Researchers conclusion was no drug effect! - USMLE Forums
USMLE Forums Logo
USMLE Forums         Your Reliable USMLE Online Community     Members     Posts
Home
USMLE Articles
USMLE News
USMLE Polls
USMLE Books
USMLE Apps
Go Back   USMLE Forums > USMLE Step 2 CK Forum

USMLE Step 2 CK Forum USMLE Step 2 CK Discussion Forum: Let's talk about anything related to USMLE Step 2 CK exam


Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 11-11-2011
USMLE Forums Scout
 
Steps History: Step 1 Only
Posts: 76
Threads: 10
Thanked 31 Times in 27 Posts
Reputation: 41
Stats Researchers conclusion was no drug effect!

A 1-year study of a new drug to treat hypertension is conducted. One hundred patients with hypertension are enrolled; 50 patients are given the new drug and another 50 patients are given hydrochlorothiazide. All patients completed the trial. One noted unexpected effect is increased growth of scalp hair which occurred in those taking the new drug, a nonstatistically significant difference (p>0.10). This effect has also been reported in studies of other similar drugs in the new therapeutic class. The investigators of the study concluded that the new drug did not cause hair growth. Which of the following features of this study is most likely to affect the validity of this conclusion?

A) Differential follow-up
B) Lead time bias
C) Length of the study
D) Sample size
E) Self-selection
Reply With Quote Quick reply to this message



  #2  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

In this case 2 answers are plausible...

1.- Differential follow-up
2.- Sample Size

But here they say that were NO drop out...so the answer should be Sample Size
Reply With Quote Quick reply to this message
  #3  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

bebix what about c...
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.
Reply With Quote Quick reply to this message
 
  #4  
Old 11-11-2011
USMLE Forums Scout
 
Steps History: Step 1 Only
Posts: 76
Threads: 10
Thanked 31 Times in 27 Posts
Reputation: 41
Default

Differential follow up aka attrition bias is a very common bias in cohort studies where cases drop out while in the study. You can call it a sub type of selection bias. In the given context it can surely thawrt validity but more so for results on blood pressures. Here. Issue is on hair growth. So i think it could just be sample size, Best study designs to find out side effects of any drug are special interventional studies that are done on large sample sizes probably in hundreds. Small sample cohort studies are not adequate for side effect evaluation of any new drug.
Reply With Quote Quick reply to this message
  #5  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

Quote:
Originally Posted by busterbee View Post
bebix what about c...
Well, C & E are more remote possibilities...we dont have all the details about the actual design...but assuming same follow up time and 0 drop out, if you have a p=0.1, maybe if you increase your power (decreasing your type II error) w/ same preset alpha you would actually get an stat. significant difference between groups
Reply With Quote Quick reply to this message
The above post was thanked by:
iron (11-11-2011)
  #6  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

Quote:
Originally Posted by usmlepak View Post
Differential follow up aka attrition bias is a very common bias in cohort studies where cases drop out while in the study. You can call it a sub type of selection bias. In the given context it can surely thawrt validity but more so for results on blood pressures. Here. Issue is on hair growth. So i think it could just be sample size, Best study designs to find out side effects of any drug are special interventional studies that are done on large sample sizes probably in hundreds. Small sample cohort studies are not adequate for side effect evaluation of any new drug.
btw, attrition is NOT a problem in a randomized trail (like this one), because the analysis always follows Intention to treat method!...this is not a cohort study!!!
Reply With Quote Quick reply to this message
  #7  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

Quote:
Originally Posted by bebix View Post
Well, C & E are more remote possibilities...we dont have all the details about the actual design...but assuming same follow up time and 0 drop out, if you have a p=0.1, maybe if you increase your power (decreasing your type II error) w/ same preset alpha you would actually get an stat. significant difference between groups
you are right ...if we increase the study sample that would increase the power of study..and it will be statistically significant...
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

Last edited by busterbee; 11-11-2011 at 10:33 AM.
Reply With Quote Quick reply to this message
  #8  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

Quote:
Originally Posted by busterbee View Post
i need to work hard regarding biostatics...bebix can you please explain this Q in layman's words for me...
so, we have 2 groups - randomized to:
D = drug n=50
no D = no drug n=50

