Fraity Model with PhReg

[iuri_leite]

Dear All,

We are analyzing data from a cohort of participants with a high risk of HIV infection. At this moment, our focus is on sexually transmitted disease infection. Given that patients in our cohort may have more than one sexually transmitted disease, we are considering using a frailty model.

We are currently thinking of using the COX Regression model, but I would like to check if the data structure I am creating is correct.  

The structure of our database would incorporate time-dependent variables.

Here we have the data for three participants. The observation of the participant 1 (ID=1) is censored at the first visit, then he experiences the event at the second visit, and does not return. The second participant experiences several events while the third one is censored at all visits.

ID  TSTART TSTOP STATUS 

1         0           91         0

1        91        187        1

2         0          115       0

2         115       210      1

2         210       321      0

2        321        430      1

2        430        542      1

3        0            122      0

3        122       312       0

3       312       431        0

The other issue I wanted to address is how to decide on the distribution of the random parameter associated with frailty. Is there a specific test for this?

Thank you..

Best regards,

Iuri leite

[sbxkoenk]

Hello,

I am not an expert in frailty models (for sure not in practice), but here's some info.

• The PHREG procedure in SAS implements the penalized likelihood method for the gamma and log-normal frailty models.
  SAS/STAT 15.3 User's Guide
  The PHREG Procedure
  The Frailty Model
  https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/statug/statug_phreg_details16.htm
• Paper 1492-2014
  Introduction to Frailty Models
  John Amrhein, McDougall Scientific Ltd.
  https://support.sas.com/resources/papers/proceedings14/1492-2014.pdf
• If you encounter recurrent disease episodes in patients, then this paper may be of interest:
  Analysis of Correlated Recurrent and Terminal Events Data in SAS®
  Li Lu(1), Chenwei Liu(2)
  1 The EMMES Corporation, Rockville, MD
  2 Core Genotyping Facility, Division of Cancer Epidemiology and Genetics, Advanced Technology Program, SAIC Frederick, Inc., NCI-Frederick, Frederick, MD
  https://www.lexjansen.com/nesug/nesug08/sa/sa16.pdf

BR, Koen