Statistics for everyone: ch.14 KM curve and log-rank test

easier R than SPSS with Rcmdr : Contents

ch.14 KM curve and log-rank test

Select ‘Kaplan-Meier survival curve and logrank test’.

Select ‘time’ and ‘event’ respectively (red square) and put the variable corresponding to the treatment, that is, a nominal variable to see the difference (green square). Set the rest of the options as shown. You can put 0, 1, or more norminal variables in the green rectangle part, but for now, you have only selected one.

The survival curve of the three groups is drawn. There is the Number at risk.

There are survival tables for each group and a 95% confidence interval of survival.

The actual number of events in each group is shown, and the total p value of the comparison of the three groups is calculated. This is the result of the logrank test.

It perform a post-hoc test for two groups each and show the value of p, which is adjusted by the Holm method. You will also see median survival and 95% CI for the three groups. This median survival and 95% CI as well as p values are indicated in the body of the paper.

Now let’s add sex to strata. At the bottom is options so that you can draw the strata separately.

The strata(sex is for now) were separated and painted. Sometimes you may need this. Strata is a subgroup analysis, in which the entire population is divided into several parts.

I drew it without separating it, and it’s quite complicated.

If the picture is complex, you can choose not to draw 95% CI, or you can change the position of the legend to ‘mouse click’.

Once the picture is drawn, resize it appropriately and click the mouse in the blank place to complete the legend.

easier R than SPSS with Rcmdr : Contents

=================================================