Stata运行结果及分析

log: e:赵耐青统计分析结果.log

log type: text

opened on: 20 Jul 2002, 10:47:03

.

. /* 调用数据库 */

. use "E:赵耐青clinical trial data(Chenfeng).dta", clear

.

. /* 描述数据库 */

. describe

Contains data from E:赵耐青clinical trial data(Chenfeng).dta

obs: 240

vars: 60 7 Feb 2002 21:41

size: 35,760 (87.5% of memory free)

-------------------------------------------------------------------------------

storage display value

variable name type format label variable label

-------------------------------------------------------------------------------

No int %8.0g 编号

Center byte %8.0g 中心编号

Name str3 %9s 姓名

Gender byte %8.0g sex 性别

Age byte %8.0g 年龄(岁)

Hight int %8.0g 身高(cm)

Weight float %9.0g 体重(kg)

Base1 byte %8.0g 糖尿病病程(年)

Base2 byte %8.0g 神经病病变病程(年)

mdns0 byte %8.0g 疗前MDNS

mdns1 byte %8.0g 疗后1月MDNS

mdns2 byte %8.0g 疗后2月MDNS

mdns3 byte %8.0g

mdns4 byte %8.0g

mdns5 byte %8.0g

mdns6 byte %8.0g

fell0 byte %8.0g

fell1 byte %8.0g

fell2 byte %8.0g

fell3 byte %8.0g

fell4 byte %8.0g

fell5 byte %8.0g

fell6 byte %8.0g

gca0 float %9.0g

gca1 float %9.0g

gcb0 float %9.0g

gcb1 float %9.0g

gcc0 float %9.0g

gcc1 float %9.0g

yca0 float %9.0g

yca1 float %9.0g

疗后3月MDNS

疗后4月MDNS

疗后5月MDNS

疗后6月MDNS

疗前感觉障碍得分

疗后1月感觉障碍得分

疗后2月感觉障碍得分

疗后3月感觉障碍得分

疗后4月感觉障碍得分

疗后5月感觉障碍得分

疗后6月感觉障碍得分

疗前正中感觉N传导速度

疗后正中感觉N传导速度

疗前尺感觉N传导速

疗后尺感觉N传导速

疗前腓感觉N传导速度

疗后腓感觉N传导速度

疗前正中运动N传导速度

疗后正中运动N传导速度

ycb0 float %9.0g 疗前腓总运动N传导速度

ycb1 float %9.0g 疗后腓总运动N传导速度

HbA1c0 float %9.0g 疗前糖化血红蛋白

HbA1c1 float %9.0g 疗后糖化血红蛋白

RBC0 float %9.0g 疗前红细胞数

RBC1 float %9.0g

WBC0 float %9.0g

WBC1 float %9.0g

CRE0 float %9.0g

CRE1 float %9.0g

ALT0 float %9.0g

ALT1 float %9.0g

TG0 float %9.0g

TG1 float %9.0g

CHO0 float %9.0g

CHO1 float %9.0g

ECG0 byte %8.0g

ECG1 byte %8.0g

side1 byte %8.0g

degree1 byte %8.0g

relate1 byte %8.0g

side2 byte %8.0g

degree2 byte %8.0g

relate2 byte %8.0g

疗后红细胞数

疗前白细胞数

疗后白细胞数

疗前肌酐

疗后肌酐

疗前谷丙转氨酶

疗后谷丙转氨酶

疗前甘油三脂

疗后甘油三脂

疗前总胆固醇

疗后总胆固醇

ECG 疗前心电图

ECG 疗后心电图

不良反应之一

degree 程度

relate 与药物的关系

不良反应之二

degree 程度

relate 与药物的关系

side3 byte %8.0g 不良反应之三

degree3 byte %8.0g degree 程度

relate3 byte %8.0g relate 与药物的关系

Group float %9.0g group 分组变量

d float %9.0g

-------------------------------------------------------------------------------

Sorted by:

.

. /* 各中心各组病例分配 */

. tab Center Group

| 分组变量

中心编号 | Treatment Placebo | Total

-----------+----------------------+----------

1 | 24 24 | 48

2 | 24 24 | 48

3 | 24 24 | 48

4 | 24 24 | 48

5 | 24 24 | 48

-----------+----------------------+----------

Total | 120 120 | 240

结果显示，各中心均按方案完成了规定的入组病例数。

. /* 产生一变量表示脱落 */

. gen loss=.

(240 missing values generated)

.

. /* 计算脱落时间 */

. for num 1/ 6 : replace loss=X if mdnsX==. & loss==.

-> replace loss=1 if mdns1==. & loss==.

(0 real changes made)

-> replace loss=2 if mdns2==. & loss==.

(0 real changes made)

-> replace loss=3 if mdns3==. & loss==.

(0 real changes made)

-> replace loss=4 if mdns4==. & loss==.

(3 real changes made)

-> replace loss=5 if mdns5==. & loss==.

(0 real changes made)

-> replace loss=6 if mdns6==. & loss==.

(0 real changes made)

.

. /* 按中心排序 */

. sort Center

.

