competing risks proportional
in [SAS]
|
Prev: How to create self join in SAS DI Studio using JOIN Transformation??
Next: Smarter Text Qualifying Than Just DSD on Import
From: siyuan li on 30 Nov 2009 02:11 hi, all, I want to do a project about the relation between commit suicide and some factors. I plan to use competing risks proportional hazard model .Who can give me any suggestions or help me? maybe someone have done the project before, can share with? though it maybe has nothing to do with the sas, I really need your help. Thanks in advance
From: Shawn Haskell on 30 Nov 2009 11:34 On Nov 30, 2:11 am, siyuan li <lisiyuan0...(a)gmail.com> wrote: > hi, all, > I want to do a project about the relation between commit suicide > and some factors. I plan to use competing risks proportional hazard > model .Who can give me any suggestions or help me? > maybe someone have done the project before, can share with? > though it maybe has nothing to do with the sas, I really need > your help. Thanks in advance Hi Siyuan, competing risks can be tricky, and there are several ways to approach it. It is best to keep the number of risk-types to minimum, so lump some together if it makes sense to do so. Texts for survival analyses seem to be talking more about competing risks these days, and the one I found most helpful was Kleinbaum and Klein's "Survival anlaysis: a self-learning text, 2nd edition" - it was this text that showed me how to do a modified Lunn-McNeil competing risks analysis using SAS PROC PHREG. The advantage of this analysis, compared to more common methods, is that you can consider covariate's potential effects on each type of risk simultaneously in a Type-3 regression context. Hope that helps - still trying to publish those final analyses, but they are in my dissertation on-line - you can Googel "Haskell deer Texas" if you are interested. SH
From: siyuan li on 10 Dec 2009 06:36 On 12ÔÂ1ÈÕ, ÉÏÎç12ʱ34·Ö, Shawn Haskell <shawn.hask....(a)state.vt.us> wrote: > On Nov 30, 2:11 am, siyuan li <lisiyuan0...(a)gmail.com> wrote: > > > hi, all, > > I want to do a project about the relation between commit suicide > > and some factors. I plan to usecompetingrisksproportionalhazard > > model .Who can give me any suggestions or help me? > > maybe someone have done the project before, can share with? > > though it maybe has nothing to do with the sas, I really need > > your help. Thanks in advance > > Hi Siyuan, > > competingriskscan be tricky, and there are several ways to approach > it. It is best to keep the number of risk-types to minimum, so lump > some together if it makes sense to do so. Texts for survival analyses > seem to be talking more aboutcompetingrisksthese days, and the one > I found most helpful was Kleinbaum and Klein's "Survival anlaysis: a > self-learning text, 2nd edition" - it was this text that showed me how > to do a modified Lunn-McNeilcompetingrisksanalysis using SAS PROC > PHREG. The advantage of this analysis, compared to more common > methods, is that you can consider covariate's potential effects on > each type of risk simultaneously in a Type-3 regression context. Hope > that helps - still trying to publish those final analyses, but they > are in my dissertation on-line - you can Googel "Haskell deer Texas" > if you are interested. SH if I want to do a project about the relation between commit suicide and some factors, what the model is better? the data is comprehensive and the data had divided into several groups by age. can somebody give me some suggestions?
