{smcl} {* 23dec2014}{...} {title:Title} {p2colset 5 15 17 2}{...} {p2col :{hi:artpep} {hline 2}}ART (Survival Outcomes) - Projection of Events and Power{p_end} {p2colreset}{...} {title:Syntax} {p 8 12 2} {cmdab:artpep} [{cmd:using} {it:filename}[{cmd:.dta}]] [ {cmd:,} {it:options} ] {synoptset 30 tabbed}{...} {synopthdr} {synoptline} {synopt :{opt date:start(ddmmmyyyy)}}Calendar date of trial opening{p_end} {synopt :{opt ed:f0(slist0)}}Survival function in group {p_end} {synopt :{opt epe:riods(#)}}Number of future periods over which projection of power and number of events is to be calculated{p_end} {synopt :{opt ept:s(numlist)}}Numbers of additional patients to be recruited in each period{p_end} {synopt :{opt pts(numlist)}}Numbers of patients recruited in each period since the start of the trial{p_end} {synopt :{opt replace}}Allows {it:filename} to be replaced if it exists{p_end} {synopt :{opt sta:rtperiod(#)}}Period to start reporting projections of events and power{p_end} {synopt :{opt stop:recruit(#)}}Number of periods after which recruitment is to cease{p_end} {synopt :{it:artsurv_options}}Options appropriate to {helpb artsurv}{p_end} {synoptline} {* synopt :{opt n(#)} Number of patients recruited by the end of the current period {p_end}} {title:Description} {pstd} {cmd:artpep} estimates the number of events and power achieved in real time as patients are recruited to a clinical trial with a time-to-event outcome. It allows projection of events and power into the future, allowing "what if?" sensitivity analyses to be evaluated. For example, the effect on power of future patterns of patient accrual and survival might be studied. Quantities calculated by {cmd:artpep} may optionally be stored in a new Stata file called {it:filename} to facilitate further analysis, plotting, etc. {pstd} Note that {cmd:artpep} is available either as a command-line program (described here) or via a dialog, invoked by typing {cmd:db artpep}. A separate help file ({helpb artpepdlg}) is available for the dialog. Running the dialog will also generate the equivalent command line for {cmd:artpep}. {pstd} See {it:Remarks} for details of {cmd:using} {it:filename}. {title:Options} {phang} {opt datestart(ddmmmyyyy)} signifies the date of opening of the trial (i.e. when recruitment started). The date of the end of each period is used to label the output, and is stored in {it:filename} if {cmd:using} is specified. Example: {cmd:datestart(14oct2009)}. {phang} {cmd:edf0(}{it:slist0}{cmd:)} is not optional and gives the survival function in the control group (group 1). This need not be one of the survival distributions to be compared in the trial, unless hratio = 1 for at least one of the groups. The format of {it:slist0} is {it:#1}[{it:#2 ...#r}{cmd:,}{it:#1 #2 ...#r}]. Thus, {cmd:edf0(}p_1 p_2 ... p_r{cmd:,}t_1 t_2 ... t_r{cmd:)} gives the value p_i for the survival function for the event time at the end of time period t_i, i=1,...,r. Instantaneous event rates (i.e. hazards) are assumed constant within time periods, that is, the distribution of time-to-event is assumed to be piecewise exponential. Note that when used in a given calculation up to period T, t_r may validly be less than, equal to or greater than T. If t_r <= T, the rules described in the {cmd:edf0()} option of {help artsurv} are applied to compute the survival function at all periods <=T. If t_r > T, the same calculation is used but estimated survival probabilities for periods > T are not used in the calculation at T, although they may of course be used in calculations (e.g. projections of sample size and events) for periods later than T. {pin} Be aware that use of the {cmd:median()} and {cmd:fp()} options of {cmd:artsurv} may modify the effects and interpretation of {cmd:edf0()}. {phang} {cmd:eperiods(}{it:#}{cmd:)} specifies the number of 'extra' (future) periods over which projection of power and number of events is to be calculated. Default {it:#} is 1. {phang} {cmd:epts(}{it:numlist}{cmd:)} specifies in {it:numlist} the numbers of 'extra' (additional) patients to be recruited in each period following the recruitment phase defined by the {cmd:pts()} option. For example, {cmd:pts(23 12 25) epts(30 30)} would specify 3 initial periods of recruitment followed by 2 further periods. A projection of events and power is required over the 2 further periods. The initial recruitment is of 23 patients in period 1, 12 in period 2 and 25 in period 3; in each of periods 4 and 5, we expect to recruit an additional 30 patients. If the number of items in (or implied by expanding) {it:numlist} is less than that specifed by {cmd:pts()}, the final value in {it:numlist} is replicated as necessary to all subsequent periods. If {cmd:epts()} is not given, the default is that the mean of the numbers of patients specified in {cmd:pts()} is used for all projections. {phang} {cmd:pts(}{it:numlist}{cmd:)} is not optional. {it:numlist} specifies the numbers of patients recruited in each period since the start of the trial, i.e. since randomisation. See {help artsurv} for the definition of a 'period'. The number of items in {it:numlist} defines the number of periods of recruitment so far. For example, {cmd:pts(23 12 25)} specifies 3 initial periods of recruitment, with recruitment of 23 patients in period 1, 12 in period 2, and 25 in period 3. The 'current' period would be period 3 and would be surrounded by parallel lines in the output. {phang} {cmd:replace} allows {it:filename} to be replaced if it exists. {phang} {cmd:startperiod(}{it:#}{cmd:)} specifies {it:#} as the period to start reporting the projections of events and power. To report from the beginning of the trial, specify {cmd:startperiod(1)}. Note that {cmd:startperiod()} does not affect the period at which the calculations are started, only how the results are reported. The default {it:#} is the last period defined by {cmd:pts()}. {phang} {cmd:stoprecruit(}{it:#}{cmd:)} specifies the number of periods after which recruitment is to cease. {it:#} must be no smaller than the number of periods of recruitment implied by {cmd:pts()}. Default {it:#} = 0, meaning continue recruiting indefinitely (no follow-up phase). {phang} {it:artsurv_options} are any of the options of {cmd:artsurv} except {cmd:recrt()}, {cmd:nperiod()}, {cmd:power()}, {cmd:n()}. {title:Remarks} {pstd} The {cmd:using} {it:filename}[{cmd:.dta] construct saves items to {it:filename}{cmd:.dta}, as in the following example: . artpep using myres, pts(15 14 15 16 19 40 44 63 65 56 65 59 78 67) epts(65) eperiods(7) startperiod(1) stoprecruit(15) datestart(24apr2001) alpha(.05) aratios(1 1) hratio(1, .75) ngroups(2) ni(0) onesided(0) trend(0) tunit(3) edf0(0.49 0.26 0.19 0.085 0.057, 4(4)20) median(0) method(l) Date | quarter | #pats #C-events #events Power ----------+----------+----------------------------------------- 23jul2001 | 1 | 15 1 2 0.03521 22oct2001 | 2 | 29 3 5 0.04761 21jan2002 | 3 | 44 5 9 0.06248 23apr2002 | 4 | 60 9 16 0.08023 23jul2002 | 5 | 79 13 23 0.10128 22oct2002 | 6 | 119 19 34 0.12981 22jan2003 | 7 | 163 28 49 0.16942 23apr2003 | 8 | 226 39 70 0.22272 23jul2003 | 9 | 291 54 97 0.29037 23oct2003 | 10 | 347 70 128 0.36653 22jan2004 | 11 | 412 89 162 0.44703 22apr2004 | 12 | 471 110 200 0.52800 23jul2004 | 13 | 549 133 242 0.60767 ----------+----------+----------------------------------------- 