{smcl}
{* version 2.1 abuxton 23sep2007.}{...}
{hline}
help for {hi:cureregr} (sep2007) {right:(see also:  {hi:[R] stset})}
{hline}

{title:cureregr}, cure model regression (split-population model) or, {title:parametric-cure model (PCM)}

{p 8 17}{cmd:cureregr} [{it:varlist}] [{cmd:if} {it:exp}] [{cmd:in} {it:range}] 
{it:,}
         [{cmdab:dist:ribution_option(}{it:weibull | lognormal | logistic | 
		 gamma | exponential}{cmd:)}
          {cmdab:class:_option(}{it:mixture | non-mixture}{cmd:)}
          {cmdab:link:_option(}{it:logistic | lml, log-minus-log | linear}
		  {cmd:)}
          {cmdab:sc:ale_option(}{it:varlist}{cmd:)}
          {cmdab:sh:ape_option(}{it:varlist}{cmd:)}
          {cmdab:noconst:ant}{cmd:}
          {cmdab:scale_noc}{cmd:}
          {cmdab:shape_noc}{cmd:}
	  {cmdab:nodesc:riptionmodel}{cmd:} 
	  {cmdab:nosh:ow}{cmd:}
          {cmdab:left}{cmd:}]
	
{title:Description}

{p 4 4}{cmd:cureregr} fits a parametric cure model in either the non-mixture or 
mixture class.{p_end}
{p 8 8} important -- {cmd:cureregr} requires that the data be {cmd:stset} prior 
to use. {cmd:cureregr} will 
correctly estimate multiple records per subject, to include 
time-varying-covariates. Please note that
for multiple records per subject the records must be in contiguous intervals 
defined in {cmd:stset}, [_t0, _t]; 
in other words, there may be no gaps (see: {cmd:stdes}).
{p_end}

{title:Options}

{p 4 8}{cmdab:dist:ribution_option(}{it:weibull | lognormal | logistic | gamma | exponential}{cmd:)}{p_end}
{p 8 12} {it:define Kernel:}{p_end}
{p 12 16} weibull: exponential in tt {p_end}
{p 12 16} lognormal: norm(ln(tt)) {p_end}
{p 12 16} logistic: tt/(1-tt) {p_end}
{p 12 16} gamma: gammap(shape, scale, time) {p_end}
{p 12 16} exponential: gammap(shape==1, scale, time) {p_end}
{p 4 8}{cmdab:class:_option(}{it:mixture | non-mixture}{cmd:)}{p_end}
{p 8 12} {it: define parametric-cure model class:}{p_end}
{p 12 16} mixture: pi+(1-pi)*(1-Kernel){p_end}
{p 12 16} non-mixture: pi^Kernel{p_end}
{p 4 8}{cmdab:link:_option(}{it:logistic | lml, log-minus-log | linear}{cmd:)}
{p_end}
{p 8 12} {it: define cure_fraction (pi) link funtion} {p_end}
{p 12 16} logistic: exp(xb_pi)/(1+exp(xb_pi)){p_end}
{p 12 16} lml: exp(-exp(xb_pi)){p_end}
{p 12 16} linear: exp(xb_pi){p_end}
{p 16 16} {it:where} xb_pi {it:is the linear model for the cure_fraction, pi}
{p_end}
{p 4 8}{cmdab:sc:ale_option(}{it:varlist}{cmd:)}{p_end}
{p 12 16}{it: varlist for scale model}{p_end}
{p 4 8}{cmdab:sh:ape_option(}{it:varlist}{cmd:)}{p_end}
{p 12 16}{it: varlist for shape model}{p_end}
{p 4 8}{cmdab:noconst:ant}{cmd:}{p_end}
{p 12 16}{it: specifies that no constant term for cure-fraction model is to be fit}{p_end}
{p 4 8}{cmdab:scale_noc}{cmd:}{p_end}
{p 12 16}{it: specifies that no constant term for scale model is to be fit}
{p_end}
{p 4 8}{cmdab:shape_noc}{cmd:}{p_end}
{p 12 16}{it: specifies that no constant term for shape model is to be fit}
{p_end}
{p 4 8} -- {p_end}
{p 4 8}{cmdab:left} specifies that there are left censored subjects.
the failures defined by the last value specified by the failure()
argument of {cmd:stset} are taken to be left censored. this {res}creates{txt} 
(or {res}overwrites{txt}) a variable called {it:{res}_leftd{txt}} and is indicated 
in e(depvar).{p_end}
{p 4 8} -- {p_end}
{p 4 8}{cmdab:nodesc:riptionmodel}{cmd:}{p_end}
{p 12 16}{it: do not show cureregr model definition information}
{p_end}
{p 4 8}{cmdab:nosh:ow}{cmd:}{p_end}
{p 12 16}{it: do not show st setting information}
{p_end}

{title:please note these definitions}
{p 4 8} tt = (scale*time)^shape {it: function of time, t-tilde}{p_end}
{p 8 12} scale=exp(xb_scale) {it:for the scale model}{p_end}
{p 8 12} shape=exp(xb_shape) {it:for the shape model}{p_end}