From: Joe Whitehurst on 10 Dec 2009 10:25 Siyuan Li, Before you jump into a "project about the relation between commit suicide and some factors', it might be prudent to first investigate existing theories of suicidal behavior even though there are damn few worthwhile theories about any aspect of human psychological functioning. I would suggest you start with http://www.pcp-net.org/encyclopaedia/constriction.html. Constricted construing is a dead give away of potential suicidal behavior and is well documented in the psychological literature. Joe 2009/12/10 siyuan li <lisiyuan0753(a)gmail.com> > On 12=D4=C21=C8=D5, =C9=CF=CE=E712=CA=B134=B7=D6, Shawn Haskell <shawn.ha= sk...(a)state.vt.us> wrote: > > On Nov 30, 2:11 am, siyuan li <lisiyuan0...(a)gmail.com> wrote: > > > > > hi, all, > > > I want to do a project about the relation between commit suicide > > > and some factors. I plan to usecompetingrisksproportionalhazard > > > model .Who can give me any suggestions or help me? > > > maybe someone have done the project before, can share with? > > > though it maybe has nothing to do with the sas, I really need > > > your help. Thanks in advance > > > > Hi Siyuan, > > > > competingriskscan be tricky, and there are several ways to approach > > it. It is best to keep the number of risk-types to minimum, so lump > > some together if it makes sense to do so. Texts for survival analyses > > seem to be talking more aboutcompetingrisksthese days, and the one > > I found most helpful was Kleinbaum and Klein's "Survival anlaysis: a > > self-learning text, 2nd edition" - it was this text that showed me how > > to do a modified Lunn-McNeilcompetingrisksanalysis using SAS PROC > > PHREG. The advantage of this analysis, compared to more common > > methods, is that you can consider covariate's potential effects on > > each type of risk simultaneously in a Type-3 regression context. Hope > > that helps - still trying to publish those final analyses, but they > > are in my dissertation on-line - you can Googel "Haskell deer Texas" > > if you are interested. SH > > if I want to do a project about the relation between commit suicide > and some factors, what the model is better? the data is comprehensive > and the data had divided into several groups by age. > can somebody give me some suggestions? >
From: siyuan li on 10 Dec 2009 23:30
On 12æ10æ¥, ä¸å11æ¶25å, joewhitehu...(a)GMAIL.COM (Joe Whitehurst) wrote: > Siyuan Li, > > Before you jump into a "project about the relation between commit suicide > and some factors', it might be prudent to first investigate existing > theories of suicidal behavior even though there are damn few worthwhile > theories about any aspect of human psychological functioning.  I would > suggest you start withhttp://www.pcp-net.org/encyclopaedia/constriction.html.  Constricted > construing is a dead give away of potential suicidal behavior and is well > documented in the psychological literature. > > Joe > > 2009/12/10 siyuan li <lisiyuan0...(a)gmail.com> > > > > > On 12=D4=C21=C8=D5, =C9=CF=CE=E712=CA=B134=B7=D6, Shawn Haskell <shawn.ha= > sk...(a)state.vt.us> wrote: > > > On Nov 30, 2:11 am, siyuan li <lisiyuan0...(a)gmail.com> wrote: > > > > > hi, all, > > > >   I want to do a project about the relation between commit  suicide > > > > and some factors. I plan to usecompetingrisksproportionalhazard > > > > model .Who can give me any suggestions or help me? > > > >   maybe someone have  done the project before, can share with? > > > >   though  it maybe has  nothing to do with the sas, I really need > > > > your help. Thanks in advance > > > > Hi Siyuan, > > > > competingriskscan be tricky, and there are several ways to approach > > > it.  It is best to keep the number of risk-types to minimum, so lump > > > some together if it makes sense to do so.  Texts for survival analyses > > > seem to be talking more aboutcompetingrisksthese days, and the one > > > I found most helpful was Kleinbaum and Klein's "Survival anlaysis: a > > > self-learning text, 2nd edition" - it was this text that showed me how > > > to do a modified Lunn-McNeilcompetingrisksanalysis using SAS PROC > > > PHREG.  The advantage of this analysis, compared to more common > > > methods, is that you can consider covariate's potential effects on > > > each type of risk simultaneously in a Type-3 regression context.  Hope > > > that helps - still trying to publish those final analyses, but they > > > are in my dissertation on-line - you can Googel "Haskell deer Texas" > > > if you are interested.  SH > > > if I want to do a project about the relation between commit  suicide > > and some factors, what the model is better? the data is comprehensive > > and the data had divided into several groups by age. > > can somebody give me some suggestions?- éè被å¼ç¨æå - > > - æ¾ç¤ºå¼ç¨çæå - thanks very much. I find most foreigners are willing to privide help for anybody,it is very commendable and different with we chinese |