重复测量试验对同一受试对象进行多次测量,各时间点数据间存在自相关性,进行方差分析和均值比较时需要进行特殊处理。虽然此方法在农业等研究领域运用十分广泛,但目前有效地相关统计方法鲜见。为了建立操作简单、实用性强、结果可靠的...重复测量试验对同一受试对象进行多次测量,各时间点数据间存在自相关性,进行方差分析和均值比较时需要进行特殊处理。虽然此方法在农业等研究领域运用十分广泛,但目前有效地相关统计方法鲜见。为了建立操作简单、实用性强、结果可靠的统计分析方法,本研究采用SAS的广义线性混合模型(Generalized Linear Mixed Models,GLIMMIX),以随机区组重复测量试验资料为例,说明了协方差结构筛选、方差分析和均值比较的具体方法。结果表明,用传统的裂区设计、多变量统计等方法会造成资料信息浪费,统计功效降低,缺区无法处理等问题,甚至会导致错误的结论。GLIMMIX能很好地处理自相关问题,功能强大,结果可靠,使用简单,允许缺区,是进行重复测量试验资料方差分析和均值比较的理想方法。目前在国内将其运用到农学类试验数据的统计分析的相关报道鲜见,该文在本领域具有很强的实用性和创新性。展开更多
Background Severity scoring systems are useful tools for measuring the severity of the disease and its outcome. This pilot study was to verify and compare the prognostic performance of the Simplified Acute Physiology ...Background Severity scoring systems are useful tools for measuring the severity of the disease and its outcome. This pilot study was to verify and compare the prognostic performance of the Simplified Acute Physiology Score II （SAPS II） and Glasgow Coma Scale （GCS） in neuro-intensive care unit （N-ICU） patients. Methods A total of 1684 patients consecutively admitted to the N-ICU at Xuanwu Hospital between January 1, 2005 and December 31, 2011 were enrolled in this study. The data-base included admission data, at 24-, 48-, and 72-hour SAPS II and GCS. Repeated measure data analysis of variance, Logistic regression analysis, the Hosmer-Lemeshow goodness-of-fit statistic, and the area under the receiver operating characteristic were used to evaluate the performance. Results There was a significant difference between the SAPS II or GCS score at four time points （F=16.110, P=0.000 or F=8.108, P=0.000）. The SAPS II scores or GCS score at four time points interacted with the outcomes with significant difference （F=116.771, P=0.000 or F=65.316, P=0.000）. Calibration of the SAPS II or GCS score at each time point on all patients was good. The percentage of a risk estimate prediction corresponding to observed mortality was also good. The 72-hour score have the greatest consistency. Discriminations of the SAPS II or GCS score at each time were all satisfactory. The 72-hour score had the greatest discriminative power. The cut-off value was 33 （sensitivity of 85.2% and specificity of 74.3%） and 6 （sensitivity of 70.6% and specificity of 65.0%）. The SAPS II at each time point on all patients showed better calibration, consistency and discrimination than GCS. The binary Logistic regression analysis identified physiological variables, GCS, age, and disease category as significant independent risk factors of death. After the two variables including underlying disease and type of admission were excluded, we built the simplified SAPS II model. A correlation was suggested between the simplified SAPS II sco展开更多
文摘重复测量试验对同一受试对象进行多次测量,各时间点数据间存在自相关性,进行方差分析和均值比较时需要进行特殊处理。虽然此方法在农业等研究领域运用十分广泛,但目前有效地相关统计方法鲜见。为了建立操作简单、实用性强、结果可靠的统计分析方法,本研究采用SAS的广义线性混合模型(Generalized Linear Mixed Models,GLIMMIX),以随机区组重复测量试验资料为例,说明了协方差结构筛选、方差分析和均值比较的具体方法。结果表明,用传统的裂区设计、多变量统计等方法会造成资料信息浪费,统计功效降低,缺区无法处理等问题,甚至会导致错误的结论。GLIMMIX能很好地处理自相关问题,功能强大,结果可靠,使用简单,允许缺区,是进行重复测量试验资料方差分析和均值比较的理想方法。目前在国内将其运用到农学类试验数据的统计分析的相关报道鲜见,该文在本领域具有很强的实用性和创新性。
文摘Background Severity scoring systems are useful tools for measuring the severity of the disease and its outcome. This pilot study was to verify and compare the prognostic performance of the Simplified Acute Physiology Score II （SAPS II） and Glasgow Coma Scale （GCS） in neuro-intensive care unit （N-ICU） patients. Methods A total of 1684 patients consecutively admitted to the N-ICU at Xuanwu Hospital between January 1, 2005 and December 31, 2011 were enrolled in this study. The data-base included admission data, at 24-, 48-, and 72-hour SAPS II and GCS. Repeated measure data analysis of variance, Logistic regression analysis, the Hosmer-Lemeshow goodness-of-fit statistic, and the area under the receiver operating characteristic were used to evaluate the performance. Results There was a significant difference between the SAPS II or GCS score at four time points （F=16.110, P=0.000 or F=8.108, P=0.000）. The SAPS II scores or GCS score at four time points interacted with the outcomes with significant difference （F=116.771, P=0.000 or F=65.316, P=0.000）. Calibration of the SAPS II or GCS score at each time point on all patients was good. The percentage of a risk estimate prediction corresponding to observed mortality was also good. The 72-hour score have the greatest consistency. Discriminations of the SAPS II or GCS score at each time were all satisfactory. The 72-hour score had the greatest discriminative power. The cut-off value was 33 （sensitivity of 85.2% and specificity of 74.3%） and 6 （sensitivity of 70.6% and specificity of 65.0%）. The SAPS II at each time point on all patients showed better calibration, consistency and discrimination than GCS. The binary Logistic regression analysis identified physiological variables, GCS, age, and disease category as significant independent risk factors of death. After the two variables including underlying disease and type of admission were excluded, we built the simplified SAPS II model. A correlation was suggested between the simplified SAPS II sco