. /* 各中心各组脱落情况 */

. by Center : table Group loss

_______________________________________________________________________________

-> Center = 1

no observations

_______________________________________________________________________________

-> Center = 2

no observations

_______________________________________________________________________________

-> Center = 3

----------------

| loss

分组变量 | 4

----------+-----

Treatment | 1

----------------

_______________________________________________________________________________

-> Center = 4

----------------

| loss

分组变量 | 4

----------+-----

Placebo | 2

----------------

____________________________________________________________________________

-> Center = 5

no observations

.

. /* 脱落情况列表 */

. list No Center Name mdns0 Group loss if loss~=.

No Center Name mdns0 Group loss

139. 139 3 LFZ 17 Treatment 4

179. 179 4 TCY 12 Placebo 4

192. 192 4 XXX 18 Placebo 4

结果显示，共有3例脱落，其中，第3中心在第4次随访时NGF组有1例脱落；第4中心在第4次随访时安慰剂组有2例脱落。

.

. /* 两组性别比较 */

. tab Group Gender, chi

| 性别

分组变量 | female male | Total

-----------+----------------------+----------

Treatment | 63 57 | 120

Placebo | 64 56 | 120

-----------+----------------------+----------

Total | 127 113 | 240

Pearson chi2(1) = 0.0167 Pr = 0.897

.

. /* 两组年龄比较 */

. ttest Age, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 60.06667 .7584014 8.307871 58.56496 61.56838

Placebo | 120 58.03333 .8088216 8.860196 56.43179 59.63488

---------+--------------------------------------------------------------------

combined | 240 59.05 .5571176 8.630829 57.95251 60.14749

---------+--------------------------------------------------------------------

diff | 2.033333 1.108767 -.1509178 4.217585

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 1.8339 t = 1.8339 t = 1.8339

P < t = 0.9660 P > |t| = 0.0679 P > t = 0.0340

.

. /* 扣除性别影响后两组的身高比较 */

. anova High Group Gender

Number of obs = 240 R-squared = 0.4022

Root MSE = 6.07288 Adj R-squared = 0.3971

Source | Partial SS df MS F Prob > F

-----------+----------------------------------------------------

Model | 5880.46954 2 2940.23477 79.72 0.0000

|

Group | 3.52527399 1 3.52527399 0.10 0.7575

Gender | 5874.13204 1 5874.13204 159.28 0.0000

|

Residual | 8740.52629 237 36.8798578

-----------+----------------------------------------------------

Total | 14620.9958 239 61.1757148

.

. /* 扣除性别影响后两组的体重比较 */

. anova Weight Group Gender

Number of obs = 240 R-squared = 0.1922

Root MSE = 9.11056 Adj R-squared = 0.1854

Source | Partial SS df MS F Prob > F

-----------+----------------------------------------------------

Model | 4681.22619 2 2340.61309 28.20 0.0000

|

Group | 188.575343 1 188.575343 2.27 0.1331

Gender | 4476.98469 1 4476.98469 53.94 0.0000

|

Residual | 19671.5512 237 83.0023256

-----------+----------------------------------------------------

Total | 24352.7774 239 101.894466

.

. /* 各组基线情况比较 */

. for var Base1 Base2 mdns0 fell0 gca0 : ttest X , by(Group)

-> ttest Base1 , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 3.55 .2119834 2.322162 3.130252 3.969748

Placebo | 120 3.258333 .2246989 2.461454 2.813407 3.70326

---------+--------------------------------------------------------------------

combined | 240 3.404167 .1544209 2.392278 3.099967 3.708366

---------+--------------------------------------------------------------------

diff | .2916667 .3089119 -.3168841 .9002175

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 0.9442 t = 0.9442 t = 0.9442

P < t = 0.8270 P > |t| = 0.3460 P > t = 0.1730

-> ttest Base2 , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 1.208333 .1673219 1.832919 .8770193 1.539647

Placebo | 120 1.066667 .1467354 1.607406 .7761159 1.357217

---------+--------------------------------------------------------------------

combined | 240 1.1375 .1111358 1.721708 .9185693 1.356431

---------+--------------------------------------------------------------------

diff | .1416667 .2225486 -.29675 .5800834

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 0.6366 t = 0.6366 t = 0.6366

P < t = 0.7375 P > |t| = 0.5250 P > t = 0.2625

-> ttest mdns0 , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 14.21667 .4322735 4.735319 13.36072 15.07261

Placebo | 120 14.79167 .4541893 4.975395 13.89233 15.69101

---------+--------------------------------------------------------------------

combined | 240 14.50417 .3134034 4.855224 13.88678 15.12155

---------+--------------------------------------------------------------------

diff | -.575 .6270154 -1.810209 .6602088

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = -0.9170 t = -0.9170 t = -0.9170

P < t = 0.1800 P > |t| = 0.3600 P > t = 0.8200

-> ttest fell0 , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 10.375 .5178958 5.673264 9.349515 11.40049

Placebo | 120 9.458333 .4752849 5.206486 8.517222 10.39944

---------+--------------------------------------------------------------------

combined | 240 9.916667 .3519802 5.452854 9.223287 10.61005

---------+--------------------------------------------------------------------

diff | .9166667 .7029309 -.4680941 2.301427

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 1.3041 t = 1.3041 t = 1.3041

