Longitudinal data

Report
Statistical Analysis of
Longitudinal Data
Ziad Taib
Biostatistics, AZ
April 2011
Name, department
1
Date
Outline of lecture 1
1. An introduction
2. Two examples
3. Principles of Inference
4. Modelling continuous longitudinal data
Name, department
2
Date
Part 1: An introduction
Name, department
3
Date
Why longitudinal data?
 Very useful for their own sake.
 With longitudinal data, we have the possibility of
understanding what mixed models are about in a relatively
simple but yet rich enough context.
___________________________________
A good reference is the book ”Designing experiments and
analyzing data” by Maxwel l& Delaney (2004)
Name, department
4
Date
Longitudinal Data
Repeated measures are obtained when a response
is measured repeatedly on a set of units
• Units:
• Subjects, patients, participants, . . .
• indivduals, plants, . . .
• Clusters: nests, families, towns, . .
•...
• Special case: Longitudinal data
Obs! Possible to handle several levels
Name, department
5
Date
A
A motivating example
B
Baseline
3 months 6 months
 Consider a randomized clinical trial with two treatment groups and
repeated measurements at baseline, 3 and 6 months later. As it
turned out some of the data was missing. Moreover patients did not
always comply with time requirements. Our first reaction is to try to
compensate for the missing values by some kind of imputation, or to
use list-wise deletion.
 Both ”methods” having their shortcomings, wouldn't it be nice to be
able to use something else? There is in fact an alternative method:
using the idea of mixed models.
 With mixed models,
1. we can use all our data having the attitude that ”what is missing is
missing”.
2. we can even account for the dependencies resulting from measurements
made on the same individuals at different times.
3. we don’t need to be consistent about time.
Name, department
6
Date
Mixed effects models
 Ordinary fixed effects linear model usually assume:
1) independence with the same variance.
2) normally distributed errors.
3) constant parameters
 Y1   1 x1 
1 
  
  0   
...  ... ...    ...
  
 
 
 1
Y n  1 x n 
  n 
Y  X    ,  is N ( 0 ,  I ),   constant.
2
 If we modify assumptions 1) and 3), then the problem becomes more
complicated and in general we need a large number of parameters
only to describe the covariance structure of the observations. Mixed
effects models deal with this type of problems.
 In general, this type of models allows us to tackle such problems as:
clustered data, repeated measures, hierarchical data.
Name, department
7
Date
Various forms of models and relation between them
Classical statistics (Observations are random, parameters are unknown constants)
LM: Assumptions:
1.
independence,
2.
normality,
3.
constant parameters
LMM:
Assumptions 1)
and 3) are modified
GLM: assumption 2)
Exponential family
Repeated measures:
Assumptions 1) and 3)
are modified
GLMM: Assumption 2) Exponential
family and assumptions 1) and 3) are
modified
Longitudinal data
Maximum likelihood
LM - Linear model
Non-linear models
GLM - Generalised linear model
LMM - Linear mixed model
GLMM - Generalised linear mixed model
Name, department
8
Date
Bayesian statistics
Part 2: Two examples
 Rat data
 Prostate data
Name, department
9
Date
Example 1: Rat Data (Verbecke et al)
Research question How does craniofacial growth in the
wistar rat depend on testosteron production?
Name, department
10 Date
Name, department
11 Date
 •Randomized experiment in which 50 male Wistar rats
are randomized to:
Prevents the production of testesterone
 Control (15 rats)
 Low dose of Decapeptyl (18 rats)
 High dose of Decapeptyl (17 rats)
 Treatment starts at the age of 45 days.
 Measurements taken every 10 days, from day 50 on.
 The responses are distances (pixels) between two well
defined points on x-ray pictures of the skull of each rat.
Here, we consider only one response, reflecting the
height of the skull.
Days
45
Name, department
12 Date
50
60
70
80
Individual profiles:
Name, department
13 Date
1.
2.
3.
Connected profiles better that scatter plots
Growth is expected but is it linear
Of interest change over time (i.e. Relationship between response and age)
Complication: Many dropouts due to anaesthesia imply less power but no bias.
Without dropouts easier problem because of balance.
Name, department
14 Date
Remarks:
 Much variability between rats
 Much less variability within rats
 Fixed number of measurements scheduled per subject,
but not all measurements available due to dropout, for
known reason.
 Measurements taken at fixed time points
Research question: How does craniofacial growth in the wistar
rat depend on testosteron production ?
Name, department
15 Date
Example 2: The BLSA Prostate Data
Name, department
16 Date
Example 2: The BLSA Prostate Data
(Pearson et al., Statistics in Medicine,1994).
 Prostate disease is one of the most common and most
costly medical problems in the world. Important to look for
biomarkers which can detect the disease at an early stage.
 Prostate-Specific Antigen is an enzyme produced by both
normal and cancerous prostate cells. It is believed that
PSA level is related to the volume of prostate tissue.
 Problem: Patients with Benign Prostatic Hyperplasia also
have an increased PSA level
 Overlap in PSA distribution for cancer and BPH cases
seriously complicates the detection of prostate cancer.
Name, department
17 Date
 Research question: Can longitudinal PSA profiles be used
to detect prostate cancer in an early stage ?
 A retrospective case-control study based on frozen serum
samples:




