{smcl}
{hline}
help for {hi:stlin_compet}
{hline}
{title:Calculating Lin censoring-adjusted estimates of mean costs, KM censoring-adjusted mean effects and incremental cost-effectiveness ratio, with adjustment for competing risks}
{p 8 17 2}
{cmd:stlin_compet}
{it:group}
{it:eventdate(eventcost)}
{cmd:,}
{cmd:time}
{cmd:fail}
[ {cmdab:len:gth}
{cmdab:disce:ffects}
{cmdab:discc:osts}
{cmdab:l:evel}
{cmd:inb}
{cmd:ceac}
{cmd:ellipse}
{cmd:grmax}
{cmd:grmin}
{cmd:scale}
{cmd:compet}
]
{p 4 4 2}
{cmd:stlin_compet} is for use with survival-time data. You must {cmd:stset} your
data before using this command; see help {help stset}.
{p 4 4 2}
{it:group} contains the name of the variable holding the intervention
assignment. {it:group} must contain binary data (two-arm trials only) coded 1,2.
{p 4 4 2}
{it:group} must be followed by costed event date variables and costs in the format {it:eventdate(eventcost)}. The number of costed events indicated in this manner is not limited.
{title:Description}
{p 4 4 2}
{cmd:stlin_compet} calculates the Lin estimates and confidence intervals of
censoring-adjusted mean costs, together with Kaplan-Meier estimates of censoring-adjusted (restricted) mean effects, and overall cost-effectiveness.
Results can be adjusted for competing risks (combined across both randomised groups).
{title:Options}
{p 4 8 2}
{cmd:time(}{it:#}{cmd:)} is where the user must specify the length of
follow-up (in months). {cmd:time} is not optional, and there is no default
value. Events are truncated at this time-point past the origin specified
in {cmd:stset}.
{p 4 8 2}
{cmd:fail(}{it:varname}{cmd:)} is where the user must specify the variable holding the failure indictor. {cmd:fail} is not optional. Analyses adjusted for competing risks through use of the {cmd:compet} option must use a fail variable holding 0 for no failure, 1 for failure of interest and 2 for competing risk failure.
{p 4 8 2}
{cmd:length(}{it:#}{cmd:)} enables the user to specify the length of each interval in months. The Lin method for censoring-adjustment of mean costs splits the data into time intervals, with smaller intervals providing more accurate results. Default is 1 month.
{p 4 8 2}
{cmd:disceffects(}{it:#}{cmd:)} specifies the annual discount rate for effects. Default is 0.035 (3.5%).
{p 4 8 2}
{cmd:disccosts(}{it:#}{cmd:)} specifies the annual discount rate for costs. Default is 0.035 (3.5%).
{p 4 8 2}
{cmd:level} specifies the confidence level, in percent, for the
confidence intervals of the estimates for mean differences in costs and effects, and the ICER; see help {help level}. The default is the current value of $S_level.
{p 4 8 2}
{cmd:inb} requests a graph showing the change in incremental net benefit over values of willingness to pay. Alternatively, {cmd:inb()} may be issued with options specifying details for the graph.
{p 4 8 2}{it:graph_options} are options of {help twoway_scatter:twoway scatter}.
{p 4 8 2}
{cmd:ceac} requests a graph showing the cost-effectiveness acceptability curve over values of willingness to pay. Alternatively, {cmd:ceac()} may be issued with options specifying details for the graph.
{p 4 8 2}{it:graph_options} are options of {help twoway_scatter:twoway scatter}.
{p 4 8 2}
{cmd:ellipse} requests a graph showing the confidence ellipse of mean costs and mean effects. Alternatively, {cmd:ellipse()} may be issued with options specifying details for the graph.
{p 4 8 2}{it:graph_options} are options of {help twoway_scatter:twoway scatter}.
{p 4 8 2}
{cmd:grmin} and {cmd:grmax} specify the minimum and maximum values of willingness to pay to be plotted on both the {cmd:inb} and {cmd:ceac} plots. Defaults are 0 and 100,000.
{p 4 8 2}
{cmd:scale} specifies the scale of the difference in effects to be plotted on the {cmd:ceac} plot, given in years. Default is days (i.e. 1/365).
{title:Remarks}
{p 4 4 2}
For detailed information on this estimation of censoring-adjusted mean
costs, see Lin, Feuer, Etzioni and Wax (1997) Biometrics 53:419-434.
{title:Examples}
{p 4 8 2}{cmd:. stlin_compet group event1(1000) event2(9000), time(24) fail(failvar)}{p_end}
{p 4 8 2}{cmd:. stlin_compet group event1(1000) event2(9000), time(24) fail(failvar) len(12) disce(0.015) discc(0.06) inb(saving(inbplot)) grmin(-20000) grmax(80000) compet}
{title:Authors}
{p 4 4 2}L Kim, MRC Biostatistics Unit, Cambridge, UK.
{p 4 4 2}lois.kim@mrc-bsu.cam.ac.uk