P < t = 0.9033 P > |t| = 0.1935 P > t = 0.0967

-> ttest gca0 , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 114 49.62544 1.05538 11.26837 47.53454 51.71634

Placebo | 115 47.36783 1.112621 11.93153 45.16373 49.57192

---------+--------------------------------------------------------------------

combined | 229 48.4917 .7689147 11.63579 46.97662 50.00679

---------+--------------------------------------------------------------------

diff | 2.257613 1.533926 -.7649423 5.280168

------------------------------------------------------------------------------

Degrees of freedom: 227

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 1.4718 t = 1.4718 t = 1.4718

P < t = 0.9288 P > |t| = 0.1425 P > t = 0.0712

从上述分析可知，两组的基本情况和基线情况是一致的。即两组具有可比性。

. /* 计算各时点MDNS的减分率 */

. for var mdns1-mdns6 : gen pX = (mdns0-X)/mdns0

-> gen pmdns1 = (mdns0-mdns1)/mdns0

-> gen pmdns2 = (mdns0-mdns2)/mdns0

-> gen pmdns3 = (mdns0-mdns3)/mdns0

-> gen pmdns4 = (mdns0-mdns4)/mdns0

(3 missing values generated)

-> gen pmdns5 = (mdns0-mdns5)/mdns0

(3 missing values generated)

-> gen pmdns6 = (mdns0-mdns6)/mdns0

(3 missing values generated)

.

. /* 按 Group 排序 */

. sort Group

.

. /* 各组各时间点的减分率描述 */

. by Group : summ pmdns*

_______________________________________________________________________________

-> Group = Treatment

Variable | Obs Mean Std. Dev. Min Max

-------------+-----------------------------------------------------

pmdns1 | 120 .0977496 .1776434 -.375 .6

pmdns2 | 120 .1943255 .2145272 -.375 .8571429

pmdns3 | 120 .2586773 .2252976 -.4285714 .7333333

pmdns4 | 119 .3505427 .2422572 -.3333333 1

pmdns5 | 119 .4200117 .2478205 -.3333333 1

pmdns6 | 119 .4998484 .2489689 -.3333333 1

_______________________________________________________________________________

-> Group = Placebo

Variable | Obs Mean Std. Dev. Min Max

-------------+-----------------------------------------------------

pmdns1 | 120 .036402 .1624928 -.5714286 .4615385

pmdns2 | 120 .0810332 .221016 -.5714286 .7142857

pmdns3 | 120 .1347084 .2422806 -.5714286 1

pmdns4 | 118 .1680638 .2485104 -.5714286 1

pmdns5 | 118 .1877292 .2597217 -.5 1

pmdns6 | 118 .2154119 .2516401 -.5 1

.

. /* 保存现有数据库 */

. save "E:赵耐青clinical trial data(ChenFeng)"

file E:赵耐青clinical trial data(ChenFeng) saved

.

. /* 疗前的减分率为0 */

. gen pmdns0=0

.

. /* 产生各时间点的均数以便画图 */

. collapse pmdns0 pmdns1-pmdns6 , by(Group)

.

. drop Group

.

. /*将横向排列的数据库转为列向排列 */

. xpose, clear

.

. /* 产生时间变量 */

. gen week=2*(_n-1)

.

. /* 绘制MDNS平均增加值随时间变化的曲线 */

. gra v1 v2 week , c(ll[-]) xlab(0,2,4,6,8,10,12) ylab saving("e:赵耐青图1", replace)

. /*调用原数据库 */

. use "E:赵耐青clinical trial data(ChenFeng)", clear

.

. /* 各组各时间点MDNS的减分率比较 */

. for var pmdns1-pmdns6 : ttest X, by(Group)

-> ttest pmdns1, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 .0977496 .0162165 .1776434 .0656392 .12986

Placebo | 120 .036402 .0148335 .1624928 .0070302 .0657738

---------+--------------------------------------------------------------------

combined | 240 .0670758 .0111438 .1726386 .0451232 .0890284

---------+--------------------------------------------------------------------

diff | .0613475 .0219775 .0180523 .1046427

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 2.7914 t = 2.7914 t = 2.7914

P < t = 0.9972 P > |t| = 0.0057 P > t = 0.0028

-> ttest pmdns2, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 .1943255 .0195836 .2145272 .1555481 .233103

Placebo | 120 .0810332 .0201759 .221016 .0410829 .1209835

---------+--------------------------------------------------------------------

combined | 240 .1376794 .0144998 .2246302 .1091156 .1662431

---------+--------------------------------------------------------------------

diff | .1132924 .0281173 .0579018 .168683

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 4.0293 t = 4.0293 t = 4.0293

P < t = 1.0000 P > |t| = 0.0001 P > t = 0.0000

-> ttest pmdns3, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 120 .2586773 .0205668 .2252976 .217953 .2994015

Placebo | 120 .1347084 .0221171 .2422806 .0909144 .1785025

---------+--------------------------------------------------------------------

combined | 240 .1966929 .0155936 .2415752 .1659744 .2274113

---------+--------------------------------------------------------------------

diff | .1239689 .0302019 .0644716 .1834661

------------------------------------------------------------------------------

Degrees of freedom: 238

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 4.1047 t = 4.1047 t = 4.1047

