You can display the 95% confidence intervals for the survival rate (by time point) by selecting [Plot Options] > [Show Confidence Intervals] from the red triangle button ( ▼ ) in the upper left corner of the Kaplan-Meier fit report. The confidence interval values (lower and upper limits) displayed here can be calculated using the following method.
For JMP 9 and above
Click the red triangle button ( ▼ ) in the upper left of the Kaplan-Meier fit report and select Save Estimates. These will be the values in the columns "Lower Survival Rate" and "Upper Survival Rate" in the newly output data table.
For JMP 8 and below
Since it is not possible to directly output values like in JMP 9, the calculation is performed using the following operations:
- When you draw a Kaplan-Meier plot, if there is a column specified as a "combination" or group variable at the bottom of the report, a report corresponding to the group name of that column will be displayed. Right-click on any report and select "Export to Data Table."
*Note: If your JMP version is earlier than 8.0.0, please also do the following:
In the newly created table, change the name of the column "Survival Standard Deviation" to "Survival Standard Error."
- Next, with the newly created data active, select Columns > Create New Column.
- In the "Create New Column" window, enter "Lower 95%" in "Column Name" and select "Column Properties" -> "Formula".
- 計算式エディタウィンドウが表示されたら、右のパネルの ［計算式がありません］の青いボックスをダブルクリックして入力モードになったら、以下の計算式を貼り付けます。
Squish(Logit(:生存率) + (Normal Quantile(0.025) * :生存標準誤差) / (:生存率 * (1 - :生存率)))
- Close the Formula Editor window and close the "Create New Column" window.
- 3〜5 の手順を再度行います。列名は、「上側95％」にして 以下の計算式を貼り付けます。
Squish(Logit(:生存率) + (Normal Quantile(0.975) * :生存標準誤差) / (:生存率 * (1 - :生存率)))
References
Meeker, WQ and Escobar, LA (1998), Statistical Methods for Reliability Data,
New York: John Wiley & Sons
FAQ #3315
