{smcl} {* 04sep2009}{...} {hline} help for {hi:stepp_tail}, {hi:stepp_window}, {hi:stepp_plot}{right:Patrick Royston} {hline} {title:Subpopulation treatment effect pattern plot (STEPP)} {phang2} {cmd:stepp_tail} {it:regression_cmd} [{it:yvar}] {it:zvar} [{it:adjvars}] {ifin} {cmd:,} {it:options} {phang2} {cmd:stepp_window} {it:regression_cmd} [{it:yvar}] {it:zvar} [{it:adjvars}] {ifin} {cmd:,} {it:options} {phang2} {cmd:stepp_plot} {it:stubname} [{cmd:, } {it:options}] {pstd} {marker syntax}where {pin} {it:regression_cmd} may be {helpb clogit}, {helpb cnreg}, {helpb glm}, {helpb intreg}, {helpb logistic}, {helpb logit}, {helpb mlogit}, {helpb nbreg}, {helpb ologit}, {helpb oprobit}, {helpb poisson}, {helpb probit}, {helpb qreg}, {helpb regress}, {helpb stcox}, {helpb streg}, or {helpb xtgee}. {synoptset 28 tabbed}{...} {synopthdr} {synoptline} {synopt :{opt gen(stubname)}}creates five new variables containing results of STEPP analysis{p_end} {synopt :{opt g(#)}}({cmd:stepp_tail} only) defines the number of subpopulation groups{p_end} {synopt :{opt n1(#)}}({cmd:stepp_window} only) defines the number of individuals belonging only to one of two neighbouring subpopulations{p_end} {synopt :{opt n2(#)}}({cmd:stepp_window} only) defines the number of individuals in a subpopulation{p_end} {synopt :{opt treatment(trt_varlist)}}defines the list of variables whose interactions with {it:zvar} are to be studied{p_end} {synopt :{it:regression_cmd_options}}options for {it:regression_cmd}{p_end} {syntab :{it:options} for {cmd:stepp_plot}} {synopt :{opt vn(#)}}variable number in {cmd:treatment()}{p_end} {synopt :{opt plot(plot)}}adds other plots to the generated graph{p_end} {synopt :{it:graph_options}}options of {cmd:graph twoway}{p_end} {synoptline} {p2colreset}{...} {phang} All weight types supported by {it:regression_cmd} are allowed; see {help weight}.{p_end} {phang} {it:yvar} is not allowed for {opt streg} and {opt stcox}. For these commands, you must first {helpb stset} your data. {title:Description} {pstd} {cmd:ststep_tail} and {cmd:ststep_window} compute Bonetti & Gelber (2000, 2004)'s STEPP estimators for graphical exploration of a treatment/covariate interaction. {cmd:ststep_tail} provides the tail-oriented estimator, and {cmd:ststep_window} the sliding-window estimator. Plotting the results may be done by using {cmd:stepp_plot}, in which case {it:stubname} is the same as in the {opt gen(stubname)} option of {cmd:stepp_tail} and {cmd:stepp_window}. {pstd} {it:zvar} is the continuous covariate whose interaction with treatment is to be studied, and {it:adjvars} is a list of other covariates used to linearly adjust each model fitted to the treatment variable(s) defined by {cmd:treatment()}. {title:Options} {pstd}{ul:Options for {cmd:stepp_tail} and {cmd:stepp_window}}: {phang} {opt gen(stubname)} creates five new variables called {it:stubname}{hi:b}, {it:stubname}{hi:se}, {it:stubname}{hi:mean}, {it:stubname}{hi:lb}, {it:stubname}{hi:ub}. {it:stubname}{hi:b} is the estimated regression coefficient in each subpopulation, {it:stubname}{hi:se} is its standard error, {it:stubname}{hi:mean} contains the mean of {it:zvar} in each subpopulation, and {it:stubname}{hi:lb} and {it:stubname}{hi:ub} are pointwise 95% confidence limits for {it:stubname}{hi:b}. If {cmd:treatment()} includes more than one variable, the created variables have 2, 3, ... appended to the names, e.g. {it:stubname}{hi:b2}. {phang} {opt g(#)} ({cmd:stepp_tail} only) defines the number of subpopulation groups. The actual number of subpopulations used is 2 * # - 1. {phang} {opt n1(#)} ({cmd:stepp_window} only) defines the number of individuals belonging only to one of two neighbouring subpopulations. {phang} {opt n2(#)} ({cmd:stepp_window} only) defines the number of individuals in a subpopulation. The overlap between two neighbouring subpopulations is {cmd:n2()} minus {cmd:n1()} individuals. {phang} {opt treatment(trt_varlist)} defines the list of variables whose interactions with {it:zvar} are to be studied. Typically {it:trt_varlist} will comprise just one binary variable, representing the two arms of a parallel-group clinical trial. {phang} {it:regression_cmd_options} are options for {it:regression_cmd}. {pstd}{ul:Options for {cmd:stepp_plot}}: {phang} {opt vn(#)} {it:#} is an integer defining the variable number in {cmd:treatment()}, when more than one variable is specified. When only one variable is specified, {opt vn()} is not required. {phang} {opt plot(plot)} provides a way to add other plots to the generated graph; see help {help plot_option:plot option}. {phang} {it:graph_options} are options of {cmd:graph twoway}, such as {cmd:xtitle()}, {cmd:ytitle()}, etc. {title:Examples} {phang}{cmd:. stepp_tail regress y x a1 a2, g(10) gen(z) treatment(t)}{p_end} {phang}{cmd:. stepp_window stcox x a1 a2, n1(40) n2(50) gen(z) treatment(t)}{p_end} {phang}{cmd:. stepp_plot z, xtitle("Serum rhubarb") ytitle("log relative hazard") name(myplot)}{p_end} {title:Author} {pstd} Patrick Royston, MRC Clinical Trials Unit, London.{break} pr@ctu.mrc.ac.uk {title:References} {phang} M. Bonetti and R. D. Gelber. 2000. A graphical method to assess treatment-covariate interactions using the Cox model on subsets of the data. {it:Statistics in Medicine} 19: 2595-2609. {phang} M. Bonetti and R. D. Gelber. 2004. Patterns of treatment effects in subsets of patients in clinical trials. {it:Biostatistics} 5: 465-481. {phang} P. Royston and W. Sauerbre. 2009. Two techniques for investigating interactions between treatment and continuous covariates in clinical trials. {it:Stata Journal} 9(2): 230-251. {phang} W. Sauerbrei, P. Royston and K. Zapien. 2007. Detecting an interaction between treatment and a continuous covariate: a comparison of two approaches. {it:Computational Statistics and Data Analysis} 51: 4054-4063. {title:Also see} {psee}Article: {it:Stata Journal}, volume 9, number 2: {browse "http://www.stata-journal.com/article.html?article=st0164":st0164} {psee} Online: {help mfpi}{p_end}