Professional Documents
Culture Documents
Further Issues For Cox Regression (As07) : Course: PG Diploma/ MSC Epidemiology
Further Issues For Cox Regression (As07) : Course: PG Diploma/ MSC Epidemiology
(AS07)
EPM304 Advanced Statistical Methods in Epidemiology
This document contains a copy of the study material located within the computer
assisted learning (CAL) session.
If you have any questions regarding this document or your course, please contact
DLsupport via DLsupport@lshtm.ac.uk.
Important note: this document does not replace the CAL material found on your
module CDROM. When studying this session, please ensure you work through the
CDROM material first. This document can then be used for revision purposes to
refer back to specific sessions.
These study materials have been prepared by the London School of Hygiene & Tropical Medicine as part of
the PG Diploma/MSc Epidemiology distance learning course. This material is not licensed either for resale
or further copying.
London School of Hygiene & Tropical Medicine September 2013 v2.0
To review Cox regression, and learn how to check whether the effect of an
exposure is constant over time, the proportional hazards assumption.
Objectives
By the end of this session you will be able to:
SM03
To illustrate methods in this session the example below is used. Click on the example
to see further details below.
Study of risk factors for mortality
among males in Trinidad
Interaction: Hyperlink: Study of risk factors for mortalityamong males in
Trinidad (card appears on RHS):
All males aged 3574 years who were living in two neighbouring suburbs of Port of
Spain, Trinidad, in March 1977 were eligible and entered into the study. Baseline
data were recorded for 1,343 men on a range of risk factors including ethnic group,
cigarette and alcohol consumption.
All subjects were then visited annually at home, and morbidity and mortality records
were compiled. Regular inspection of hospital records, death registers and obituaries
were also used to update the records. Those who had moved away (or abroad) were
contacted annually by postal questionnaire and were also seen if they returned to
Port of Spain. Follow-up of the study cohort finished at the end of 1987, giving a
study period of ten years.
event. This produces a constant rate within each very small interval but also allows
the rate to vary over longer time intervals.
So, for rapidly changing rates Cox is the appropriate model.
z
2.023
0.535
P>
|z|
0.043
0.593
95% confidence
interval
1.034366
8.362418
0.4948687
3.426243
2.9410
54
1.3021
29
Standa
rd
Error
1.5680
6
0.6427
479
P > |z|
95% confidence
interval
2.023
0.043
0.535
0.593
1.0343
66
0.4948
687
8.3624
18
3.4262
43
Interaction: Hotspot: The hazard ratio for the effect of smoking does not change
over time:
Interaction: Hotspot: The baseline rate for the non-smoking group is constant:
Incorrect Response (pop up box appears):
Interaction: Hotspot: The rate of deaths, adjusted for 'CHD at entry', in men who
smoke, is the same as the rate of deaths in men who do not smoke, i.e. the hazard
ratio = 1.:
Correct Response (pop up box appears):
(t,i)
= (constant over time)
(t,0)
This means that the variation over time for any individual is the same as the
variation in the baseline.
Forget about Cox regression to begin with and think back to a classical method for
survival analysis, namely Kaplan-Meier estimates and survival curves.
This is a non-parametric method and therefore makes few assumptions about your
data.
To illustrate how you can assess whether the proportional hazards assumption is
valid you will use the Trinidad study of mortality risk factors. The focus is on whether
there is a change in the effect of an exposure over time.
Total
252
Death
s
7
Lost
245
240
234
10
223
10
212
18
190
23
165
11
21
133
73
10
56
55
Survi
val
0.972
2
0.952
4
0.932
5
0.892
6
0.852
4
0.835
7
0.826
3
0.767
5
0.735
6
0.709
8
Std
Error
0.010
4
0.013
4
0.015
8
0.019
5
0.022
4
0.023
5
0.024
1
0.028
2
0.031
2
0.039
4
[95% Conf.
Int.]
0.942 0.986
6
7
0.917 0.972
7
7
0.893 0.957
7
5
0.847 0.925
2
0
0.802 0.890
2
8
0.783 0.876
5
2
0.772 0.868
9
2
0.706 0.817
5
4
0.668 0.791
7
2
0.624 0.779
6
1
Std
Error
0.043
7
0.054
8
0.062
9
0.071
4
[95% Conf.
Int.]
