在职研究生医学统计学作业一峰王.docx

1、在职研究生医学统计学作业一峰王2013级南昌大学医学院在职研究生(南昌半脱产)医学统计学作业姓名： 学号： 得分：题目：以某治疗胃炎疼痛症状的临床试验数据库为例，评价A和B两种治疗胃炎疼痛症状药物的疗效是否有差别。该临床试验采用完全随机设计，数据库包括人口学特征、体检及基线指标(共13项)和疗效指标(2项)，其中，人口学特征、体检及疗前基线指标包括：年龄、性别、呼吸、心率、收缩压、舒张压、体温、身高、体重、病程、过敏史、治疗史、治疗前胃部疼痛评分；疗效指标包括：治疗前后胃部疼痛评分差值、综合疗效(痊愈、显效、有效、无效)。要求：1）拟一份统计分析计划(字数不限)，内容包括分析哪些指标来评价A和

2、B两种药物疗效的差别、采用何种统计指标或方法分析这些指标。 答：在此次实验中实验者随机抽取了100例实验对象并采用了完全随机设计方法来评价A和B两种治疗胃炎疼痛症状药物的疗效是否有差别。因此首先应该判定随机分组的两组病人各项人口学特征、体检及基线指标的可比性分析；分析判定各项指标是否具有可比性；分析各项指标的类别并选用不同的统计学检验方法描述并评价各项指标及疗效。统计分析计划：通过分析各项指标的类别及可比性后，对治疗前后胃部疼痛评分差值、综合疗效分别进行t检验或单因素方差分析及秩和检验分析A、B两组药物治疗效果有无差别，选取当P0.05时，认为两组药物在疗效上有统计学差异，当P0.05时，认为

3、两组药物在疗效上无显著性差异。2）对每一个须分析指标进行统计描述和推断，列出主要的描述性指标及统计推断结果（检验统计量及P值）。 答：（一）、各处理因素的统计描述：1、根据题目要求创建两个纵列分别命名为“group1”、“d”；2、各项指标分别绘制成P-P图分析数据分布形态；经统计SPSS软件绘制出的P-P图，我们可以得出以下结论：符合近似正态分布的计数指标有：age、heart、breathe、dp、temp、weight、height、tt_1、d；另有sp、bc不符合正态分布。2、利用统计SPSS软件绘制图标描述各项指标：各项计数/量指标的均数、方差、标准差、四分位数间距、中位数表格；

4、Statisticsgroup ageheartbreathespdptempweightheightbctt_1dANValid4949494949494949494949Missing00000000000Mean45.4576.5119.02117.6576.2036.56359.79164.7127.4316.8611.6939Median47.0078.0019.00120.0076.0036.50060.00165.0014.0015.0011.0000Std. Deviation12.0758.7781.6018.0598.818.21389.8586.25247.1855.40

5、85.60953Percentiles2536.0070.5018.00110.0070.0036.40052.50160.007.0012.008.00005047.0078.0019.00120.0076.0036.50060.00165.0014.0015.0011.00007554.0080.0020.00120.0080.0036.80065.50169.0024.0020.5016.0000BNValid5050505050505050505050Missing00000000000Mean46.4872.8619.24116.1075.0036.53258.06162.5226.

6、3018.5013.1600Median45.5073.5020.00120.0075.0036.50059.00161.5013.0019.0013.0000Std. Deviation12.4307.2171.3939.5526.350.18348.0826.37738.7336.3295.65093Percentiles2536.0069.5018.00110.0070.0036.40050.00158.008.5012.758.75005045.5073.5020.00120.0075.0036.50059.00161.5013.0019.0013.00007557.2579.0020

7、.00120.0080.0036.70062.25168.0024.0024.2517.2500标准差描述，不符合近似正态分布指标用中位数及四分位数间距描述；ageheartbreathedptempweightheighttt_1dA均数45.4576.5119.02117.6576.2036.56359.79164.7127.43B均数46.4872.8619.24116.1075.0036.53258.06162.5226.30A标准差12.0758.7781.6018.0598.818.21389.8586.25247.185B标准差12.4307.2171.3939.5526.350

8、.18348.0826.37738.733A组中位数B组中位数A组四分位数间距B四分位数间距Sp120.00120.00120-110=10120-110=10bc14.0013.0024.00-8.25=15.7524.00-9.00=15.00分类变量指标的描述： Case Processing Summary CasesValidMissingTotalNPercentNPercentNPercentgroup_1 * gender99100.0%0.0%99100.0%group_1 * gms99100.0%0.0%99100.0%group_1 * zls99100.0%0.0%9

9、9100.0%group_1 * zylx99100.0%0.0%99100.0% group_1 * gender Crosstabulation genderTotal12group_11.00Count222749% within group_144.9%55.1%100.0%2.00Count163450% within group_132.0%68.0%100.0%TotalCount386199% within group_138.4%61.6%100.0% group_1 * gms Crosstabulation gmsTotal12group_11.00Count49049%

10、 within group_1100.0%.0%100.0%2.00Count44650% within group_188.0%12.0%100.0%TotalCount93699% within group_193.9%6.1%100.0% group_1 * zls Crosstabulation zlsTotal12group_11.00Count40949% within group_181.6%18.4%100.0%2.00Count401050% within group_180.0%20.0%100.0%TotalCount801999% within group_180.8%

11、19.2%100.0% group_1 * zylx Crosstabulation zylxTotal1234group_11.00Count22321349% within group_14.1%46.9%42.9%6.1%100.0%2.00Count52123150% within group_110.0%42.0%46.0%2.0%100.0%TotalCount74444499% within group_17.1%44.4%44.4%4.0%100.0%（二）、实验统计的推断分析：1、利用t检验或单因素方差齐性检验分析各计数指标是否方差齐；SPSS统计软件对计数指标进行t检验可得