16 control patients
20 BPH cases
14 local cancer cases
4 metastatic cancer cases
Name, department
18 Date
Individual profiles:
Name, department
19 Date
Remarks:
 Much variability between subjects
 Little variability within subjects
 Highly unbalanced data
Research question: Can longitudinal PSA profiles be used to
detect prostate cancer in an early stage ?
Name, department
20 Date
Part 3: Principles of Inference
Name, department
21 Date
Fisher´s likelihood Inference for
observable y and fixed parameter q
 Data Generation : Given a stochastic model
Generate data, y, from
fq ( y )
,
fq ( y )
 Parameter Estimation : Given the data y, make inference
about q by using the likelihood L ( y / q )
q
 Connection between two processes :
Lq ( y / q )  f q ( y )
Name, department
22 Date
(Classical) Likelihood Principle
 Birnbaum (1962) All the evidence or information about the
parameters in the data is in the likelihood.
Conditionality principle
& Sufficiency principle
Name, department
23 Date
Likelihood principle
Bayesian Inference for observable y and
unobservable n
 Data Generation : Generate data according to
1. n, from
f (n )
prior
2. For n fixed generate y from
f ( y /n )
 Combine into f (n ) f ( y / n )
 Parameter Estimation : Given the data y, make
inference about n by using f (n / y )
posterior
 The connection between two processes:
f (n ) f ( y / n )  f ( y ) f (n / y )
f (n / y ) 
f ( y ,n )
f ( y)
Name, department
24 Date
Compare with Lq ( y / q )
 f ( y ) f (n / y )  f ( y ,n )  f (n ) f ( y / n )
Extended likelihood inference: (Lee and
Nelder) for observable y, fixed
parameter q and unobservable n
Name, department
25 Date
Parameter estimation
Name, department
26 Date
L ( y / q )  fq ( y )
Extended Likelihood Principle
 Björnstad (1996) All information in the data about the
unobservables and the parameters is in the “likelihood”.
Conditionality principle
& Sufficiency principle
Name, department
27 Date
Likelihood principle
Prediction: predict the number of
seizures during the next week
Name, department
28 Date
Name, department
29 Date
Bayesian Predictive Inference
 Given n, the observations y are assumed to be
independent. How do we predict the next value, Y, of the
observable? In a Bayesian setting we may determine the
posterior f (n / y ) and define the predictive density of Y
given y as: f Y ( x / y )
Jefreys’ Priors
Obs!
Name, department
30 Date
Bayesian inference (Pearson, 1920)
Name, department
31 Date
Name, department
32 Date
Nelder and Lee (1996)
?
Name, department
33 Date
Name, department
34 Date
Part 4: A Model for Longitudinal Data
Name, department
35 Date
Introduction
 In practice: often unbalanced data due to
 (i) unequal number of measurements per subject
 (ii) measurements not taken at fixed time points. Therefore,
ordinary multivariate regression techniques are often not
applicable.
 Often, subject-specific longitudinal profiles can be well
approximated by linear regression functions. This leads to
a 2-stage model formulation:
 Stage 1: A linear (e.g. regression) model for each subject
separately
 Stage 2: Explain variability in the subject-specific (regression)
coefficients using known covariates
Name, department
36 Date
A 2-stage Model Formulation:
Stage 1
 Response Yij for ith subject, measured at time tij, i = 1, . . . , N, j = 1, . . .
, ni
 Response vector Yi for ith subject:
Y i  (Y i1 , Y i 2 ,..., Y in i )'
Possibly after some convenient transformation
Y i  Z i  i   i ,  i ~ N ( 0 ,  i ), often  i   In i
2
 Zi is a (ni x q) matrix of known covariates and
 i is a (ni x q) matrix of parameters
 Note that the above model describes the observed variability within
subjects
Name, department
37 Date
Stage 2
 Between-subject variability can now be studied from
relating the parameters i to known covariates
 i  K i   bi
 Ki is a (q x p) matrix of known covariates and
  is a (p-dimensional vector of unknown regression
parameters
 Finally
Name, department
38 Date
bi ~ N ( 0 ,  i )
The General Linear Mixed-effects
Model
 The 2-stages of the 2-stage approach can now be
combined into one model:
Average evolution
Name, department
39 Date
Subject specific
The general mixed effects models can be summarized by:
Convenient using multivariate normal.
Very difficult with other distributions
Terminology:
• Fixed effects: 
• Random effects: bi
• Variance components: elements in D and i
Name, department
40 Date
Remarks
1. It is occasionally unclear if we should treat an effect as
a fixed or a mixed effect. For example in clinical trials with ?
treatment and clinic as “factors” should we consider
clinics as random?
2. Considering the general form of a mixed effects model
Yi  X i   Z i bi   i
notice that the fixed effects are involved only in mean
values (just like in ordinary linear models) while random
effects modify the covariance matrix of the observations.
Name, department
41 Date
Example: The Rat Data
Name, department
42 Date
 Transformation of the time scale to linearize the profiles:
Age ij  t ij  ln[ 1 
( Age ij  45 )
]
10
 Note that t = 0 corresponds to the start of the treatment
(moment of randomization)
 • Stage 1 model:
Y ij   1i   2 i t ij   ij , j  1,... , n i
Name, department
43 Date
Stage 1
  1i 
i  