P < t = 1.0000 P > |t| = 0.0001 P > t = 0.0000

-> ttest pmdns4, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 119 .3505427 .0222077 .2422572 .3065655 .39452

Placebo | 118 .1680638 .0228772 .2485104 .1227567 .213371

---------+--------------------------------------------------------------------

combined | 237 .2596883 .0169786 .261383 .2262392 .2931373

---------+--------------------------------------------------------------------

diff | .1824789 .0318799 .1196719 .2452859

------------------------------------------------------------------------------

Degrees of freedom: 235

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 5.7239 t = 5.7239 t = 5.7239

P < t = 1.0000 P > |t| = 0.0000 P > t = 0.0000

-> ttest pmdns5, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 119 .4200117 .0227177 .2478205 .3750246 .4649989

Placebo | 118 .1877292 .0239093 .2597217 .140378 .2350803

---------+--------------------------------------------------------------------

combined | 237 .3043605 .018106 .2787384 .2686905 .3400305

---------+--------------------------------------------------------------------

diff | .2322826 .0329745 .1673193 .2972459

------------------------------------------------------------------------------

Degrees of freedom: 235

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 7.0443 t = 7.0443 t = 7.0443

P < t = 1.0000 P > |t| = 0.0000 P > t = 0.0000

-> ttest pmdns6, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 119 .4998484 .0228229 .2489689 .4546528 .545044

Placebo | 118 .2154119 .0231653 .2516401 .1695342 .2612897

---------+--------------------------------------------------------------------

combined | 237 .3582302 .0186798 .2875711 .3214299 .3950306

---------+--------------------------------------------------------------------

diff | .2844365 .0325181 .2203724 .3485007

------------------------------------------------------------------------------

Degrees of freedom: 235

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 8.7470 t = 8.7470 t = 8.7470

P < t = 1.0000 P > |t| = 0.0000 P > t = 0.0000

. /* 产生疗效等级变量 */

. for num 1/ 6 : gen lxX=recode(pmdnsX,0.1599,0.4999,0.7499,1)

-> gen lx1=recode(pmdns1,0.1599,0.4999,0.7499,1)

-> gen lx2=recode(pmdns2,0.1599,0.4999,0.7499,1)

-> gen lx3=recode(pmdns3,0.1599,0.4999,0.7499,1)

-> gen lx4=recode(pmdns4,0.1599,0.4999,0.7499,1)

(3 missing values generated)

-> gen lx5=recode(pmdns5,0.1599,0.4999,0.7499,1)

(3 missing values generated)

-> gen lx6=recode(pmdns6,0.1599,0.4999,0.7499,1)

(3 missing values generated)

.

. /* 1表示<0.16 2表示 0.16～ 表示 0.50～0 4表示 0.75～1 */

. for num 1/6 : recode lxX 0.1599=1 0.4999=2 0.7499=3 1=4

-> recode lx1 0.1599=1 0.4999=2 0.7499=3 1=4

(240 changes made)

-> recode lx2 0.1599=1 0.4999=2 0.7499=3 1=4

(240 changes made)

-> recode lx3 0.1599=1 0.4999=2 0.7499=3 1=4

(240 changes made)

-> recode lx4 0.1599=1 0.4999=2 0.7499=3 1=4

(237 changes made)

-> recode lx5 0.1599=1 0.4999=2 0.7499=3 1=4

(237 changes made)

-> recode lx6 0.1599=1 0.4999=2 0.7499=3 1=4

(237 changes made)

.

. /* 变量名定义 */

. for num 1/6 : label var lxX "第X次随访疗效等级"

-> label var lx1 `"第1次随访疗效等级"'

-> label var lx2 `"第2次随访疗效等级"'

-> label var lx3 `"第3次随访疗效等级"'

-> label var lx4 `"第4次随访疗效等级"'

-> label var lx5 `"第5次随访疗效等级"'

-> label var lx6 `"第6次随访疗效等级"'

.

. /* 定义疗效等级变量的数值标签 */

. lab define lxlab 1 "无效" 2 "好转" 3 "显效" 4 "近愈"

. for num 1/6 : label value lxX lxlab

-> label value lx1 lxlab

-> label value lx2 lxlab

-> label value lx3 lxlab

-> label value lx4 lxlab

-> label value lx5 lxlab

-> label value lx6 lxlab

.