12、下表： Levenes Test for Equality of Variancest-test for Equality of MeansFSig.tdfSig. (2-tailed)Mean DifferenceStd. Error Difference95% Confidence Interval of the DifferenceLowerUpperageEqual variances assumed.268.606-.41997.676-1.0312.464-5.9203.858Equal variances not assumed -.41996.993.676-1.0312.46

13、3-5.9193.857genderEqual variances assumed5.549.021-1.31897.191-.129.098-.323.065Equal variances not assumed -1.31796.310.191-.129.098-.323.065heartEqual variances assumed.384.5372.26297.0263.6501.614.4486.853Equal variances not assumed 2.25892.777.0263.6501.617.4396.861breatheEqual variances assumed

14、.005.946-.72997.468-.220.301-.818.379Equal variances not assumed -.72794.635.469-.220.302-.819.380dpEqual variances assumed.102.750.78197.4371.2041.542-1.8564.265Equal variances not assumed .77887.143.4381.2041.547-1.8714.279tempEqual variances assumed3.095.082.78197.436.0313.0400-.0481.1107Equal va

15、riances not assumed .78094.209.437.0313.0401-.0483.1108weightEqual variances assumed2.341.129.95697.3421.7301.810-1.8635.322Equal variances not assumed .95492.672.3431.7301.814-1.8725.332heightEqual variances assumed.215.6441.72997.0872.1941.269-.3254.714Equal variances not assumed 1.72997.000.0872.

16、1941.269-.3254.713tt_1Equal variances assumed2.804.097-1.38797.169-1.6431.184-3.993.708Equal variances not assumed -1.38995.247.168-1.6431.182-3.990.704dEqual variances assumed.031.860-1.29597.198-1.466121.13183-3.71249.78024Equal variances not assumed -1.29596.983.198-1.466121.13174-3.71232.78008 上

17、表可知 heart 指数的P值为0.025=0.05，故选用协方差对其进行校正分析。heart 指数协方差如下：Tests of Between-Subjects EffectsDependent Variable: d SourceType III Sum of SquaresdfMean SquareFSig.Corrected Model98.274(a)249.1371.541.219Intercept49.031149.0311.538.218heart37.525137.5251.177.281group181.496180.4962.528.115Error3060.474963

18、1.880 Total18615.00099 Corrected Total3158.74798 a R Squared = .031 (Adjusted R Squared = .011)从上表可知：经协方差校正后 heart 指数的F值为1.541，P值为0. 219=0.05。综上所述经t检验后可知：age、heart、breathe、dp、temp、weight、height、tt_1、d 的P值均=0.05，按=0.05水准，可认为A、B两组资料方差齐；A、B两组药物治疗前后胃部疼痛评分差值 d 无统计学差别。2、无序分类变量需选用卡方检验分析：利用SPSS统计软件对无序分类变量ge

19、nder、gms、zls 进行卡方检验得下表有： Case Processing Summary CasesValidMissingTotalNPercentNPercentNPercentgroup_1 * gender99100.0%0.0%99100.0%group_1 * gms99100.0%0.0%99100.0%group_1 * zls99100.0%0.0%99100.0%group_1 * zylx99100.0%0.0%99100.0%group_1 * gender Crosstab genderTotal12group_11.00Count222749Expected

20、 Count18.830.249.02.00Count163450Expected Count19.230.850.0TotalCount386199Expected Count38.061.099.0 Chi-Square Tests ValuedfAsymp. Sig. (2-sided)Exact Sig. (2-sided)Exact Sig. (1-sided)Point ProbabilityPearson Chi-Square1.741(b)1.187.218.133 Continuity Correction(a)1.2381.266 Likelihood Ratio1.746

21、1.186.218.133 Fishers Exact Test .218.133 Linear-by-Linear Association1.723(c)1.189.218.133.070N of Valid Cases99 a Computed only for a 2x2 tableb 0 cells (.0%) have expected count less than 5. The minimum expected count is 18.81.c The standardized statistic is 1.313.group_1 * gms Crosstab gmsTotal1

22、2group_11.00Count49049Expected Count46.03.049.02.00Count44650Expected Count47.03.050.0TotalCount93699Expected Count93.06.099.0 Chi-Square Tests ValuedfAsymp. Sig. (2-sided)Exact Sig. (2-sided)Exact Sig. (1-sided)Point ProbabilityPearson Chi-Square6.259(b)1.012.027.014 Continuity Correction(a)4.3291.

23、037 Likelihood Ratio8.5771.003.027.014 Fishers Exact Test .027.014 Linear-by-Linear Association6.196(c)1.013.027.014.014N of Valid Cases99 a Computed only for a 2x2 tableb 2 cells (50.0%) have expected count less than 5. The minimum expected count is 2.97.c The standardized statistic is 2.489.group_

24、1 * zls Crosstab zlsTotal12group_11.00Count40949Expected Count39.69.449.02.00Count401050Expected Count40.49.650.0TotalCount801999Expected Count80.019.099.0 Chi-Square Tests ValuedfAsymp. Sig. (2-sided)Exact Sig. (2-sided)Exact Sig. (1-sided)Point ProbabilityPearson Chi-Square.043(b)1.8371.000.520 Continuity Correction(a).00011.000 Likelihood Ratio.0431.8371.000.520 Fishers Exact Test 1.000.520 Linear-by-Linear Association.042(c)1.8371.000.520.197N of Valid Cases99 a Computed only for a 2x2 tableb 0 cells (.0%) have expected count less than 5. The minimum expected count is 9.40.c The