Professional Documents
Culture Documents
171 242 PDF
171 242 PDF
171
172
1:
mortality
/
5 (quality of recurrence
life scale)
PANSS BPRS
/ ICH
guideline (E9)[1]
173
Do Tirofiban
and ReoPro Give Similar Efficacy Trial.
Comparison of two platelet glycoprotein
IIb/IIIa inhibitors, tirofiban and abciximab,
for the prevention of ischemic events
with percutaneous coronary revas-
cularization New England
(com- Journal of Medicine. Vol. 344 (2001)
posite variables) 2
(noninferiority trial)
3 events (death +
myocardial infarction + urgent target
vessel revascularization)
( + + ) 30 ( 1)[2]
2.
( + (Efficacy & safety variables)
+ )
(validity) (
(reliability) )
2 (effectiveness)
(safety)
174
1
( New England Journal of Medicine Massachusetts Medical
Society)
10%
8% 7.6%
6.0% Abciximab
6% Better
4%
2% Tiroban
Better
0%
Tiroban Abciximab
Tiroban Abciximab
8% 7.6% 7.2%
6.9%
6.0% 5.7%
6% 5.4%
Event Rate
4%
p = 0.038 p = 0 .6 6 p = 0.04 p = 0.04 p = 0 .4 9
2%
0.5% 0.4% 0.8% 0.7%
0%
Composite Death MI Death/MI Urgent
TVR
175
3 (phase III)
3.
(independent
variable) /
3 (dependent variable)
(external/control variables)
[4,5]
(US-FDA)[1,3]
3.1 (Binary)
12
3
(propor-
176
( ) ous)
3.2 (Count)
CD4
30
(mean)
2
3.6 (Time to event)
(count per unit
time)
(incidence density ratios)
3.3 (Categorical)
2
3
2
3.4 (Ordinal)
( )
4 (censored)
177
(averaged relative risk)
(hazard risk ratios)
3.7
3
57 3
(likert scale)
(pain score)
10
3
4.
Generalized
Estimating Equation GEE
/
(randomeffects model) 2
(multiple
measurements)
[6,7,8,9]
178
17 9
VKC
1 A Randomized
(not eye specific)
toconjunctivitis Ophthalmo-
VKC
VKC
(eye) ( 2)
26
(placebo) 2:1
2
( Ophthalmology American Academy of
Opthelmology)
179
ICSI Laser-
assisted ICSI
59
p < 0.01 (mature oocytes)
( p-value 6
775
)
2 No advan- ICSI (
tage of laser-assisted over conventional ) ICSI
intracytoplasmic sperm injection: a ran- Laser-assisted ( )
domized controlled trial Journal
of Experimental & Clinical Assisted
Reproduction (2006)[11] (oocytes)
2
Intracytoplas-
mic Sperm Injection ICSI ( 3)
3
Table I: Mean clinical outcomes per cycle compared betwee laser-assisted and conventional ICSI
180
3 A rand- intervention)
3 (practices) cluster
(sample dressing),
(phenytonin
cream)
(hydrocolloid dressing)
83
91
(baseline)
()
( 4) ()
4 Rational GEE
Prescribing in Primary Care (RaPP): A cluster effect
(passive guideline dissemination)
(tailored
181
4
( Open Access BMC
Dermatology)
182
5
( Open Access PLoS Medicine)
a
b
Adjusted for baseline differences clustering effects (GEE).
Fewer data on exclusion criteria were available for the baseline period, explaining the higher number of prescription than during follow-
up.
c
Recommended treatment goal for blood pressure is less than 140/90 mm Hg; for lipids is total cholesterol < 5 mmol/l (190 mg/dl) and/or
LDL cholesterol < 3 mmol/l (115 mg/dl).
DOI: 10.1371/journal.pmed.0030134.t002
183
H0
2
(standard treatment)
( 5 100 )
(research hypothesis)
(statistical H0
hypothesis) (not reject) H0
1.
