topic Re: Plotting survival sample size calculation in Statistical Procedures

Plotting survival sample size calculation

UcheOkoro — Thu, 12 May 2022 18:13:05 GMT

Hello,

Please, I need help plotting this survival sample size calculation. The problem is with the plot x=effect. Please, what is the right thing?

Below is my code.

Thank you

ods listing style=htmlbluecml;
ods graphics on;
   proc power;
   twosamplesurvival test=logrank
      curve("Standard") = 90 : 0.7
      curve("Proposed") = (10 to 90 by 10):(0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.60 0.55)
      groupsurvival = "Standard" | "Proposed"
      accrualtime = 0
      followuptime = 90
       power = 0.8
      npergroup = .;
   plot x=effect min=0 max=90;run;

Re: Plotting survival sample size calculation

ballardw — Thu, 12 May 2022 21:24:57 GMT

That is not a "problem". That is an ERROR.

Running your code:

115     proc power;
116     twosamplesurvival test=logrank
117        curve("Standard") = 90 : 0.7
118        curve("Proposed") = (10 to 90 by 10):(0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.60 0.55)
119        groupsurvival = "Standard" | "Proposed"
120        accrualtime = 0
121        followuptime = 90
122         power = 0.8
123        npergroup = .;
124     plot x=effect min=0 max=90;
125  run;

ERROR: X=EFFECT is invalid for this analysis.

With Twosamplesurvival the effect supported is Hazard ratio if used. Which means a Hazardratio paremeter or list is required. The Hazardratio is incompatible with Groupsurvival. So to get plots you need to rephrase your approach. I don't know enough about the survival stuff to make a viable suggestion.

Re: Plotting survival sample size calculation

UcheOkoro — Mon, 16 May 2022 13:50:49 GMT

ods listing style=htmlbluecml;
ods graphics on;
proc power;
twosamplesurvival test=logrank
curve("Standard") = 90 : 0.7
curve("Proposed") = (10 to 90 by 10):(0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.60 0.55)
groupsurvival = "Standard" | "Proposed"
accrualtime = 0
followuptime = 90
power = 0.8
npergroup = .;
plot x=power min=0 max=1;run;