Forest Plots in Survival Subgroup Analyses Agenda CRO getting - - PowerPoint PPT Presentation

▶
forest plots in survival subgroup analyses
SMART_READER_LITE
LIVE PREVIEW

Forest Plots in Survival Subgroup Analyses Agenda CRO getting - - PowerPoint PPT Presentation

Dec 13, 2022 378 likes •952 views

Your statistical health consultancy CHM CRO TEMP WORK staying healthy getting healthy team health Forest Plots in Survival Subgroup Analyses Agenda CRO getting healthy Motivation ADTTE dummy data Macro for forest plots Summary

slide-1
SLIDE 1

Your statistical health consultancy

CHM staying healthy CRO getting healthy TEMP WORK team health

Forest Plots in Survival Subgroup Analyses

slide-2
SLIDE 2

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses 1/12

Motivation ADTTE dummy data Macro for forest plots Summary

Agenda

slide-3
SLIDE 3

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Survival subgroup analysis, e.g. study from oncology

Motivation

2/12

slide-4
SLIDE 4

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Survival subgroup analysis, e.g. study from oncology
  • Effect of study drug vs. placebo:

Hazard ratio (HR) and confidence interval (CI)

Motivation

2/12

slide-5
SLIDE 5

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Survival subgroup analysis, e.g. study from oncology
  • Effect of study drug vs. placebo:

Hazard ratio (HR) and confidence interval (CI)

  • Comparing HRs and CIs among various subgroups

Motivation

2/12

slide-6
SLIDE 6

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Survival subgroup analysis, e.g. study from oncology
  • Effect of study drug vs. placebo:

Hazard ratio (HR) and confidence interval (CI)

  • Comparing HRs and CIs among various subgroups

à Forest plot

Motivation

2/12

slide-7
SLIDE 7

CRO – getting healthy

Simple example:

http://genometoolbox.blogspot.com/2014/06/easy-forest-plots-in-r.html

3/12

Motivation

slide-8
SLIDE 8

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

Motivation

4/12

slide-9
SLIDE 9

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

  • Different confidence levels (e.g. 95% à 80%)

Motivation

4/12

slide-10
SLIDE 10

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

  • Different confidence levels (e.g. 95% à 80%)
  • Stratification (à different model)

Motivation

4/12

slide-11
SLIDE 11

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

  • Different confidence levels (e.g. 95% à 80%)
  • Stratification (à different model)
  • In our case: about 40 different forest plots

Motivation

4/12

slide-12
SLIDE 12

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

  • Different confidence levels (e.g. 95% à 80%)
  • Stratification (à different model)
  • In our case: about 40 different forest plots

à Macro needed!

Motivation

4/12

slide-13
SLIDE 13

CRO – getting healthy

Forest Plots in Survival Subgroup Analyses

  • Many endpoints in the study (Death, Progression free

survival, ...)

  • Different confidence levels (e.g. 95% à 80%)
  • Stratification (à different model)
  • In our case: about 40 different forest plots

à Macro needed!

  • Macro: ADTTE data as input à Forest plot as output

Motivation

4/12

slide-14
SLIDE 14

CRO – getting healthy

ADTTE dummy data

Ob s Subject identifier Gender Age group BMI at baseline Treatment Parameter code Censor variable Analysis value Analysis value unit 1 123450000 1 Female >= 85 < 25 Study drug PFS 301 Days 2 123450000 1 Female >= 85 < 25 Study drug DEATH 1 315 Days 3 123450004 2 Male 65-74 25-30 Placebo PFS 1 1 Days 4 123450004 2 Male 65-74 25-30 Placebo DEATH 1 1 Days 5 123450011 4 Male < 65 > 30 Study drug PFS 51 Days 6 123450011 4 Male < 65 > 30 Study drug DEATH 94 Days 7 123450021 Female 75-84 . Placebo PFS 23 Days 8 123450021 Female 75-84 . Placebo DEATH 1 23 Days

5/12

slide-15
SLIDE 15

CRO – getting healthyMacro for forest plots

%macro_forest_plot()

Forest Plots in Survival Subgroup Analyses

slide-16
SLIDE 16

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

7/12

slide-17
SLIDE 17

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

7/12

slide-18
SLIDE 18

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

7/12

slide-19
SLIDE 19

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times

7/12

slide-20
SLIDE 20

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

7/12

slide-21
SLIDE 21

CRO – getting healthy

Requirements & layout:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

7/12

slide-22
SLIDE 22

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

Options

8/12

slide-23
SLIDE 23

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

PROC PHREG Options

8/12

slide-24
SLIDE 24

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options

8/12

slide-25
SLIDE 25

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options Standard footnotes

8/12

slide-26
SLIDE 26

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options Standard footnotes

8/12

slide-27
SLIDE 27

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Main input:

8/12

slide-28
SLIDE 28

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Main input:

8/12

slide-29
SLIDE 29

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Survival data Main input:

8/12

slide-30
SLIDE 30

CRO – getting healthy

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC SGPLOT PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Main input: Survival data