22oct2004 | 14 | 616 157 288 0.68285 ----------+----------+----------------------------------------- 21jan2005 | 15 | 681 183 336 0.74893 23apr2005 | 16 | 681 206 381 0.79989 23jul2005 | 17 | 681 226 418 0.83529 22oct2005 | 18 | 681 241 449 0.86056 21jan2006 | 19 | 681 254 475 0.87914 23apr2006 | 20 | 681 265 497 0.89332 23jul2006 | 21 | 681 275 517 0.90446 file myres.dta saved . use myres, replace . describe Contains data from myres.dta obs: 21 vars: 6 19 Sep 2012 09:31 size: 504 --------------------------------------------------------------------------- storage display value variable name type format label variable label --------------------------------------------------------------------------- date float %td date of end of period period float %9.0g quarter period patients float %9.0g Cumulative patients Events float %9.0g Cumulative control-arm events events float %9.0g Cumulative total events power float %9.0g Power --------------------------------------------------------------------------- . list +------------------------------------------------------------+ | date period patients Events events power | |------------------------------------------------------------| 1. | 23jul2001 1 15 1 2 .0352053 | 2. | 22oct2001 2 29 3 5 .0476124 | 3. | 21jan2002 3 44 5 9 .0624845 | 4. | 23apr2002 4 60 9 16 .0802253 | 5. | 23jul2002 5 79 13 23 .1012759 | |------------------------------------------------------------| 6. | 22oct2002 6 119 19 34 .129815 | 7. | 22jan2003 7 163 28 49 .1694168 | 8. | 23apr2003 8 226 39 70 .2227229 | 9. | 23jul2003 9 291 54 97 .2903701 | 10. | 23oct2003 10 347 70 128 .3665293 | |------------------------------------------------------------| 11. | 22jan2004 11 412 89 162 .4470252 | 12. | 22apr2004 12 471 110 200 .5280019 | 13. | 23jul2004 13 549 133 242 .6076685 | 14. | 22oct2004 14 616 157 288 .6828461 | 15. | 21jan2005 15 681 183 336 .7489257 | |------------------------------------------------------------| 16. | 23apr2005 16 681 206 381 .7998888 | 17. | 23jul2005 17 681 226 418 .8352927 | 18. | 22oct2005 18 681 241 449 .8605611 | 19. | 21jan2006 19 681 254 475 .8791357 | 20. | 23apr2006 20 681 265 497 .8933169 | |------------------------------------------------------------| 21. | 23jul2006 21 681 275 517 .9044613 | +------------------------------------------------------------+ {title:Examples of artpep command lines} {phang}{cmd:. artpep, pts(25 30 40) edf0(0.5, 2) hr(1, 0.7)} {phang}{cmd:. artpep, pts(25 30 40) datestart(15jan2008) epts(50) eperiods(3) edf0(0.5, 2) hr(1, 0.7)} {phang}{cmd:. artpep, pts(5 7 12 18 25 30 40) eperiods(8) stoprecruit(12) edf0(0.5, 2) hr(1, 0.7)} {phang}{cmd:. artpep using myres, pts(5 7 12 18 25 30 40) datestart(01nov1995) epts(45) eperiods(8) startperiod(3) stoprecruit(12) edf0(0.9 0.7 0.1, 1 2 9) hr(1, 0.7) replace} {phang}{cmd:. use myres, replace} {phang}{cmd:. line patients events date, sort} {title:Authors} {pstd}Abdel Babiker, MRC Clinical Trials Unit at UCL{break} {browse "mailto:a.babiker@ucl.ac.uk":Ab Babiker} {pstd}Friederike Maria-Sophie Barthel, formerly MRC Clinical Trials Unit{break} {browse "mailto:sophie@fm-sbarthel.de":Sophie Barthel} {pstd}Babak Choodari-Oskooei, MRC Clinical Trials Unit at UCL{break} {browse "mailto:b.choodari-oskooei@ucl.ac.uk":Babak Oskooei} {pstd}Patrick Royston, MRC Clinical Trials Unit at UCL{break} {browse "mailto:j.royston@ucl.ac.uk":Patrick Royston} {title:Reference} {phang} Royston, P., and F. M.-S. Barthel. 2010. Projection of power and events in clinical trials with a time-to-event outcome. {it:Stata Journal} 10: 386-394. {title:Also see} {psee} Online: help for {help artpepdlg}, {help artsurv}