Print

Print


          Dear William 
when you have several covariates that you believe contribute to survival?
            you can use stratified Kaplan Meier  curves – for 2 or maybe 3 covariates
            If you have more than 3 covariate you need another approach which is Cox proportional hazards model
Example: Smoking, hyperlipidemia, diabetes, hypertension,  contribute to development of myocardial infarction
            Cox proportional hazards model used  to assess  the effect of multiple covariates on survival
            it is the most commonly used multivariate survival method
            Easy to implement in SPSS, Stata, or SAS , Can handle both continuous and categorical predictor variables (logistic, linear regression)
            Without knowing baseline hazard, can still calculate coefficients for each covariate, and therefore hazard ratio

 Best wishes
 
Rajaa Mohammad Al-Raddadi
MBBS,ABCM, RICR
Consultant Community Medicine
Head of Research Department,PHC,Jeddah ,Saudi Arabia
Trainer, Postgraduate Center for Family and Community Medicine
Vice President Saudi Epidemiology Association 
http://health.groups.yahoo.com/group/Saudiresearch/

From: William Grant <[log in to unmask]>
To: [log in to unmask]
Sent: Wednesday, February 15, 2012 5:59 PM
Subject: Re: Kaplain Meire vs Cox regression analysis

Dr. Tiwan
 
Whoa, Halt.  Carefully step away from the data.
 
Is this data in search of an analysis or are you at the design specification stage of determining research design and analysis?
 
It is a very important question as the specification of the research question and identification of outcomes, and predictors as well as
the selection of analysis will determine what data is required to be collected.
You may have a data set already collected but it may not include the data you need for a specific analysis.
 
Basically:
 
Kaplan-Meier is a straight forward analysis of survival.  Subjects all start at a baseline and then you record the time
to a specific event such as death. 
 
Cos survival analysis allows you to determine the effect of one or more risk factors on survival times.
 
Then there are various approaches to assessing differences in two or more groups.
 
The approaches answer similar questions but in different ways. 
 
If you are interested only in survival the Kaplan-Meier might suffice.  If, for example, you are interested in determining
which life style issues might influence survival so that you may counsel patients on life style changes then you
should consider applying the Cox approach.
 
If you are a beginner in this, my advice is to find a friendly statistician who will help you with your research.
 
Bill
 
 
 
William D. Grant, EdD
Associate Dean, Graduate Medical Education
Professor, Emergency Medicine
Professor, Family Medicine
SUNY Upstate Medical University
750 E. Adams St.  EmStat
Syracuse, NY 13210
315-464-4861 (p)    315-464-4854 (f)

STATEMENT OF CONFIDENTIALITY
The information contained in this electronic message and any attachments to the message are intended for the exclusive use of the addressee(s) and may contain confidential or privileged information that is protected by law.  If you are not the intended recipient, please notify the sender at (315) 464-6997 or at [log in to unmask], and destroy all copies of this message and any attachments.  This communication may not be forwarded without the express written permission of the original author.


>>> "Dr. Kaushal Tiwari" <[log in to unmask]> 2/15/2012 9:30 AM >>>
Dear colleagues
Can someone explain in simple language the difference between Kaplain Meier survival curve and Cox Proportional-Hazards Regression for Survival Data.
I am almost beginner in the statistics and need these methods to apply for my study.
Thank you all for your contribution in advance.
Best regards
Kaushal
 
_______________________________________
Dr Kaushal K Tiwari, MD (Hons), MS, IMCS
PhD Fellow in Cardiac Surgery
Sant' Anna School for Higher Studies
G. Pasquinucci Heart Hospital
Via Aurelia Sud, 303
Massa, 54100
Italy
Mobile - +393881246366