中华检验医学杂志
中華檢驗醫學雜誌
중화검험의학잡지
CHINESE JOURNAL OF LABORATORY MEDICINE
2015年
5期
292-295
,共4页
曾洁%汪静%张传宝%张江涛%赵海建%刘倩%张天娇%闫颖%周伟燕%任思楣
曾潔%汪靜%張傳寶%張江濤%趙海建%劉倩%張天嬌%閆穎%週偉燕%任思楣
증길%왕정%장전보%장강도%조해건%류천%장천교%염영%주위연%임사미
血蛋白质类%置信区间%回归分析
血蛋白質類%置信區間%迴歸分析
혈단백질류%치신구간%회귀분석
Blood proteins%Confidence intervals%Regression analysis
目的:依据EP9-A3评价常规检测方法测定血清总蛋白的准确性,掌握EP9-A3建议的回归分析法与Bland-Altman法对血清总蛋白可比性结果判定的差异。方法方法评价性研究。收集46人份血清,分成4套血清盘,分别用于15种常规检测方法和参考方法。按EP9-A3建议的回归分析法计算常规检测方法与参考方法间的相关系数和直线回归方程,并以直线方程计算医学决定水平下(45、60、80 g/L)的相对偏差和95%可信区间。按经典Bland-Altman法或Carstensen改良Bland-Altman法计算常规检测方法与参考方法在医学决定水平下的相对偏差、95%一致限和95%可信区间。将2种方法得到的相对偏差和95%可信区间与生物学变异总误差标准5%进行比较,若都在±5%以内则在该医学决定水平下结果可比。结果(1)按回归分析法,15种常规检测方法和参考方法间的相关都>0.975( P<0.001);在医学决定水平下,15种常规检测方法和参考方法间的相对偏差和95%可信区间都在±5%以内。(2)经典和Carstensen改良Bland-Altman法处理,1种检测方法无法处理,其余14种检测方法的相对偏差和95%可信区间都在±5%以内。结论回归分析法和Bland-Altman法得到的相对偏差大小和95%可信区间大小都较一致性,15种常规血清总蛋白检测方法与参考方法在医学决定水平下检测结果差异<5%。(中华检验医学杂志,2015,38:292-295)
目的:依據EP9-A3評價常規檢測方法測定血清總蛋白的準確性,掌握EP9-A3建議的迴歸分析法與Bland-Altman法對血清總蛋白可比性結果判定的差異。方法方法評價性研究。收集46人份血清,分成4套血清盤,分彆用于15種常規檢測方法和參攷方法。按EP9-A3建議的迴歸分析法計算常規檢測方法與參攷方法間的相關繫數和直線迴歸方程,併以直線方程計算醫學決定水平下(45、60、80 g/L)的相對偏差和95%可信區間。按經典Bland-Altman法或Carstensen改良Bland-Altman法計算常規檢測方法與參攷方法在醫學決定水平下的相對偏差、95%一緻限和95%可信區間。將2種方法得到的相對偏差和95%可信區間與生物學變異總誤差標準5%進行比較,若都在±5%以內則在該醫學決定水平下結果可比。結果(1)按迴歸分析法,15種常規檢測方法和參攷方法間的相關都>0.975( P<0.001);在醫學決定水平下,15種常規檢測方法和參攷方法間的相對偏差和95%可信區間都在±5%以內。(2)經典和Carstensen改良Bland-Altman法處理,1種檢測方法無法處理,其餘14種檢測方法的相對偏差和95%可信區間都在±5%以內。結論迴歸分析法和Bland-Altman法得到的相對偏差大小和95%可信區間大小都較一緻性,15種常規血清總蛋白檢測方法與參攷方法在醫學決定水平下檢測結果差異<5%。(中華檢驗醫學雜誌,2015,38:292-295)
목적:의거EP9-A3평개상규검측방법측정혈청총단백적준학성,장악EP9-A3건의적회귀분석법여Bland-Altman법대혈청총단백가비성결과판정적차이。방법방법평개성연구。수집46인빈혈청,분성4투혈청반,분별용우15충상규검측방법화삼고방법。안EP9-A3건의적회귀분석법계산상규검측방법여삼고방법간적상관계수화직선회귀방정,병이직선방정계산의학결정수평하(45、60、80 g/L)적상대편차화95%가신구간。안경전Bland-Altman법혹Carstensen개량Bland-Altman법계산상규검측방법여삼고방법재의학결정수평하적상대편차、95%일치한화95%가신구간。장2충방법득도적상대편차화95%가신구간여생물학변이총오차표준5%진행비교,약도재±5%이내칙재해의학결정수평하결과가비。결과(1)안회귀분석법,15충상규검측방법화삼고방법간적상관도>0.975( P<0.001);재의학결정수평하,15충상규검측방법화삼고방법간적상대편차화95%가신구간도재±5%이내。(2)경전화Carstensen개량Bland-Altman법처리,1충검측방법무법처리,기여14충검측방법적상대편차화95%가신구간도재±5%이내。결론회귀분석법화Bland-Altman법득도적상대편차대소화95%가신구간대소도교일치성,15충상규혈청총단백검측방법여삼고방법재의학결정수평하검측결과차이<5%。(중화검험의학잡지,2015,38:292-295)
Objective To evaluate the difference of Doumas′method and 15 commercial serum total protein ( TP ) methods based on EP9-A3.Methods Serum panels were quantified for TP with Doumas′method and measured in parallel with 15 commercial methods.The linear regression analyses were performed, followed by calculating relative deviation and 95%CI between commercial method and Doumas′method at three different medical decision levels (45 g/L, 60 g/L, 80 g/L).We also calculated relative deviation, 95% limit of agreement ( LoA ) and 95% CI based on classical and improved Bland-Altman method at three different medical decision levels.If both the relative deviation and 95%CI were within 5%, we conside red the commercial serum total protein method was comparable to Doumas′method.Results (1) All assays presented high correlation ( r>0.975, P<0.001) with the Doumas′method.All assays showed that the relative deviations and 95%CIs were within the biological total error goal (5%) at medical decision levels based on regression analysis.(2) Based on classical and improved Bland-Altman method, fourteen of 15 commercial methods showed that the relative deviations and 95%CIs were within +/-5%. Conclusions All commercial assays are comparable to Doumas′method at medical deviation levels.There is no difference between regression analysis and Bland-Altman method for comparison study.