Follow up = 1 year
No drop out = everybody was measured = no attrition

Preset alpha (probability type I error = Probability of reject H0 when H0 is true) = 0.05

Analysis group D vs no D diff. in secondary outcome with p=0.10

"General formula" for sample size calculation:

n = [alpha * beta * variability]/[minimum significant difference]

- actually we use the Z value of alpha & beta...like alpha = 0.05 ==> Z = 1.96
- and variability we use SD (most of the time)
- low probability ==> greater Z score

So, keeping alpha, variability and min.sig. diff. equal, if we decreased beta, Z score goes up ==> sample size goes up ==> the chances of detecting an smaller min.sig.diff goes up.
Reply With Quote Quick reply to this message
The above post was thanked by:
busterbee (11-11-2011), iron (11-11-2011)
  #9  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

Quote:
Originally Posted by bebix View Post
so, we have 2 groups - randomized to:
D = drug n=50
no D = no drug n=50

Follow up = 1 year
No drop out = everybody was measured = no attrition

Preset alpha (probability type I error = Probability of reject H0 when H0 is true) = 0.05

Analysis group D vs no D diff. in secondary outcome with p=0.10

"General formula" for sample size calculation:

n = [alpha * beta * variability]/[minimum significant difference]

- actually we use the Z value of alpha & beta...like alpha = 0.05 ==> Z = 1.96
- and variability we use SD (most of the time)
- low probability ==> greater Z score

So, keeping alpha, variability and min.sig. diff. equal, if we decreased beta, Z score goes up ==> sample size goes up ==> the chances of detecting an smaller min.sig.diff goes up.
how do we decrease beta error??
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.
Reply With Quote Quick reply to this message
The above post was thanked by:
iron (11-11-2011)
  #10  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

Quote:
Originally Posted by busterbee View Post
how do we decrease beta error??
Well, alpha & beta probabilities are numbers selected by the researcher
alpha almost always is preset = 0.05
range of beta = 0.10 to 0.20
Reply With Quote Quick reply to this message
  #11  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

Quote:
Originally Posted by bebix View Post
Well, alpha & beta probabilities are numbers selected by the researcher
alpha almost always is preset = 0.05
range of beta = 0.10 to 0.20
OK BIOSTAT

THANKS
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.
Reply With Quote Quick reply to this message
  #12  
Old 11-11-2011
iron's Avatar
USMLE Forums Veteran
 
Steps History: 1+CK+CS+3
Posts: 229
Threads: 43
Thanked 52 Times in 43 Posts
Reputation: 62
Default

Quote:
Originally Posted by bebix View Post
so, we have 2 groups - randomized to:
D = drug n=50
no D = no drug n=50

Follow up = 1 year
No drop out = everybody was measured = no attrition

Preset alpha (probability type I error = Probability of reject H0 when H0 is true) = 0.05

Analysis group D vs no D diff. in secondary outcome with p=0.10

"General formula" for sample size calculation:

n = [alpha * beta * variability]/[minimum significant difference]

- actually we use the Z value of alpha & beta...like alpha = 0.05 ==> Z = 1.96
- and variability we use SD (most of the time)
- low probability ==> greater Z score

So, keeping alpha, variability and min.sig. diff. equal, if we decreased beta, Z score goes up ==> sample size goes up ==> the chances of detecting an smaller min.sig.diff goes up.
this formula i haven't seen so for, at least not in uworld..
Reply With Quote Quick reply to this message
  #13  
Old 11-11-2011
iron's Avatar
USMLE Forums Veteran
 
Steps History: 1+CK+CS+3
Posts: 229
Threads: 43
Thanked 52 Times in 43 Posts
Reputation: 62
Default

Quote:
Originally Posted by busterbee View Post
OK BIOSTAT

THANKS
same here.... I was thinking my biostat has improved but seeing this question I think I will again go into depression, will come to this question once again after revising biostat..
Reply With Quote Quick reply to this message
  #14  
Old 11-11-2011
bebix's Avatar
USMLE Forums Master
 
Steps History: 1+CK+CS+3
Posts: 1,357
Threads: 194
Thanked 3,268 Times in 881 Posts
Reputation: 3278
Default