. /* 6次随访各组的疗效 */

. for num 1/6 : tab Group lxX

-> tab Group lx1

| 第1次随访疗效等级

分组变量 | 无效 好转 显效 | Total

-----------+---------------------------------+----------

Treatment | 78 40 2 | 120

Placebo | 99 21 0 | 120

-----------+---------------------------------+----------

Total | 177 61 2 | 240

-> tab Group lx2

| 第2次随访疗效等级

分组变量 | 无效 好转 显效 近愈 | Total

-----------+--------------------------------------------+----------

Treatment | 57 50 12 1 | 120

Placebo | 82 34 4 0 | 120

-----------+--------------------------------------------+----------

Total | 139 84 16 1 | 240

-> tab Group lx3

| 第3次随访疗效等级

分组变量 | 无效 好转 显效 近愈 | Total

-----------+--------------------------------------------+----------

Treatment | 40 63 17 0 | 120

Placebo | 72 42 3 3 | 120

-----------+--------------------------------------------+----------

Total | 112 105 20 3 | 240

-> tab Group lx4

| 第4次随访疗效等级

分组变量 | 无效 好转 显效 近愈 | Total

-----------+--------------------------------------------+----------

Treatment | 24 59 33 3 | 119

Placebo | 61 50 5 2 | 118

-----------+--------------------------------------------+----------

Total | 85 109 38 5 | 237

-> tab Group lx5

| 第5次随访疗效等级

分组变量 | 无效 好转 显效 近愈 | Total

-----------+--------------------------------------------+----------

Treatment | 19 49 44 7 | 119

Placebo | 60 44 12 2 | 118

-----------+--------------------------------------------+----------

Total | 79 93 56 9 | 237

-> tab Group lx6

| 第6次随访疗效等级

分组变量 | 无效 好转 显效 近愈 | Total

-----------+--------------------------------------------+----------

Treatment | 14 29 63 13 | 119

Placebo | 53 48 14 3 | 118

-----------+--------------------------------------------+----------

Total | 67 77 77 16 | 237

.

. /* 降为4.0版本 */

. version 4.0

.

. /* 对两组疗效等级进行Wilcoxon检验 */

. wilcoxon lx1 , by(Group)

Test: Group==2 has longer survival time

Wilcoxon-Gehan statistic = -2562

z = -3.12

Pr>|z| = 0.0018

. wilcoxon lx2 , by(Group)

Test: Group==2 has longer survival time

Wilcoxon-Gehan statistic = -3246

z = -3.46

Pr>|z| = 0.0005

. wilcoxon lx3 , by(Group)

Test: Group==2 has longer survival time

Wilcoxon-Gehan statistic = -4125

z = -4.25

Pr>|z| = 0.0000

. wilcoxon lx4 , by(Group)

Test: Group==1 has longer survival time

Wilcoxon-Gehan statistic = 5763

z = 5.91

Pr>|z| = 0.0000

. wilcoxon lx5 , by(Group)

Test: Group==2 has longer survival time

Wilcoxon-Gehan statistic = -6452

z = -6.48

Pr>|z| = 0.0000

. wilcoxon lx6 , by(Group)

Test: Group==1 has longer survival time

Wilcoxon-Gehan statistic = 7803

z = 7.76

Pr>|z| = 0.0000

从上面的分析可见，MDNS的疗效等级，从第1次随访开始，两组即具有统计学差异。NGF组优于安慰剂对照组。

.

. /* 回到7.0版本 */

. version 7.0

.

. /* 定义疗效变量 */

.

. for num 1/6 : recode lxX 1 2 =1 3 4 =2

-> recode lx1 1 2 =1 3 4 =2

(63 changes made)

-> recode lx2 1 2 =1 3 4 =2

(101 changes made)

-> recode lx3 1 2 =1 3 4 =2

(128 changes made)

-> recode lx4 1 2 =1 3 4 =2

(152 changes made)

-> recode lx5 1 2 =1 3 4 =2

(158 changes made)

-> recode lx6 1 2 =1 3 4 =2

(170 changes made)

.

. for num 1/6 : label var lxX "第X次随访疗效"

-> label var lx1 `"第1次随访疗效"'

-> label var lx2 `"第2次随访疗效"'

-> label var lx3 `"第3次随访疗效"'

-> label var lx4 `"第4次随访疗效"'

-> label var lx5 `"第5次随访疗效"'

-> label var lx6 `"第6次随访疗效"'

.

. /* 定义疗效变量数据标签 */

. lab define lxlab0 1 "无效" 2 "有效"

. for num 1/6 : label value lxX lxlab0

-> label value lx1 lxlab0

-> label value lx2 lxlab0

-> label value lx3 lxlab0

-> label value lx4 lxlab0

-> label value lx5 lxlab0

-> label value lx6 lxlab0

.

