Hi,
I'm working on discrete-time event history modeling for a case of repeated events (multiple episodes). I'd like to construct life tables by age and race using this data on multiple episodes per respondent. As far as I can see, the STATA command for life tables is ltable; unfortunately, I don't think this command is set up for repeated events. I am sort of stuck here, since I am not able to find the appropriate code for drawing up a life table for repeated events. I was wondering if there is a specific way/ code/ command to handle this, and would really appreciate any help/ advice you may have in this regard. Thanks! Sowmya Rajan. |
Thanks for the suggestions! I'll try to see if I can figure this out.
Sowmya. On Mon, Sep 20, 2010 at 10:28 AM, Steve Samuels <[hidden email]> wrote: > As far as I know, there is no command that will do this. You will have > to construct the tables in a do-file, and I have no example to show > you. Before starting out, be sure to read > http://www.stata.com/support/faqs/stat/stmfail.html. > > There might well be published examples that you can imitate. A quick > Google search found: > http://library2.usask.ca/theses/available/etd-12192007-232441/unrestricted/ghosh_s.pdf; > see especially Chapter 5. > > Assuming that you want something more: > > How you would set up each table depends on whether your events are of > the same or different types and whether you consider them unordered or > ordered. I would guess yours are ordered. In that case, you have to > decide on the method for your analysis, as outlined in Section 3.2 of > the FAQ page. You might want one table for each order (1st, 2nd, 3rd > occurrence), but the start time might or might not be zero. You also > have to decide on the quantities to appear in the table. The minimum > would be: number at risk at each time point, number of events (of > possibly different types, including exits), and the rates of these > event. If you show cumulative probabilites, they should appropriate to > the table you have created. For example, you might show probabilities > for time to the 1st, 2nd, 3rd events, starting from entry or from the > time of the previous event (or both!). > > This discussion assumes that no individual experiences more than one > event at a time point. If your data are not discrete, but are grouped > (took place within intervals, with unknown order of events within > intervals ), then I suggest that you estimate the proportion of > individuals with at least one event, perhaps the proportion with 2+ > events, and the average event rate (events/person-time). > > In short, with multiple failure data, you have fundamental choices to > make, and the tables you create should reflect those choices. > > > Steve > > Steven J. Samuels > [hidden email] > 18 Cantine's Island > Saugerties NY 12477 > USA > Voice: 845-246-0774 > Fax: 206-202-4783 > > On Sun, Sep 19, 2010 at 9:43 PM, Sowmya Rajan <[hidden email]> wrote: >> Hi, >> I'm working on discrete-time event history modeling for a case of >> repeated events (multiple episodes). I'd like to construct life tables >> by age and race using this data on multiple episodes per respondent. >> As far as I can see, the STATA command for life tables is ltable; >> unfortunately, I don't think this command is set up for repeated >> events. I am sort of stuck here, since I am not able to find the >> appropriate code for drawing up a life table for repeated events. I >> was wondering if there is a specific way/ code/ command to handle >> this, and would really appreciate any help/ advice you may have in >> this regard. >> >> Thanks! >> Sowmya Rajan. >> >> -- >> View this message in context: http://statalist.1588530.n2.nabble.com/ltable-for-repeated-events-tp5548926p5548926.html >> Sent from the Statalist mailing list archive at Nabble.com. >> * >> * For searches and help try: >> * http://www.stata.com/help.cgi?search >> * http://www.stata.com/support/statalist/faq >> * http://www.ats.ucla.edu/stat/stata/ >> > > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ |
In reply to this post by Sowmya Rajan
There is a fairly broad literature on this topic. You can possibly get started by looking at stochastic models in which there are multiple visits to various states. In the survival literature, Dabrowska has been active, but her work is fairly tough.
Another possibility would be Therneau and Grambsch "Modeling Survival Data: Extending the Cox Model" Springer 2001 - Chapter 8 is entitled "Multiple Events per Subject" This book is moderately theoretical so you should have a good course in math stat and basic survival analysis. Tony Peter A. Lachenbruch Department of Public Health Oregon State University Corvallis, OR 97330 Phone: 541-737-3832 FAX: 541-737-4001 -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Steve Samuels Sent: Monday, September 20, 2010 7:29 AM To: [hidden email] Subject: Re: st: ltable for repeated events As far as I know, there is no command that will do this. You will have to construct the tables in a do-file, and I have no example to show you. Before starting out, be sure to read http://www.stata.com/support/faqs/stat/stmfail.html. There might well be published examples that you can imitate. A quick Google search found: http://library2.usask.ca/theses/available/etd-12192007-232441/unrestricted/ghosh_s.pdf; see especially Chapter 5. Assuming that you want something more: How you would set up each table depends on whether your events are of the same or different types and whether you consider them unordered or ordered. I would guess yours are ordered. In that case, you have to decide on the method for your analysis, as outlined in Section 3.2 of the FAQ page. You might want one table for each order (1st, 2nd, 3rd occurrence), but the start time might or might not be zero. You also have to decide on the quantities to appear in the table. The minimum would be: number at risk at each time point, number of events (of possibly different types, including exits), and the rates of these event. If you show cumulative probabilites, they should appropriate to the table you have created. For example, you might show probabilities for time to the 1st, 2nd, 3rd events, starting from entry or from the time of the previous event (or both!). This discussion assumes that no individual experiences more than one event at a time point. If your data are not discrete, but are grouped (took place within intervals, with unknown order of events within intervals ), then I suggest that you estimate the proportion of individuals with at least one event, perhaps the proportion with 2+ events, and the average event rate (events/person-time). In short, with multiple failure data, you have fundamental choices to make, and the tables you create should reflect those choices. Steve Steven J. Samuels [hidden email] 18 Cantine's Island Saugerties NY 12477 USA Voice: 845-246-0774 Fax: 206-202-4783 On Sun, Sep 19, 2010 at 9:43 PM, Sowmya Rajan <[hidden email]> wrote: > Hi, > I'm working on discrete-time event history modeling for a case of > repeated events (multiple episodes). I'd like to construct life tables > by age and race using this data on multiple episodes per respondent. > As far as I can see, the STATA command for life tables is ltable; > unfortunately, I don't think this command is set up for repeated > events. I am sort of stuck here, since I am not able to find the > appropriate code for drawing up a life table for repeated events. I > was wondering if there is a specific way/ code/ command to handle > this, and would really appreciate any help/ advice you may have in > this regard. > > Thanks! > Sowmya Rajan. > > -- > View this message in context: http://statalist.1588530.n2.nabble.com/ltable-for-repeated-events-tp5548926p5548926.html > Sent from the Statalist mailing list archive at Nabble.com. > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ |
Free forum by Nabble | Edit this page |