 2i 
Name, department
44 Date
Stage 2 model:
 In the second stage, the subject-specific intercepts and time effects are related to the
treatment of the rats
Name, department
45 Date
The hierarchical versus the marginal
Model
The general mixed model is given by
It can be written as
It is therefore also called a hierarchical model
Name, department
46 Date
Marginally we have that
is distributed as
Hence
Name, department
47 Date
f(yi I bi)
f(bi)
f(yi)
Example: The Rat Data
Can be negative or positive
reflecting individual deviation
from average
Name, department
48 Date
Linear model where each
rat has its own intercept
and its own slope
Comments:
• Linear average evolution in each group
• Equal average intercepts
• Different average slopes
Moreover, taking
Notice that the model assumes that the
variance function is quadratic over time.
Name, department
49 Date
Cov i ( Y ( t1 ), Y ( t 2 ))
  1i 
  1i 
 Cov ( 1, t1 
   i1 , 1, t 2 
   i2 )
  2i 
  2i 
  1i   1 
 1, t1  cov( 
 )    cov(  i1 ,  i1 )
  2 i  t 2 
 d 11
 1, t1 
 d 12
d 12   1 
    cov(  i1 ,  i1 )
d 22   t 2 
 d 11  t1 d 12 , d 12
1
 t1 d 22    cov(  i1 ,  i1 )
t 2 
 d 11  t1 d 12  t 2 d 12  t1t 2 d 22   cov(  i1 ,  i1 )
 d 11  ( t1  t 2 ) d 12  t1t 2 d 22   cov(  i1 ,  i1 )
Name, department
50 Date
Name, department
51 Date
Name, department
52 Date
The prostate data
A model for the prostate cancer
Stage 1
Y ij
 ln( PSA ij  1)
  1 i   2 i t ij   t   ij , j  1,... , n i
2
3 i ij
Name, department
53 Date
The prostate data
A model for the prostate cancer
Stage 2
Age could not be matched
  1i    1 Age i   2 C i   3 B i   4 L i   5 M i  b1 j 


