Richardson, Patrick
2008-Dec-04 13:49 UTC
[R] Comparing survival curves with "survdiff" "strata" help
ExpeRts, I'm trying to compare three survival curves using the function "survdiff" in the survival package. Following is my code and corresponding error message.> survdiff(Surv(st_months, status) ~ strata(BOR), data=mydata)Error in survdiff(Surv(st_months, status) ~ strata(BOR), data = mydata) : No groups to test When I check the "strata" of the variable. I get . . .> strata(BOR)Levels: BOR=PD BOR=PR BOR=SD So I know there are levels associated with the BOR variable, but I can't seem to get the strata statement to work within "survdiff". I checked ?strata and (unless I am not understanding) the help, it seems to be working as I get the following output> levels(strata(BOR))[1] "BOR=PD" "BOR=PR" "BOR=SD" I can get the survival tables and KM plot with no problem, but I can't seem to be able to compare the survival curves. I can get an overall comparison if I remove the "strata" statement> survdiff(Surv(st_months, status) ~ BOR, data=mydata)Call: survdiff(formula = Surv(st_months, status) ~ BOR, data = mydata) n=35, 5 observations deleted due to missingness. N Observed Expected (O-E)^2/E (O-E)^2/V BOR=PD 13 13 6.51 6.464 8.87 BOR=PR 10 10 18.48 3.888 10.59 BOR=SD 12 12 10.01 0.395 0.61 Chisq= 13.5 on 2 degrees of freedom, p= 0.00117 But when I use survdiff(Surv(st_months, status) ~ strata(BOR), data=mydata) it doesn't want to work. Any suggestions? Many thanks for your assistance! _____________________________________________ Patrick Richardson Biostatistician - Program of Translational Medicine Van Andel Research Institute - Webb Lab 333 Bostwick Avenue NE Grand Rapids, MI 49503 This email message, including any attachments, is for th...{{dropped:9}}
Richardson, Patrick
2008-Dec-05 13:00 UTC
[R] Comparing survival curves with "survdiff" "strata" help
ExpeRts, I'm trying to compare three survival curves using the function "survdiff" in the survival package. Following is my code and corresponding error message.> survdiff(Surv(st_months, status) ~ strata(BOR), data=mydata)Error in survdiff(Surv(st_months, status) ~ strata(BOR), data = mydata) : No groups to test When I check the "strata" of the variable. I get . . .> strata(BOR)Levels: BOR=PD BOR=PR BOR=SD So I know there are levels associated with the BOR variable, but I can't seem to get the strata statement to work within "survdiff". I checked ?strata and (unless I am not understanding) the help, it seems to be working as I get the following output> levels(strata(BOR))[1] "BOR=PD" "BOR=PR" "BOR=SD" I can get the survival tables and KM plot with no problem, but I can't seem to be able to compare the survival curves. I can get an overall comparison if I remove the "strata" statement> survdiff(Surv(st_months, status) ~ BOR, data=mydata)Call: survdiff(formula = Surv(st_months, status) ~ BOR, data = mydata) n=35, 5 observations deleted due to missingness. N Observed Expected (O-E)^2/E (O-E)^2/V BOR=PD 13 13 6.51 6.464 8.87 BOR=PR 10 10 18.48 3.888 10.59 BOR=SD 12 12 10.01 0.395 0.61 Chisq= 13.5 on 2 degrees of freedom, p= 0.00117 But when I use survdiff(Surv(st_months, status) ~ strata(BOR), data=mydata) it doesn't want to work. Any suggestions? Many thanks for your assistance! _____________________________________________ Patrick Richardson Biostatistician - Program of Translational Medicine Van Andel Research Institute - Webb Lab 333 Bostwick Avenue NE Grand Rapids, MI 49503 This email message, including any attachments, is for th...{{dropped:9}}