(significance testing)
(onetailed
onesided)
Null (twotailed twosided)
Hypothesis H0
2 (hypothesis formulation)
()
(probability)
(acceptance) H 0
95 (pro-
(rejection) H0 5 bability estimate) H0
184
H0
H 0
( p-value = 0.05) [4,14]
2
H0 (null hypothesis) H0 : 1 2
Ha (alternative hypothesis) H0 : 1 2
H0 (null hypothesis) H0 : 1 2
Ha (alternative hypothesis) H0 : 1 2
H0 (null hypothesis) H0 : 1 2
Ha (alternative hypothesis) H0 : 1 2
(predetermined direc- ()
tion)
185
[15,16,17]
p-value=0.05
(post-marketing surveillance)
X X
(relative risk) ( Z statistics) p=0.05
Y +1.64 -1.64
X
+1.96
-1.96
1.85
H0
X
Y H 0 ( 1.85>1.64)
H0 (
1.85<1.96)( 6)
H0
(a priori)
186
2
H 0
(Significance level)
significant
(conservative) [16,17]
impor-
tant ()
significant
(probably
true, not due to chance)
(equality) (superiority)
(non-inferiority) highly significant
(equivalence) 4
( 4)
(statistical significance)
(clinical public
[18,19]
health importance)
2.
Intelligence Quotient IQ
1,000
187
IQ
(= 100) IQ
(= 98)
0.001
(so what?) p-value (probability
2 value) p-value
?
25
p-value
[20,21]
(inferential
statistics)
( ) 5
2 5
(population) 95
(sample) p
p=0.027
(p<0.05) (p>0.05)
p-value 3
188
p=0.003
1 2 p=0.1 p=0.39
p
p<0.001 p<0.0001
[22]
(Confidence
level)
(Confidence Interval CI)
2
2
(plausibility) (estimate)
2
(inference)
p value
point estimate
95
2 ( ( 95CI)
)
2
2 interval estimate
95CI
95CI
95
2 2
Roald Koffman
Nobel Prize .. 1981
189
2
95CI
2 X (t 0.05,9 x SD
1 n) 6.3 1.94
2 0
95CI /
( 10
20 = 20 x 10 ) 6.0
95CI 95CI 48
95 6 48
20 95CI
100 95
95CI (6 ) 19 ( 95
[4,23] 20 )
1 ( 5 20 )
20 95CI
( 3 7)
95CI
20 10
1
p-value
10
95CI
6.3 6 /
2.71
95
190
7 95CI
1
95CI
[16,18,24] / (compliance/adherence)
/
(Analysis data sets)
(violation/deviation)
(longitudinal study)
1, 3, 6
4
3
191
(IntentionToTreat
ITT)
ITT
(
ICH Guideline E3)[25]
ITT
(1)
,
(2)
1 (Type I error)
( Type I error
)
1.
(inclusion criteria eligibility
violation)
(
ITT
192
- -
( GEE)
( / /)
)
- - 2.
(valid cases)
ITT (effectiveness
sample)
(evaluable samples)
-
(imputation technique) (
193
3. Intention-
)
To-Treat Pre-Protocol
-
- / (sensitive)
(confirmatory trial)
Per-Protocol PP
( ITT PP)
PP
2
ITT PP
2
(
)
ITT
PP
PP
PP
(validity)
PP
ITT PP
PP
(
PP
194
ITT PP
ITT (
as-treated
[ 15,
17,18]
( 8)
8
( British Medical Journal British Medical
Journal Group, Ltd)
195
2 A phase
PP According
9)
2
FFM + ME TRAP
Rabies Vaccine
?
