Dear Malgorzata,

You may check the following links:

http://www.medicine.ox.ac.uk/bandolier/painres/download/whatis/cox_model.pdf

http://www.gog.org/sdcstaff/mikesill/classes/sta575/lectures/LectureNotesChp5.pdf

http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-cox-regression.pdf

http://www.naturalspublishing.com/files/published/988u9y98k5r14f.pdf

I hope it helps!

I am suggesting a simplified formula for 95.0% confidence Interval. If data is qualitative we deal with proportions, the size of the sample can be determined by using the formula with desired Allowable Error (AE). The assumption is that the AE doesn’t exceed 10% or 20% of the positive character (P).

Sample Size=4*P*Q/L2

Where P is the positive character (Q=1-P) and L is the Allowable Error that is10% or 20% of P.

Thank you very much for your help. I have a bit of a problem using the method of Kaplan -Meier, because its methodology is the assumption of mortality (data decrease). And I want to make statistics attendance of traning's health , where people come back. I do not know how to do it and I am looking for articles that can help me to solve this task.

Thank you for the links.

Thank you for the tip, I'll try to take advantage of this assumption.

I greet

Małgorzata

First,

Collet, D. Modelling survival data in medical research. London: Chapman and Hall, 1994, 347p

Next,

The manual  are in BMDP Statistical Sofware, 1990 p.769-806

Life table and Survival function.

More references can be found in my page in the Researchgate and most of them can be found in the internet:

Authors:

Terezinha Aparecida Guedes

and José Walter Pedroza Carneiro.

You may want to have a look at these materials where the event can be repeated.

Multilevel discrete-time event history analysis

http://www.bristol.ac.uk/cmm/software/support/workshops/materials/eha.html

Dear Malgorzata,

You may also check the following links too:

http://www.springer.com/mathematics/probability/book/978-0-387-72834-6

http://www.springer.com/mathematics/probability/book/978-0-387-98784-2

Regards,

Rungo

This is great help. I wish that free is solely introduction, but thank you.

Dear Malgorzata,

You may check the following links:

http://www.tandfonline.com/doi/full/10.1080/15598608.2014.847766#.VKT7C8nuQ6w

Hi,

Your question: Is there any work about the use of statistic survival models (for example, hazard Cox) in health sciences ?

Kindly see the following links:  Its very much helpful to you.

http://www.math.umd.edu/~slud/CourseFil.pdf

http://onlinelibrary.wiley.com/doi/10.1002/0471458546.fmatter/pdf

http://www.amstat.org/chapters/northeasternillinois/pastevents/presentations/summer05_Ibrahim_J.pdf

http://www.cscu.cornell.edu/news/statnews/stnews78.pdf

http://www.medicine.ox.ac.uk/bandolier/painres/download/whatis/cox_model.pdf

https://www.researchgate.net/post/What_is_survival_analysis_And_what_are_the_statistical_methods_to_estimate_survival_time_from_data

Oooo! Thank you very much. All articles are online in full and free of charge. Two works of those I have, but others are very important for me. Great!

Many references have be given. In summary:

If you only need estimate “exception of the estimated mortality”, you do not need survival analyses. The death status is known: yes or no, you can use logistic regression,

The survival analysis analyzes the out outcome: time to death, study not only the death rate, but also how fast the event occurs (say death). A simple example, two treatments, each have 50 patients, both have 10 deaths following in a year, however, the 10 deaths occurred in months 1 for treatment 1, but the 10 deaths occurred in months 12. Using logistic model, study the death rate, there is no mortality rate difference for two treatments, but using survival, there is the difference on survival probability. Treatment 2 is better.

The survival data is incomplete, for those deaths, we know the survival time, but for those survival subjects, we only know they time they survived so far, they may die tomorrow, they may survival a long l time. We call those subjects “cancer”.   

Three methods: Kaplan-Meir (non-parameter), Proportional Hazard Model (Semi-parameter) and life regression are common used (parameter).

For the above, each subject only has one outcome: one death status and the time to death.

Read your statement “And I want to make statistics attendance of traning's health , where people come back”, it seems for each subject may have multiple events, such as hospital visiting.  For multiple time to event data:, you may use:

Marginal model, Conditional A model, Conditional B model.  By using different time to event coding.