8/12

slide-31
SLIDE 31

CRO – getting healthy

6/9

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC SGPLOT PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Forest plot with title & table Main input: Survival data

slide-32
SLIDE 32

CRO – getting healthy

6/9

Requirements:

  • CI level & endpoint

apparent in the title

  • HRs & CIs for each

subgroup category + Subgroup size

  • Survival data: Events

& censored patients

  • Median survival times
  • Footnotes

& PROC SGPLOT PROC LIFETEST PROC PHREG Options ADTT E data Standard footnotes Forest plot with title, table & footnotes Main input: Survival data

slide-33
SLIDE 33

CRO – getting healthy

%macro_forest_plot( endpoint = pfs , ci_level = 95 , sgrp_vlst = sex agegr01n , sgrp_lbls = Gender~Age Group , footnote1 = Up to five additional footnotes (footnote1- footnote5) possible. );

Macro call – example 1:

9/12

slide-34
SLIDE 34

CRO – getting healthy

%macro_forest_plot( endpoint = pfs , ci_level = 95 , sgrp_vlst = sex agegr01n , sgrp_lbls = Gender~Age Group , footnote1 = Up to five additional footnotes (footnote1- footnote5) possible. );

Macro call – example 1:

9/12

slide-35
SLIDE 35

CRO – getting healthy

%macro_forest_plot( endpoint = pfs , ci_level = 95 , sgrp_vlst = sex agegr01n , sgrp_lbls = Gender~Age Group , footnote1 = Up to five additional footnotes (footnote1- footnote5) possible. );

Macro call – example 1:

9/12

slide-36
SLIDE 36

CRO – getting healthy

%macro_forest_plot( endpoint = pfs , ci_level = 95 , sgrp_vlst = sex agegr01n , sgrp_lbls = Gender~Age Group , footnote1 = Up to five additional footnotes (footnote1- footnote5) possible. ); %macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Changing to: Macro call – example 1:

9/12

slide-37
SLIDE 37

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

9/12

slide-38
SLIDE 38

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

9/12

slide-39
SLIDE 39

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

9/12

slide-40
SLIDE 40

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

9/12

slide-41
SLIDE 41

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

9/12

slide-42
SLIDE 42

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

10/12

slide-43
SLIDE 43

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2: Further possibilities – additional macro vars:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

10/12

slide-44
SLIDE 44

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

Further possibilities – additional macro vars:

10/12

slide-45
SLIDE 45

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

Further possibilities – additional macro vars:

10/12

slide-46
SLIDE 46

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

Further possibilities – additional macro vars:

10/12

slide-47
SLIDE 47

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

Further possibilities – additional macro vars:

10/12

slide-48
SLIDE 48

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI );

Macro call – example 2:

subj_lst_ny = N strat_facts = sex agegr01n strat_lbls = Gender and Age time_unit = weeks

  • verall_ny

= Y

Further possibilities – additional macro vars:

10/12

slide-49
SLIDE 49

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI , time_unit = months , evnts_cnsr_ny = N , surv_rates_ny = Y , time_points = 3 6 );

Macro call – example 3:

11/12

slide-50
SLIDE 50

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI , time_unit = months , evnts_cnsr_ny = N , surv_rates_ny = Y , time_points = 3 6 );

Macro call – example 3:

11/12

slide-51
SLIDE 51

CRO – getting healthy

%macro_forest_plot( endpoint = death , ci_level = 80 , sgrp_vlst = sex agegr01n bmigr1 , sgrp_lbls = Gender~Age Group~BMI , time_unit = months , evnts_cnsr_ny = N , surv_rates_ny = Y , time_points = 3 6 );

Macro call – example 3:

11/12

slide-52
SLIDE 52

CRO – getting healthy

Summary

Ø Macro that generates forest plots for a survival subgroup analysis

Forest Plots in Survival Subgroup Analyses 12/12

slide-53
SLIDE 53

CRO – getting healthy

Summary

Ø Macro that generates forest plots for a survival subgroup analysis Ø Easy handling of

  • Many different endpoints, confidence levels, etc.
  • Stratification
  • Table with additional survival data

Forest Plots in Survival Subgroup Analyses 12/12

slide-54
SLIDE 54

CRO – getting healthy

Summary

Ø Macro that generates forest plots for a survival subgroup analysis Ø Easy handling of

  • Many different endpoints, confidence levels, etc.
  • Stratification
  • Table with additional survival data

Ø Automatically generated titles, footnotes, ...(“Dynamical output generation“) à Easy to maintain

Forest Plots in Survival Subgroup Analyses 12/12

slide-55
SLIDE 55

CRO – getting healthy

Summary

Ø Macro that generates forest plots for a survival subgroup analysis Ø Easy handling of

  • Many different endpoints, confidence levels, etc.
  • Stratification
  • Table with additional survival data

Ø Automatically generated titles, footnotes, ...(“Dynamical output generation“) à Easy to maintain Ø If you have to create an extensive amount of similar plots or tables, write a macro! It simplifies your work.

12/12

slide-56
SLIDE 56

CRO – getting healthy

Thank you for your attention! Questions?

Forest Plots in Survival Subgroup Analyses