close

Вход

Забыли?

вход по аккаунту

код для вставкиСкачать
Asthma Trial
– a double blinded, randomized,
placebo-controlled study
Team Moser:
Jing Dong
Yan Yan Wu
Haipeng Yao
• Purpose
The new puffer is effective or not
• Target population:
Physician-diagnosed severe asthma patients
Objectives and Endpoints
• Primary objective: the new puffer is effective or not
Primary endpoint: the measurement of FEV1
• Secondary objective :
- if the puffer is effective for patients of different age group
- if the puffer will reduce symptoms of different severity levels
Secondary endpoint:
- record the numbers of Hospitalization, emergency visits, the use of
rescue medication etc
* Rescue medication is allowed in this trial. And the # of the use of rescue
medication will be one of the measures of the secondary endpoint
Evaluation design:
A double blinded, randomized, placebo-controlled study
1. Control
- Intervention Group: patients take the new puffer twice a day for 12
weeks and record FEV1 weekly.
- Control Group
: patients take placebo puffer twice a day for 12
weeks and record FEV1 weekly
* 12 weeks later, all patients visit physician and physician diagnose if the
puffer is effective based on the FEV1 value
The type of design is superiority because our goal is that the treatment if
more effective than placebo
2. Randomization
•
Sample size
α=0.05 β=1- power(0.8) =0.2
(Zα=1.645, Zβ=1.28)
Assume : effective rate of the intervention group (Pi ) = 0.4
effective rate of the placebo group
(Pc) = 0.2
2N = 2
•
Z  *
2 Pc (1  Pc )  Z  *
 Pc  Pi ^ 2

Pi (1  Pi ) ^ 2
= 126
Randomize
Assign trial subject to intervention or control group using computer-generated
randomization with 50-50 chance for each group for a total of 126 patients
3. Double - Blind
Both patients and physician are unaware of treatment assignment
Drop - outs
•
complete analysis
P = # effective case / N
•
weighting method
P = # effective case / N – drop outs
•
imputation method
- LOCF (Last Observation Carry Forward)
•
incomplete case analysis
- Analyze the odds ratio of drop-outs in different groups
- Use regression analysis to analyze the reasons of drop outs
Repeated Measures
- Pre-post treatment on the same observation
Intervention
group
Pre-treatment
measure
treatment
measure
Placebo
Group
Pre-treatment
measure
Analyze the correlation and the superiority
for the pre-post treatment on the same
observation for each group
treatment
measure
Statistical Analysis
1. Pi – Pc ≥0.2 (difference in proportion)
Pi (1  Pi )
( Pi  Pc ) Z  *
  Z *
2
Pi

Nc
Pi /( 1  Pi )
 
1
Pc (1  Pc )
Ni
2
2. Ө > 2 ( Ө: odds ratio)

Pc /( 1  Pc )
1
1  Pi

1
Pc

1
1  Pc
1/--страниц
Пожаловаться на содержимое документа