9
2 Intention-To-
9
( Open Access PLoS Clinical Trials)
196
(CD4)
A B
2 - (t-test)
3 ANOVA
(assumption) 2
2 (independent variable)
A B
(dependent variable)
90
-
( , , A B,
)
( ,
2 , )
2
1
(nominal)
/
197
(ordinal)
(continuous) (gaussian
normal distribution)
(time to event)
(skew) (
)
/
(
10)
10
198
2
(apgar) (titre)
http://www.graphpad.com
( 2)
(
30 100
)
( Kolmogorov Smirnov test)
(parametric)
1. Parametric
Non-parametric
(
- (logarithm) )
(non-parametric statistics)
) [5,28,29]
199
2
,
2 ,
,
(survival time)
Mean, SD
Median,
Proportion
Kaplan-Meier
Inter-quartile range
survival curve
One-sample t test
Wilcoxons test
Chi-square Binomial
test
Unpaired t test
Mann-Whitney test
Fishers test (chi- Log-rank test or
2
square for large Mantel-Haenszel
samples)
Paired t test
Wilcoxons test
McNemars stest
Conditional
2 proportional hazards
regression
2
One-way ANOVA
Kruskal-Wallis test
Chi-square test
Coxs proportional
3
hazard regression
Repeated Friedmans test
Cochrans Q
Conditional
3
measures ANOVA
proportional hazards
regression
Pearson correlation
Spearman Spearman correlation.
2
correlation
Odds Ratio (OR)
Relative Risk (RR)
Simple linear / Simple non- Simple logistic Coxs proportional
non-linear parametric regression
hazard regression
regression
regression
Multiple linear/ Multiple non-linear Multiple logistic Coxs proportional
non-linear regression
regression
hazard regression
regression
Multivariate ANO- General Linear Mixed Multivariate Coxs
VA, General Linear Model (GLMM) proportional hazard
Mixed Model regression
(GLMM)
/ Generalized Estima- Generalized Estimation Multivariate Coxs
tion Equation (GEE) Equation (GEE) proportional hazard
regression
200
11
( Open Access PLoS Clinical Trials)
6 2
/
12
(paired ttest)
- (Wilcoxontest) (
Chi-square test, ttest, ANOVA
12)
2.
/
2
202
12
( Open Access PLoS Pathogens)
RR ( VE = 1 RR)
2 x 2
2
()
2
A B
/
2
(Relative Risk RR) chi-square
/ (Odds test 2
Ratio OR)
203
1
2 (strength of 2 OR RR 1
association) OR 95%CI OR RR
RR
1
2
OR RR
1 2 /
OR RR
(study
OR RR 1 design)
OR RR 16
16 2
OR RR
( Randomized Controlled Trial
RCT) cohort study
OR RR
RR
95%CI
OR RR 95%CI
OR RR
p-value case-control study
H2test 95%CI OR
RR 1 cross-sectional
study OR
p-value H2
test RR
(p-value > 0.05)
OR RR 1 RR / OR
2 OR RR
204
RCT
OR
RR ( (before-and-after trial)
logistic regression OR
RR) OR (ICU)
RCT 48
4
(control)
(Hazard 4
Ratio HR) RR HR
(Medical Emergency Team MET)
Cox regression
( survival
2
analysis )
[14,32,33,34] ( chi-square
1 Effect of the test) MET
medical emergency team on long-term
mortality following major surgery
Critical Care (2007) [35]
odds ratio ( 13)
13 OR 2
( Open Access Critical Care)
205
2 Cholesterol
lowering with simvastatin improves
prognosis of diabetic patients with (RCT)
coronary heart disease. A subgroup
analysis of the Scandanavian Simvasta- (Risk
tion Study Diabetes care Ratio RR)
(1997) [36] RR
-
(simvastatin)
( RR < 1)
( 95CI RR 1)
p-value <0.05 RR
1 (
(Pooled RR = 0.94, 95CI 0.89
) ( 14) 0.98) ( 15)
3 Fracture 3.
prevention with vitamin D supplementation:
considering the inconsistent results
BMC Musculoskeletal
Disorders (2007) [37]
meta survival analysis
analysis
(/)
(/ )
()
206
14 RR 2
(
Diabetes Care The American
Diabetes Association)
15 RR
( Open Access BMC Musculoskeletal
Disorders)
207
/
5 (five-year survival probability)
2
(remission/re-infection) ()
survival time
regression
2 [7,33,34,38]
(outcome)
/ /
(step curve)
Kaplan-Meier ( 16)
16
208
hazard probability
2
A
B
/ 1 Mortality
and morbidity in patients receiving
1 encainide, flecanide, or placebo
(one year survival probability) The New England Journal of
A 19 Medicine. Vol.324 (1991) [39]
B 69
50 Cardiac Arrhythmia Suppression Tria
(median survival CAST
time) A 0.7 (
2 ( 17)
2 Safety and
209
17
2
( New England Journal of Medicine
HRV 3 1.