0.774 0.973
9
8
0.712 0.943
3
0
0.652 0.907
1
6
0.566 0.848
1
8
Total
38
Death
s
3
Lost
35
33
31
Survi
val
0.921
1
0.868
4
0.815
8
0.736
8
28
26
23
18
12
10
0.684
2
0.684
2
0.620
6
0.620
6
0.551
6
0.551
6
0.075
4
0.075
4
0.080
7
0.080
7
0.096
8
0.096
8
0.511
5
0.511
5
0.443
0
0.443
0
0.346
2
0.346
2
0.806
7
0.806
7
0.756
1
0.756
1
0.716
3
0.716
3
(t,i)
= (constant over time)
(t,0)
(t,i) = i (t,0)
Interaction: Button: More:
Then, summing both sides over time:
(t,i) = i (t,0)
Cumulative hazard ( i ) = i x cumul. hazard ( 0 )
The log cumulative incidence curves can be assumed parallel throughout the study
period.
You can therefore say that the effect of 'CHD at entry' is constant throughout the
follow-up time and the proportional hazards assumption is valid.
Correct
Yes, the assumption of a constant exposure effect throughout the follow-up period
means that the effect is constant on a log scale. Therefore, on a hazard scale, the
rates will be proportional hence the term proportional hazards.
Incorrect Response variable (pop up box appears):
The hazards are variable, but that is not what the assumption of constant exposure
effect means. The effect is the ratio between the rates (hazards), so what does a
constant effect imply? Please try again.
Incorrect Response constant (pop up box appears):
No, it is the effect that is constant, not the hazards. The effect is the ratio between
the rates (hazards), so what does a constant effect imply? Please try again.
Notice the three intervals of time are not equal because they are chosen to contain
the same number of events.
Number of deaths
27
33
28
88
Table
Plot
2 to < 5 years:
Table
Plot
5 or more years:
Table
Plot
Examine the hazard ratio estimates and their confidence intervals, do you think you
can assume proportional hazards for 'CHD at entry'?
Interaction: Hyperlink: Table( 0 to < 2 years:):
Time interval: 0 to < 2 years
Variable: Time1
chdstart
Hazard
ratio
2.907241
Standard
error
1.547554
P < |z|
2.005
0.045
95% confidence
interval
1.024183
8.252478
chdstart
Hazard
ratio
2.158332
Standard
error
0.923029
P < |z|
1.799
0.072
95% confidence
interval
0.933458
4.990475
chdstart
Hazard
ratio
1.360623
Standard
error
0.838447
z
0.500
P < |z|
0.617
95% confidence
interval
0.406638
4.552683
The likelihood ratio statistic (LRS) = 0.91, P = 0.6. Therefore, there is no evidence
for non-proportional hazards for 'CHD at entry'.
Cox model with time interval interaction for 'CHD at entry'
chdstart
time2.chd
time3.chd
Hazard
ratio
2.907241
0.742399
0.468012
Standard
error
1.547554
0.506926
0.381102
P < |z|
2.005
0.436
0.932
0.045
0.663
0.351
95% confidence
interval
1.024183
8.252478
0.194723
2.830457
0.094869
2.308821
Table
2 to < 5 years:
Table
5 or more years:
Plot
Plot
Table
Plot
Examine the estimates and the corresponding plots. Do you think you can assume
proportional hazards for a model that includes exposure to smoking?
Interaction: Hyperlink: Table (0 to < 2 years):
Time interval: 0 to < 2 years
Variable: Time1
smoking
Hazard
ratio
0.938849
Standard
error
0.383288
z
0.155
P < |z|
0.877
95% confidence
interval
0.421784
2.089784
smoking
Hazard
ratio
1.840149
Standard
error
0.640995
P < |z|
1.751
0.080
95% confidence
interval
0.929708 3.642163
2.21650
5
Standar
d
error
0.840160
z
2.100
P < |z|
0.036
95% confidence
interval
1.054443
4.659229
smoking
smok.tim
e2
smok.tim
e3
Hazard
ratio
0.938849
1.960006
Standard
error
0.383288
1.051869
0.155
1.254
0.877
0.210
95% confidence
interval
0.421784
2.089784
0.684615
5.611364
2.360875
1.315216
1.542
1.542
0.792280
P < |z|
7.035057
Section 8: Summary
This is the end of AS07. When you are happy with the material covered here please
move on to session AS08 .
The main points of this session will appear below as you click on the relevant title.
Use of Cox regression
Cox regression is used to model rapidly changing rates.
In survival analysis, a rate is more commonly referred to as a hazard and the effect
of exposure is a hazard ratio.
The hazard is assumed constant within a timeclick, but is allowed to change over
time.
Proportional hazards assumption
The key assumption for Cox regression is that the effect of an exposure is constant
over time. In other words, the variation in an exposed group is the same as the
variation in the baseline group, resulting in the effect of exposure being constant.
This is called the proportional hazards assumption.
Nelson-Aalen plots