. /* 各组有效率比较 */

. for num 1/6 : tab Group lxX , exact row

-> tab Group lx1 , exact row

| 第1次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 118 2 | 120

| 98.33 1.67 | 100.00

-----------+----------------------+----------

Placebo | 120 0 | 120

| 100.00 0.00 | 100.00

-----------+----------------------+----------

Total | 238 2 | 240

| 99.17 0.83 | 100.00

Fisher's exact = 0.498

1-sided Fisher's exact = 0.249

-> tab Group lx2 , exact row

| 第2次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 107 13 | 120

| 89.17 10.83 | 100.00

-----------+----------------------+----------

Placebo | 116 4 | 120

| 96.67 3.33 | 100.00

-----------+----------------------+----------

Total | 223 17 | 240

| 92.92 7.08 | 100.00

Fisher's exact = 0.041

1-sided Fisher's exact = 0.021

-> tab Group lx3 , exact row

| 第3次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 103 17 | 120

| 85.83 14.17 | 100.00

-----------+----------------------+----------

Placebo | 114 6 | 120

| 95.00 5.00 | 100.00

-----------+----------------------+----------

Total | 217 23 | 240

| 90.42 9.58 | 100.00

Fisher's exact = 0.026

1-sided Fisher's exact = 0.013

-> tab Group lx4 , exact row

| 第4次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 83 36 | 119

| 69.75 30.25 | 100.00

-----------+----------------------+----------

Placebo | 111 7 | 118

| 94.07 5.93 | 100.00

-----------+----------------------+----------

Total | 194 43 | 237

| 81.86 18.14 | 100.00

Fisher's exact = 0.000

1-sided Fisher's exact = 0.000

-> tab Group lx5 , exact row

| 第5次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 68 51 | 119

| 57.14 42.86 | 100.00

-----------+----------------------+----------

Placebo | 104 14 | 118

| 88.14 11.86 | 100.00

-----------+----------------------+----------

Total | 172 65 | 237

| 72.57 27.43 | 100.00

Fisher's exact = 0.000

1-sided Fisher's exact = 0.000

-> tab Group lx6 , exact row

| 第6次随访疗效

分组变量 | 无效 有效 | Total

-----------+----------------------+----------

Treatment | 43 76 | 119

| 36.13 63.87 | 100.00

-----------+----------------------+----------

Placebo | 101 17 | 118

| 85.59 14.41 | 100.00

-----------+----------------------+----------

Total | 144 93 | 237

| 60.76 39.24 | 100.00

Fisher's exact = 0.000

1-sided Fisher's exact = 0.000

.

. /* 计算有效率的95%CI */

. cii 119 76

-- Binomial Exact --

Variable | Obs Mean Std. Err. [95% Conf. Interval]

-------------+-------------------------------------------------------------

| 119 .6386555 .0440373 .5455312 .7246909

从上面的分析可见，从第1 次随访开始，NGF组的有效率就优于安慰剂对照组。

. /* 产生结果变量，y=1 表示有效，y=0 表示无效 */

. gen y=1 if pmdns6>=0.5

(144 missing values generated)

. replace y=0 if pmdns5<0.5

(172 real changes made)

/*********************************

/* CMH 分析

/*********************************

.

. /* 影响疗效的因素分析其中，中心变量Center按哑变量形式进入模型 */

. xi : logistic y Gender Age Base1 Base2 mdns0 fell0 Group

_ICenter_1-5 (naturally coded; _ICenter_1 omitted)

Logit estimates Number of obs = 233

LR chi2(11) = 53.83

Prob > chi2 = 0.0000

Log likelihood = -107.0422 Pseudo R2 = 0.2009

------------------------------------------------------------------------------

y | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]

-------------+----------------------------------------------------------------

Gender | .4888366 .1764132 -1.98 0.047 .2409789 .9916275

Age | .9980924 .0214442 -0.09 0.929 .9569351 1.04102

Base1 | .9870993 .0779154 -0.16 0.869 .8456144 1.152257

Base2 | 1.007274 .1030168 0.07 0.944 .8243145 1.230843

mdns0 | 1.005151 .0425848 0.12 0.903 .9250574 1.092179

fell0 | .9536386 .0352244 -1.29 0.199 .8870398 1.025238

_ICenter_2 | .384867 .2208795 -1.66 0.096 .1249674 1.18529

_ICenter_3 | 1.653209 .9146822 0.91 0.364 .5589559 4.889651

_ICenter_4 | .3796829 .2254166 -1.63 0.103 .1185947 1.215561

_ICenter_5 | 1.188417 .6194094 0.33 0.740 .4278763 3.300804

Group | .1075251 .0431229 -5.56 0.000 .0489932 .2359848

------------------------------------------------------------------------------

. /* 各中心间效应比较 */

. test _b[_ICenter_2]=_b[_ICenter_3]

( 1) _ICenter_2 - _ICenter_3 = 0.0

chi2( 1) = 6.20

Prob > chi2 = 0.0128

. test _b[_ICenter_2]=_b[_ICenter_4]

( 1) _ICenter_2 - _ICenter_4 = 0.0

chi2( 1) = 0.00

Prob > chi2 = 0.9825

. test _b[_ICenter_2]=_b[_ICenter_5]

( 1) _ICenter_2 - _ICenter_5 = 0.0

chi2( 1) = 3.77

Prob > chi2 = 0.0521

. test _b[_ICenter_3]=_b[_ICenter_4]

( 1) _ICenter_3 - _ICenter_4 = 0.0

chi2( 1) = 5.74

Prob > chi2 = 0.0166

. test _b[_ICenter_3]=_b[_ICenter_5]

( 1) _ICenter_3 - _ICenter_5 = 0.0

chi2( 1) = 0.36

Prob > chi2 = 0.5486

. test _b[_ICenter_4]=_b[_ICenter_5]

( 1) _ICenter_4 - _ICenter_5 = 0.0

chi2( 1) = 3.73

Prob > chi2 = 0.0536

可见，疗效与性别无关，由于减分率考虑了疗前的基线水平，故所得疗效与基线(疗前MDNS，感觉得分FELL0)无关，各中心间差异无统计学意义。但疗效与年龄有关，年龄越大，疗效越差。在考虑以上有效因素后，分组变量有统计学意义，即NGF组的疗效优于安慰剂对照组。

.