severe gastroenteritis
hazard
(dose) (subgroup)
2
( 18)
/
data
dredging data trawling
210
18
2
( New England Journal of Medicine
- / (biologi-
cal/clinical rationale)
- (prior
5 100 evidence/belief)
(pilot
study)
[14,18,
41,42]
211
- (statistical
evidence)
-
( power
)
/
-
(independent con-
firmation) -
/
212
HIV 3
-
( 19)
2 Randomised
trial of home-based psychological
- nursing intervention for patients
recovering from myocardial infarction
Lancet, Vol. 350 (2002)
Subgroup
Analysis, Covariate adjustment and
/
baseline comparisons in clinical trial
reporting: Current practice and
problems. [Statistics in Medicine,
Vol.21 (2002)] [44]
1 Under-
213
19
( VaxGen)
()
(
20)
p-value 3 Prevention
=0.064
of heart failure by antihypertensive
drug treatment in older persons with
(interaction)
Journal of the American
Medical Association, Vol.278 (1997)
22 12 (OR=
2.0) Subgroup
2 Analysis, Covariate adjustment and
11 11 (OR=1.0) baseline comparisons in clinical trial
interaction OR reporting: Current practice and
2 (-) problems. [Statistics in Medicine,
p=0.21 Vol.21 (2002)] [44]
214
20
( Statistics in Medicine John Wiley &
Sons, Ltd)
RR
RR ( 95CI)
2 MI ( yes)
Myocardial 1892
Infarction MI ( yes vs. no) ()
5 17
(active vs. placebo) interaction
MI ( yes) 2 p=0.24
RR (/ MI)
(active
yes) (placebo yes) /
76 33 ( 21)
215
21
( Statistics in Medicine John Wiley &
Sons, Ltd)
2.
/
/ (baseline)
(change from
baseline)
216
( baseline )
(demographic
)
(precise) [14,15,32,33]
(personal characteristics
, /
) (study
- p-value site)
(adjusted/controlled for)
(covariables)
(unbiased
estimate)
Toremifene
and tamoxifen are equally effective for
early-stage breast cancer: first results
217
Press)
218
3.
H 0
(p-value) p=0.05
5
[14]
3
(multiple
null hypotheses)
H0
p-value
1
p-value < 0.05
Type I error 0.05
1
5%
0.0500
2
10%
0.0253
3
14%
0.0170
4
19%
0.0127
5
23%
0.0102
10
40%
0.0056
20
64%
0.0026
50
92%
0.0010
100
99%
0.0005
N
100 (1.00 0.95N)
1.00-0.95(1/N)
219
1 Medroxy-
progesterone improves nocturnal
breathing in postmenopausal women
H0 0.05 with chronic obstructive pulmonary
/ H0 disease Respiratory
0.05 Research (2005) [46]
3 H0
/ H0 (singleblinded)
0.05
14 15
1
2
3 0.05 MPA 2
3
H0 p-value H0 0.017
4
S1, S2, SWS (Slow Wave Sleep)
Rapid Eye Movement REM
(multiple comparisons)
(SaO2 )
p-value
p-value
Bonferroni
adjustment ( 23)
( ) 2 Effects of
(interim analysis) glaucoma drugs on ocular hemody-
namics in normal tension glaucoma: a
(final analysis)
randomized trial comparing bimatoprost
[38,42]
and latanoprost with dorzolamide
220
23 Bon-
ferroni
( Open Access Respiratory Research)
BMC Ophthalmology
(2005) [47] (hemodynamics) Peak Systolic
Velocity PSV End Diastolic
Velocity EDV
1 3 2
(bimatoprost), 4
(latanoprost), 8
(dorzalamide) pvalue
(negative 0.05
control) 1 pvalue Bonferroni
221
p <0.006 ( 2 )
(SAP)
PSV
( ) /
pvalue <0.001 EDV
p=0.04 p=0.02
( 24)
4.