{title:Example}

{p 8 12}{inp:. cureregr g1_ageyr, dist(weibull) class(n) link(logi) noconstant}
{p_end}

{p 8 12}{inp:. xi: cureregr  g0_ageyr i.g0_wbc i.g0_plt, sc(i.g0_wbc i.g0_plt) sh(i.g0_plt) dist(lognormal) class(non) link(logi) scale_noc shape_noc}{p_end}

{p 8 12}{inp:. xi: cureregr (i.study_r) i.g_liver i.race2 if analset1 ,}
		{inp: dist(weibull) link(lml) sc((i.study_r) ) sh((i.study_r) )}
		{inp: class(nonmixture) technique(nr)}{p_end}

{p 8 12}{res}{it: predicting }{txt}{p_end}

{p 8 12}{cmd:predict} [{cmd:if} {it:exp}] [{cmd:in} {it:range}] {it:,}
         [{cmdab:at:(}{txt}{it:numlist}{cmd:)}{txt} | 
		  {cmdab:all:}{txt} |
          {cmdab:s:urvival}{txt} |
          {cmdab:se:s}{txt} |
          {cmdab:u:cs}{txt} |
          {cmdab:l:cs}{txt} |
          {cmdab:h:azard}{txt} |
          {cmdab:fd:}{txt}]
		  [{cmdab:gen:(}{txt}{it:varlist}{cmd:)}{txt}]
		  [{cmdab:level:(}{txt}{it:number}{cmd:)}{txt}]
		  [{cmdab:fmt:print(}{txt}{it:%8.6f}{cmd:)}{txt}]
		  {p_end}

{p 16 12} After model estimation the survival, standard error, and confidence 
interval can be displayed with {cmd:predict}, 
{cmdab:at:()} specified time points. The survival estimates are displayed 
for all covariate patterns and {cmd:if} or {cmd:in} may be used to limit the 
number of covariate patterns displayed; An example can be found in the ancillary file, cureregr_predict_eg.pdf. The {cmd:fmt()} option allows any acceptable numeric format specification for the displayed numeric results. Please note also that the 
confidence limits are based on a ln(-ln(S)) transform. {p_end}

{p 16 12} Six quantities may be estimated and placed into a new variable: 
survival (S), and its standard error (seS), lower confidence limit (lciS), 
and upper confidence limit (uciS). In addition, the hazard (haz), or the fail 
density (fd) may be estimated. One of the key words (s, se, l, u, h, fd, 
or all) can be specified. With the option {cmdab:gen:()}, the estimate 
requested by the keyword will be placed in the new variable, while the 
keyword {cmdab:all:} requires six new variable names in order:
survival, ses, lcs, ucs, hazard, and fd. Where the default names are: {it:S}, 
{it:seS}, {it:lciS}, {it:uciS}, {it:haz}, and {it:fd}, respectively.{p_end}

{p 8 12}{inp:. predict if study_r==1 & g_liver==1 & race2==2, at(30(30)365 2922)}{p_end}
{p 8 12}{inp:. predict ,  all gen(s_v2 se_v2 lci90_v2 uci90_v2 haz_v2 fd_v2) level(90)}{p_end}
{p 8 12}{inp:. predict ,  hazard gen(myhaz)}{p_end}
{p 8 12}{inp:. predict ,  lcs gen(lci95_v1) level(95)}{p_end}
{p 8 12}{inp:. predict ,  ucs gen(uci95_v1) level(95)}{p_end}

{title:Author}
{p 8 12} Allen Buxton, CureSearch, Children's Oncology Group, Arcadia, CA{p_end}
{p 4 8} - acknowledgements - {p_end}
{p 8 12} Richard Sposto, Mark Krailo, & Todd Alonzo, Keck School of Medicine, 
University of Southern California, USA{p_end}
{p 8 12} John Thompson & Clarie Weston, Department of Health Sciences, 
University of Leicester, UK{p_end}

{title:References}
{p 8 12} Sposto R. Cure model analysis in cancer: An application to data from 
the Children's Cancer Group. Statistics in Medicine, 21: 293-312, 2002{p_end}
{p 8 12} Stephen Jenkins, spsurv (spsurv.ado for Stata, 23may2001), University 
of Essex, UK{p_end}
{p 8 12} Maller RA. & Zhou, X. Survival Analysis with Long Term Survivors, 
John Wiley, 1997. {p_end}
{p 8 12} Schmidt P. & Witte A. Predicting criminal recidivism using 
'split-population' survival time models, Journal of Econometrics, 40: 141-159, 
1989{p_end}

{title:Files}
{p 8 12} main: cureregr.ado, & ml {it:evaluators}: PCM{it:mmkkll}.ado{p_end}
{p 8 12} predict: _cureregr.ado & _cureregr_gen.ado{p_end}
{p 4 8} - ancillary - {p_end}
{p 8 12} pcm_key.pdf, a description of the models {cmd:cureregr} will fit.
{p_end}
{p 8 12} cureregr_predict_eg.pdf, example of predict.{p_end}
{p 8 12} example_stata.dta, dataset in the above example.{p_end}

{title:Also see}
{p 1 14}Manual: {hi:[R] stset} {p_end}