. /* 计算神经传导速度的差值 */

. gen speed1=gca1-gca0

(15 missing values generated)

. gen speed2=gcb1-gcb0

(15 missing values generated)

. gen speed3=gcc1-gcc0

(16 missing values generated)

.

. /* 神经传导速度差值的组间比较 */

. for var speed1 - speed3 : ttest X, by(Group)

-> ttest speed1, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 113 2.836283 .7502392 7.975152 1.34978 4.322786

Placebo | 112 .1500001 .6843089 7.242044 -1.206004 1.506004

---------+--------------------------------------------------------------------

combined | 225 1.499111 .5146867 7.7203 .4848639 2.513358

---------+--------------------------------------------------------------------

diff | 2.686283 1.015885 .6843196 4.688246

------------------------------------------------------------------------------

Degrees of freedom: 223

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 2.6443 t = 2.6443 t = 2.6443

P < t = 0.9956 P > |t| = 0.0088 P > t = 0.0044

-> ttest speed2, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 116 4.003448 .8380118 9.025663 2.343508 5.663388

Placebo | 109 -.6688075 .7054635 7.365255 -2.067158 .7295435

---------+--------------------------------------------------------------------

combined | 225 1.74 .5713674 8.570511 .6140571 2.865943

---------+--------------------------------------------------------------------

diff | 4.672256 1.102305 2.499988 6.844524

------------------------------------------------------------------------------

Degrees of freedom: 223

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 4.2386 t = 4.2386 t = 4.2386

P < t = 1.0000 P > |t| = 0.0000 P > t = 0.0000

-> ttest speed3, by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 114 2.173684 .6231747 6.653685 .9390629 3.408306

Placebo | 110 -.2227273 .7345729 7.704266 -1.678627 1.233172

---------+--------------------------------------------------------------------

combined | 224 .996875 .485907 7.27239 .039318 1.954432

---------+--------------------------------------------------------------------

diff | 2.396412 .9607866 .5029824 4.289841

------------------------------------------------------------------------------

Degrees of freedom: 222

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = 2.4942 t = 2.4942 t = 2.4942

P < t = 0.9933 P > |t| = 0.0134 P > t = 0.0067

. 由上可信区间可见，NGF组三根神经的传导速度均有提高，而对照组无改善；假设检验结果表明，NGF组的神经传导速度与安慰剂组的差异有统计学意义。

. /* 治疗前后实验室指标WBC、RBC的差值 */

. gen dWBC=WBC1-WBC0

(3 missing values generated)

. gen dRBC=RBC1-RBC0

(3 missing values generated)

.

. /* WBC差值的两组比较 */

. ttest dWBC , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 119 .0421849 .1433136 1.563367 -.241615 .3259848

Placebo | 118 .1077966 .1491779 1.620487 -.1876424 .4032356

---------+--------------------------------------------------------------------

combined | 237 .0748523 .1032183 1.589025 -.1284946 .2781992

---------+--------------------------------------------------------------------

diff | -.0656117 .2068329 -.4730952 .3418718

------------------------------------------------------------------------------

Degrees of freedom: 235

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = -0.3172 t = -0.3172 t = -0.3172

P < t = 0.3757 P > |t| = 0.7514 P > t = 0.6243

.

. /* RBC差值的两组比较 */

. ttest dRBC , by(Group)

Two-sample t test with equal variances

------------------------------------------------------------------------------

Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]

---------+--------------------------------------------------------------------

Treatmen | 119 -.1057983 .047912 .5226586 -.2006772 -.0109195

Placebo | 118 .0416949 .0631882 .6863994 -.083446 .1668358

---------+--------------------------------------------------------------------

combined | 237 -.0323629 .0398109 .6128811 -.110793 .0460672

---------+--------------------------------------------------------------------

diff | -.1474932 .0792095 -.3035447 .0085583

------------------------------------------------------------------------------

Degrees of freedom: 235

Ho: mean(Treatmen) - mean(Placebo) = diff = 0

Ha: diff < 0 Ha: diff ~= 0 Ha: diff > 0

t = -1.8621 t = -1.8621 t = -1.8621

P < t = 0.0319 P > |t| = 0.0638 P > t = 0.9681

从RBC差值的分析可见，处理组治疗后RBC有所降低，治疗后比治疗前降低0.1058,差值的95%CI为：-0.2007～-0.0109(这种差别在临床上是否有意义？)；对照组治疗前后差值无统计学意义。

.

. /* 两组心电图指标疗前正常、疗后异常率比较 */

. tab Group ECG1 if ECG0==1 , exact row

| 疗后心电图

分组变量 | normal unnormal | Total

-----------+----------------------+----------

Treatment | 54 18 | 72

| 75.00 25.00 | 100.00

-----------+----------------------+----------

Placebo | 67 20 | 87

| 77.01 22.99 | 100.00

-----------+----------------------+----------

Total | 121 38 | 159

| 76.10 23.90 | 100.00

Fisher's exact = 0.852

1-sided Fisher's exact = 0.455

可见，试验组对受试者的心电图的影响于对照组相比并无特别之处。

. /* 将三个不良事件变量接到一起，便于分析 */

. stack No Center Group Name Gender Age mdns0 side1 degree1 relate1 No Center Group Name Gender Age

mdns0 side2

> degree2 relate2 No Center Group Name Gender Age mdns0 side3 degree3 relate3 , into(No Center Group

Name Gende

> r Age mdns0 side degree relate)

Warning: data in memory will be lost.