ICH guideline [1,25] [4,14,38,41]
24
Bonferroni
( Open Access BMC Ophthalmology)
222
Last Observation Carried Forward Missing
LOCF Completely At Random MCAR,
Missing At Random MAR
non-random/non-ignorable missing-
(placebo-controlled su- ners at random
periority trial)
non-
LOCF ignorable
(H 0) (multiple imputations)
LOCF ITT PP
1 A clinical
trial of the use of Sibutramine for
223
(2004) [49]
day-15, day 0, day 30,
day 60, day 90, day 120, day 10,
day 180 (gabapentin)
2 ITT ( 3
6 ) 3
(wash out)
LOCF. 2
(VAS)
( 12
25)
2 Randomised
controlled trial of gabapentin in
25
LOCF
( International Journal of Obesity Nature
Publishing Group)
224
4 (open-label)
(proof-of-concept)
(17 )
Hydroxyurea
( (HU) 3
26)
3 Abacavir, ( HIVRNA
efavirenz, didanosine, with or without 400 copies/ml
hydroxyurea, in HIV-infected adults 16
failing initial nucleoside/protease ) (primary efficacy
inhibitor-containing regimens endpoints)
BMC Infectious Diseases HIVRNA <400 copies/ml
(2005) [50]
24 48
26
( Open Access BMC Neurology)
225
ml
Intention-To-Treat
24 48
ITT ITT 2
(ITT:Missing = 27)
treatment failure analysis)
(subgroup analysis)
HU HU
24
8
48
CD4 ( 28)
27
2 (:
: )
( Open Access BMC Infectious
Diseases)
226
28
( Open Access BMC Infectious
Diseases)
-
(outlier) (lot)
- -
227
(guassian/
normal distribution)
2
[14,15,32,38]
- /
(-)
4 Supervised
( versus unsupervised antimalarial
treatment with six-dose artemether-
)
lumefantrine: pharmacokinetic and
dosage-related findings from a clinical
trial in Uganda Malaria
Journal (2006) [51]
228
(Plas-
modium falciparum)
- (Artemeter (missing)
Lumefantrine A/L)
3 7
(lumefan- ( 29)
trine concentration) 2
5.
(supervised)
29
( Open Access Malaria Journal)
229
(scale)
101001000
log(10)log(100)...log(1000)
123
10 1, 2, 3
10
(skew
curve)
(
) ( 4)
30
( titre)
-
(log)
A B
(non-normal non-guassian
abnormal)
30 ()
[5,23,28]
(count) (value)
230
30 () (synergistic interaction)
(ketamine)
(opioids)
(count) (value)
10
10
10
1 Investiga-
tion of the potentiation of the
analgesic effects of fentanyl by 5 5
ketamine in humans: a double-blinded,
randomised, placebo controlled,
crossover study of experimental pain
BMC Anesthesiology (Subjective Sedation
(2005) [52]
VAS)
(crossover)
231
30
()
()
232
31
( Open Access BMC
Anesthesiology)
233
32 log 10
( Pediatrics American Academy of
Pediatrics)
4
24
pearson correlation
R2
(linear regression)
( 33)
(data transformation)
4 Managing
an effective treatment for neuroleptic
malignant Syndrome -
Critical Care (2007) [55]
Neuro-
leptic Malignant Syndrome NMS
( 34)
234
33 log 10
( Open Access PLoS Pathogens)
34
( Open Access Critical Care)
235
1. US FDA. Guidance for Industry:
(E9) Statistical Principles for Clini-
cal Trails; September,1998.
2. Topol EJ, Moliterno DJ, Herrmann
HC, et al. TARGET I: Do Tirofiban
and ReoPro Give Similar Efficacy
Trial. Comparison of two platelet
(peer-
glycoprotein IIb/IIIa inhibitors,
review journal)
tirofiban and abciximab, for the
prevention of ischemic events
with percutaneous coronary revas-
cularization. N Engl J Med. 2001;
:
344:1888-1894.
[56,57] 3. US FDA Guidance for Industry: (E8)
// General Considerations for Clinical
Open Access Trials;1999.
236
237
238
239
240
241