Quote:
Originally Posted by iron View Post
this formula i haven't seen so for, at least not in uworld..
YOU DONT have to know this formula....and btw, this is only the general formula (just to explain a concept)...they never going to ask u this!!!
Here is the actual formula
http://en.wikipedia.org/wiki/Sample_size_determination

Reply With Quote Quick reply to this message
  #15  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

Quote:
Originally Posted by iron View Post
same here.... I was thinking my biostat has improved but seeing this question I think I will again go into depression, will come to this question once again after revising biostat..
you know me..i get it n forget it..its easy but only once you get it
seriously after all the assessments i did i gave time to biostat n still dont cant get it..so its all left to fate
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.
Reply With Quote Quick reply to this message



  #16  
Old 11-11-2011
iron's Avatar
USMLE Forums Veteran
 
Steps History: 1+CK+CS+3
Posts: 229
Threads: 43
Thanked 52 Times in 43 Posts
Reputation: 62
Default

Quote:
Originally Posted by bebix View Post
YOU DONT have to know this formula....and btw, this is only the general formula (just to explain a concept)...they never going to ask u this!!!
Here is the actual formula
http://en.wikipedia.org/wiki/Sample_size_determination

heheh, seeing the real formula there I just did miss a beat ,anyway thanks a lot ..
Reply With Quote Quick reply to this message
  #17  
Old 11-11-2011
iron's Avatar
USMLE Forums Veteran
 
Steps History: 1+CK+CS+3
Posts: 229
Threads: 43
Thanked 52 Times in 43 Posts
Reputation: 62
Default

[QUOTE=busterbee;78379]you know me..i get it n forget it..its easy but only once you get it
seriously after all the assessments i did i gave time to biostat n still dont cant get it..so its all left to fate[

I know, just don't worry abt it anymore, do well in other sectors.
Reply With Quote Quick reply to this message
  #18  
Old 11-11-2011
busterbee's Avatar
USMLE Forums Veteran
 
Steps History: 1 + CK
Posts: 218
Threads: 16
Thanked 110 Times in 59 Posts
Reputation: 120
Default

[QUOTE=iron;78382]
Quote:
Originally Posted by busterbee View Post
you know me..i get it n forget it..its easy but only once you get it
seriously after all the assessments i did i gave time to biostat n still dont cant get it..so its all left to fate[

I know, just don't worry abt it anymore, do well in other sectors.
thankyou for your support guys...i will ace...INSHALLAH
__________________
you are born with wings, don't walk through life
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.
Reply With Quote Quick reply to this message



Reply

Tags
Biostatistics-Epidemiology, Step-2-Questions

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the USMLE Forums forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:
Medical School
Choose "---" if you don't want to tell. AMG for US & Canadian medical schools. IMG for all other medical schools.
USMLE Steps History
What steps finished! Example: 1+CK+CS+3 = Passed Step 1, Step 2 CK, Step 2 CS, and Step 3.

Choose "---" if you don't want to tell.

Favorite USMLE Books
What USMLE books you really think are useful. Leave blank if you don't want to tell.
Location
Where you live. Leave blank if you don't want to tell.

Log-in

Human Verification

In order to verify that you are a human and not a spam bot, please enter the answer into the following box below based on the instructions contained in the graphic.



Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes


Similar Threads
Thread Thread Starter Forum Replies Last Post
The effect of MRCP (UK) on US residency? Dr. N Elham IMG Residency Match Forum 12 11-02-2013 02:15 PM
Drug-induced Adverse Effect Evergreen USMLE Step 1 Forum 2 11-05-2011 11:22 PM
Anti-Parkinson Drug Side Effect! usmleman2020 USMLE Step 1 Forum 5 08-10-2011 10:56 AM
effect of prednisone ricko335 USMLE Step 1 Forum 7 08-07-2011 01:10 AM
question - lithium toxicity drug-drug interaction mtoi USMLE Step 1 Forum 5 10-31-2010 12:48 PM

RSS Feed
Find Us on Facebook
vBulletin Security provided by vBSecurity v2.2.2 (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.

USMLE® & other trade marks belong to their respective owners, read full disclaimer
USMLE Forums created under Creative Commons 3.0 License. (2009-2014)