sx was equal to 8044.39,and the small error probability was equal to 1 from 2009 to 2018.Conclusion The prediction model of annual clinical blood supply of various blood components created by the gray prediction model GM(1,1)was very useful to provided scientific basis for the recruitment of voluntary blood donors and the preparation of blood components when evaluated by the fitting effect test.
作者
毕晓琳
孙嘉良
范亚欣
苏明慧
李玉
王亮
BI Xiaolin;SUN Jialiang;FAN Yaxin;SU Minghui;LI Yu;WANG Liang(Dalian Blood Center,Dalian,116001,China)
出处
《临床血液学杂志》
CAS
2023年第4期279-283,共5页
Journal of Clinical Hematology
关键词
年度供血量
灰色预测模型GM(1
1)
统计分析
供血计划
预测数据模型
annual blood supply quantity
grey prediction model GM(1
1)
statistical analysis
blood supply plan
prediction data model
