01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014...

27
01/2014 1 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology & Community Medicine, University of Ottawa

description

01/20143 Cox methods do not require an equation for h(t) –do not produce direct estimates of h(t), H(t) or S(t). If we could estimate any of these, we could get the others: Introduction (1) Assuming h(t) is constant in interval

Transcript of 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014...

Page 1: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

101/2014

EPI 5344:Survival Analysis in

EpidemiologyEstimating S(t) from Cox models

April 1, 2014

Dr. N. Birkett,Department of Epidemiology & Community

Medicine,University of Ottawa

Page 2: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

201/2014

Objectives

• Theory of estimating S(t)• SAS methods

Page 3: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

301/2014

• Cox methods do not require an equation for h(t)– do not produce direct estimates of h(t), H(t) or S(t).

• If we could estimate any of these, we could get the others:

Introduction (1)

Assuming h(t) is constant in interval

Page 4: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

401/2014

• Since hazards are proportional, all we need to know is h(t) or S(t) for the baseline group: h0(t) or S0(t)

Introduction (2)

So, estimate S0(t)and we can get S(t)for any ‘x’s’

Page 5: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

501/2014

• Must use methods outside Cox regression.• Two common approaches are used:

– Generalize the Kaplan-Meier method to estimate S0(t)– Generalize Nelson-Aalen method to estimate H0(t)

• Implemented by using the BASELINE statement.

• First, a quick review of the technical background

Introduction (3)

Page 6: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

6

A bit of technical stuff

• We assume a piecewise constant hazard model– It keeps coming up, doesn’t it

01/2014

Page 7: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

701/2014

We don’t care about the origin of the next formulae.Assume there are no ties

Where:‘l’ is subject having event at tj

‘k’ is subject whose survival curve is needed

Page 8: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

801/2014

• Non-RCT study of therapy• Hypernephroma (type of kidney cancer)• All patients treated with

– chemotherapy and – Immunotherapy

• Some also had their affected kidney removed.• Questions

– Does having a nephrectomy affect survival? – Does age affect survival in patients having a

nephrectomy?

Example (1)

Page 9: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

901/2014

• Answer is YES

A2i = 1 if age is 60-70

= 0 otherwiseA3i = 1 if age is 70+

= 0 otherwiseNi = 1 if had nephrectomy

= 0 if no nephrectomy

• A3 & N are statistically significant.• What do survival curves look like?

Example (2)

Page 10: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1001/2014

Page 11: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1101/2014

Page 12: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1201/2014

Page 13: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1301/2014

Could try smoothing the h(t) curve

Estimated baseline hazard curve

Page 14: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

14

• How did we produce these curves?• Using the formulae given earlier.• SAS uses the BASELINE statement

01/2014

Example (3)

Page 15: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1501/2014

• SAS implements these methods in Phreg using the BASELINE statement

• By default, this statement produces S(t) at a pre-defined set of covariate values:– The reference level of any variable mentioned

in a class statement– The mean value of each other variable

SAS Methods (1)

Page 16: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

16

ODS GRAPHICS ON;ODS RTF;

PROC PHREG DATA=allison.recid PLOTS=S; MODEL week*arrest(0)=fin age prio / TIES=EFRON; BASELINE OUT=a SURVIVAL=s LOWER=lcl UPPER=ucl;RUN;

proc print data=a;run;

ODS RTF close;ODS GRAPHICS OFF;

01/2014

Page 17: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1701/2014

Page 18: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

1801/2014

Page 19: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

19

ODS GRAPHICS ON;ODS RTF;

PROC PHREG DATA=allison.recid PLOTS=S; Class fin; MODEL week*arrest(0)=fin age prio / TIES=EFRON; BASELINE OUT=a SURVIVAL=s LOWER=lcl UPPER=ucl;RUN;

proc print data=a;run;

ODS RTF close;ODS GRAPHICS OFF;

01/2014

Page 20: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

2001/2014

Page 21: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

21

• What if you want S(t) for some other set of covariates?– Interested in specific target group– Contrast extremes of the range of variables

• Use COVARIATES statement in SAS

01/2014

Page 22: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

22

DATA covals; INPUT fin age prio; DATALINES;0 40 3;run;

PROC PHREG DATA=allison.recid; MODEL week*arrest(0)=fin age prio / TIES=EFRON; BASELINE OUT=a COVARIATES=covals SURVIVAL=s LOWER=lcl UPPER=ucl;run;

proc print data=a;run;

01/2014

Page 23: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

2301/2014

Page 24: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

24

• What if you want to display the results of multiple covariate sets on the same graph?– Method #1

• Run several of previous models, for different covariates.

• Combine the output datasets into on dataset• Plot using SAS Graph, etc.

– Method #2• list more than one set of covariates in the covariate

data set

01/2014

Page 25: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

25

DATA covals; INPUT fin age prio; DATALINES;0 40 31 50 2;run;

ODS graphics ON;

PROC PHREG DATA=allison.recid PLOTS(OVERLAY)=S; MODEL week*arrest(0)=fin age prio / TIES=EFRON; BASELINE OUT=a COVARIATES=covals SURVIVAL=s LOWER=lcl UPPER=ucl;run;

ODS graphics off;

proc print data=a;run;

01/2014

Page 26: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

2601/2014

Page 27: 01/20141 EPI 5344: Survival Analysis in Epidemiology Estimating S(t) from Cox models April 1, 2014 Dr. N. Birkett, Department of Epidemiology  Community.

2701/2014