 
  2 i     6 Age i   7 C i   8 B i   9 L i   10 M i  b 2 j 
     Age   C   B   L   M  b 
i
12
i
13 i
14 i
15
i
3j
 3 i   11
Ci, Bi, Li, Mi are indicators of the classes: control, BPH, local or
metastatic cancer. Agei is the subject’s age at diagnosis. The
parameters in the first row are the average intercepts for the different
classes.
Name, department
54 Date
The prostate data
This gives the following model
ij
Name, department
55 Date
Stochastic components in general linear
mixed model
Response
Subject 1
Average evolution
Subject 2
Time
Name, department
56 Date
References
 Aerts, M., Geys, H., Molenberghs, G., and Ryan, L.M.(2002). Topics in
Modelling of Clustered Data. London: Chapman and Hall.
 • Brown, H. and Prescott, R. (1999). Applied Mixed Models in
Medicine. New-York: John Wiley & Sons.
 • Crowder, M.J. and Hand, D.J. (1990). Analysis of Repeated
Measures. London: Chapman and Hall.
 • Davidian, M. and Giltinan, D.M. (1995). Nonlinear Models For
Repeated Measurement Data. London: Chapman and Hall.
 Davis, C.S. (2002). Statistical Methods for the Analysis of Repeated
Measurements. New York: Springer-Verlag.
 Diggle, P.J., Heagerty, P.J., Liang, K.Y. and Zeger, S.L. (2002).
Analysis of Longitudinal Data. (2nd edition). Oxford: Oxford University
Press.
Name, department
57 Date
References
 Fahrmeir, L. and Tutz, G. (2002). Multivariate Statistical Modelling
Based on Generalized Linear Models, (2nd edition). Springer Series in
Statistics. New-York: Springer-Verlag.
 Goldstein, H. (1979). The Design and Analysis of Longitudinal Studies.
London: Academic Press.
 Goldstein, H. (1995). Multilevel Statistical Models. London: Edward
Arnold.
 Hand, D.J. and Crowder, M.J. (1995). Practical Longitudinal Data
Analysis. London: Chapman and Hall.
 Jones, B. and Kenward, M.G. (1989). Design and Analysis of
Crossover Trials. London: Chapman and Hall.
 Kshirsagar, A.M. and Smith, W.B. (1995). Growth Curves. New-York:
Marcel Dekker.
 Lindsey, J.K. (1993). Models for Repeated Measurements. Oxford:
Oxford University Press.
Name, department
58 Date
Longford, N.T. (1993). Random Coefficient Models. Oxford: Oxford
University Press.
References
 Pinheiro, J.C. and Bates D.M. (2000). Mixed effects models in S and
S-Plus, Springer Series in Statistics and Computing. New-York:
Springer-Verlag.
 Searle, S.R., Casella, G., and McCulloch, C.E. (1992). Variance
Components. New-York: Wiley.
 Senn, S.J. (1993). Cross-over Trials in Clinical Research. Chichester:
Wiley.
 Verbeke, G. and Molenberghs, G. (1997). Linear Mixed Models In
Practice: A SAS Oriented Approach, Lecture Notes in Statistics 126.
New-York: Springer-Verlag.
 Verbeke, G. and Molenberghs, G. (2000). Linear Mixed Models for
Longitudinal Data. Springer Series in Statistics. New-York: SpringerVerlag.
 Vonesh, E.F. and Chinchilli, V.M. (1997). Linear and Non-linear Models
Name, department
59 Date
for the Analysis of Repeated Measurements. Marcel Dekker: Basel.
Any Questions
Name, department
60 Date
?

similar documents