Press any key to continue, Ctrl-Break to abort.

.

. /* 删除无不良事件者 */

. drop if side==.

(557 observations deleted)

.

. /* 定义不良事件名称 */

. label define sidelab 1 "搔痒" 2 "浮肿" 3 "皮疹" 4 "针孔周围感觉过敏" 5 "针孔周围痛" 6 "注射侧肢痛" 7 "胸背腰*

> *" 8 "其他部位痛" 9 "头晕" 10 "头痛" 11 "失眠" 12 "倦睡" 13 "发热" 14 "胃寒" 15 "恶性" 16 "呕吐"

17 "食欲下*

> *" 18 "其他"

. label value side sidelab

.

. /* 定义组别 */

. label define group 1 "Treatment" 2 "Placebo"

. label value Group group

.

. /* 两组不良事件 */

. tab side Group

| Group

side | Treatment Placebo | Total

-----------------+----------------------+----------

搔痒 | 6 5 | 11

浮肿 | 3 6 | 9

皮疹 | 2 3 | 5

针孔周围感觉过敏 | 1 0 | 1

针孔周围痛 | 33 17 | 50

注射侧肢痛 | 6 5 | 11

胸背腰痛 | 5 6 | 11

其他部位痛 | 1 0 | 1

头晕 | 10 7 | 17

头痛 | 3 3 | 6

失眠 | 6 7 | 13

倦睡 | 7 4 | 11

发热 | 0 2 | 2

食欲下降 | 1 4 | 5

其他 | 6 4 | 10

-----------------+----------------------+----------

Total | 90 73 | 163

.

. /* 定义不良事件与试验药物的关系变量 */

. label define relatelab 0 "无关" 1 "可能无关" 2 "可能有关" 3 "有关"

. label value relate relatelab

.

. /* 定义不良事件的程度 */

. label define deg 1 "轻" 2 "中" 3 "重"

. label value degree deg

.

. /* 两组不良反应结果 */

. tab side Group if relate==2 | relate==3

| Group

side | Treatment Placebo | Total

-----------------+----------------------+----------

搔痒 | 2 1 | 3

浮肿 | 0 1 | 1

针孔周围感觉过敏 | 1

针孔周围痛 | 14 6 | 20

注射侧肢痛 | 2 0 | 2

倦睡 | 2 2 | 4

其他 | 0 1 | 1

-----------------+----------------------+----------

Total | 21 11 | 32

.

. /* 不良事件列表 */

. list

Observation 1

_stack 1 No 15 Center 1

Group Treatment Name HFY Gender

Age 60 mdns0 13 side 搔痒 degree 轻 relate 无关

0 |0

1

Observation 2

_stack 1 No 74 Center 2

Group Treatment Name JBL Gender 0

Age 67 mdns0 9 side 注射侧肢痛

degree 中 relate 可能有关

Observation 3

_stack 1 No 137 Center 3

Group Treatment Name JJH Gender 0

Age 55 mdns0 17 side 针孔周围痛

degree 轻 relate 可能无关

Observation 4

_stack 1 No 111 Center 3

Group Treatment Name ZFK Gender 1

Age 65 mdns0 17 side 针孔周围痛

degree 轻 relate 可能无关

Observation 5

_stack 1 No 22 Center 1

Group Treatment Name YJH Gender 1

Age 69 mdns0 13 side 针孔周围痛

degree 轻 relate 可能无关

Observation 6

_stack 1 No 129 Center 3

Group Treatment Name ZYL Gender 1

Age 62 mdns0 21 side 胸背腰痛

degree 轻 relate 无关

Observation 7

_stack 1 No 142 Center 3

Group Treatment Name FES Gender

Age 55 mdns0 16 side 头晕 degree 轻 relate 无关

Observation 8

_stack 1 No 71 Center 2

Group Treatment Name HJY Gender

Age 58 mdns0 7 side 其他

degree 轻 relate 可能无关

Observation 9

_stack 1 No 127 Center 3

Group Treatment Name PSM Gender

Age 59 mdns0 13 side 头晕 degree 轻 relate 无关

1

0

1

Observation 10

_stack 1 No 138 Center 3

Group Treatment Name LSN Gender 1

Age 58 mdns0 17 side 针孔周围痛

degree 轻 relate 可能无关

Observation 11

_stack 1 No 6 Center 1

Group Treatment Name SWP Gender 1

Age 62 mdns0 18 side 浮肿

degree 重 relate 无关

Observation 12

_stack 1 No 18 Center 1

Group Treatment Name XLM Gender 0

Age 51 mdns0 18 side 针孔周围痛

degree 轻 relate 有关

Observation 13

_stack 1 No 25 Center 1

Group Treatment Name LDZ Gender 0

Age 58 mdns0 12 side 头晕

degree 轻 relate .