1 Classification Schemes for “Ways of Getting the Wrong Answer” Szklo and Nieto –Bias...
If you can't read please download the document
1 Classification Schemes for “Ways of Getting the Wrong Answer” Szklo and Nieto –Bias »Selection Bias »Information/Measurement Bias –Confounding –Chance
1 Classification Schemes for Ways of Getting the Wrong Answer
Szklo and Nieto Bias Selection Bias Information/Measurement Bias
Confounding Chance Other Common Approach Bias Selection Bias
Information/Measurement Bias Confounding Bias Chance
Slide 2
2 Confounding and Interaction Confounding: one of the central
problems in observational clinical research What is it? What does
it do? What kind of variables act as confounders? Which variables
to consider as confounders? Which variables not to consider as
confounders?
Slide 3
3 Smoking, Matches, and Lung Cancer A tobacco company
researcher believes that exposure to matches is the cause of lung
cancer He conducts a large case-control study to test this
hypothesis
Slide 4
4 Smoking, Matches, and Lung Cancer Your colleague has located
1000 cases of lung cancer from a population-based registry, of whom
820 have a history of carrying matches. Among 1000 reference
(control) patients (selected randomly from the population and found
to have normal chest x-rays), 340 carry matches. Quantitate the
relationship between matches and lung cancer in your colleagues
data.
Slide 5
5 Matches and Lung cancer Exposure odds ratio = (820/180) /
(340/660) = disease odds ratio OR = 8.8 95% CI (7.2, 10.9)
Slide 6
6 Smoking, Matches, and Lung Cancer You decide to look at the
relationship between matches and lung cancer in the smokers
separately from the non-smokers. You find that among the 1000
cases, 900 are smokers and 810 (OF THE 900) carry matches. Among
the 1000 reference patients, 300 are smokers and 270 (OF THE 300)
carry matches. Draw the necessary stratified tables and calculate
the relevant measure of association
Slide 7
7 Smoking, Matches, and Lung Cancer Stratified Crude
Non-SmokersSmokers OR crude OR CF+ = OR smokers OR CF- = OR non -
smokers OR crude = 8.8 (7.2, 10.9) OR smokers = 1.0 (0.6, 1.5) OR
non-smoker = 1.0 (0.5, 2.0)
Slide 8
8 Confounding: Smoking, Matches, and Lung Cancer Illustrates
how confounding can create an apparent effect even when there is no
actual true effect In the relationship between matches and lung
cancer, smoking is a ________ factor or a ______________ Smoking
_______ the relationship between matches and lung cancer
Slide 9
9 Estes continues to be confounding puzzle Ray RATTO Friday,
July 20, 2001 2001 San Francisco Chronicle Ray RATTO 2001 San
Francisco Chronicle SHAWN ESTES seemed loath to analyze his own
performance last night, for fear that people would see the first
three innings and use them to obscure the last four. But that's
what made his outing so perfectly Estes-like -- an ongoing argument
with himself that he eventually won. Well, an argument in which he
held his own and his teammates won for him in the bottom of the
ninth. Ramon Martinez lined a game-tying single with two outs, and
Jeff Kent followed two batters later with a bases- loaded walk off
Juan Acevedo to give the Giants a 2-1 victory against Colorado and
move them to within 4 1/2 games of division leader Arizona. It was
in many ways an eye-opening victory for a team that hadn't had one
of this type for a while.
Slide 10
10 Finding: After an initial course of post-exposure
prophylactic (PEP) medication following a sexual exposure to HIV
infection, gay men reported a decrease in the practice of high-risk
behavior over the following year. Reviewer: Perhaps the men simply
withheld the real amount of high-risk behavior they had in order to
be eligible for future courses of PEP. How do you account for this
confounding?
Slide 11
11 Smoking, Matches, and Lung Cancer To be complete, you also
decide to examine the relationship between smoking and lung cancer
independent from the use of matches. What tables should you
construct to do this?
Slide 12
12 Smoking, Matches, and Lung Cancer Stratified Crude Matches
Absent Matches Present OR crude OR CF+ = OR matches OR CF+ = OR no
matches OR crude = 21.0 (16.4, 26.9) OR matches = 21.0 (10.7, 41.3)
OR no matches = 21.0 (13.1, 33.6)
Slide 13
13 Confounding: Smoking, Matches, and Lung Cancer What is the
effect of matches on the relationship between smoking and lung
cancer? Illustrates one important component in the requirements of
a ____________________ (aka a _______________ factor)
Slide 14
14 Confounding: Examples of Magnitude and Direction Stratified
Crude Potential Confounder Absent Potential Confounder Present RR
crude RR CF+ RR CF-
Slide 15
15 Nightlights
Slide 16
16 Nightlights and Myopia Quinn et al. Nature 1999 Prevalence
Ratio =
Slide 17
17 Insert picture with nightlight off
Slide 18
18 Nightlights and Myopia: Two subsequent studies found no
association Zadnik et al. and Gwiazda et al. Nature, 2000
Slide 19
19 Childs Myopia Night Light Parental Myopia X X
Slide 20
20 Insert picture with nightlight on again
Slide 21
21 What kind of variables act as confounders? Properties of a
True Confounder A true confounder (C) must be associated with: the
exposure (E) in question and the disease (D) under study Confounder
D D ANOTHER PATHWAY TO GET TO THE DISEASE ANOTHER PATHWAY TO GET TO
THE DISEASE
Slide 22
22 What kind of variables act as confounders? Properties of a
True Confounder A true confounder (C) must be associated with: the
exposure (E) in question and the disease (D) under study Confounder
D D ANOTHER PATHWAY TO GET TO THE DISEASE ANOTHER PATHWAY TO GET TO
THE DISEASE
Slide 23
23 Lung Cancer Matches Smoking ? ?
Slide 24
24 Properties of a True Confounder Refined Properties:
Association with Exposure A confounding variable can be either the
cause of, the result of, or simply associated in a non- causal
manner with the exposure in question Confounder D D
Slide 25
25 C causes E ? [cardiovascular work-out]
Slide 26
26 Mortality Poor Diet Poverty ? ? E causes C [access to
care]
Slide 27
27 Non-causal relationship between C and E CAD
Slide 28
28 Properties of a True Confounder Refined Properties:
Association with Exposure A confounding variable must be associated
with the exposure in question independent of its association with
the disease in question. It must be associated with the exposure
not simply through its association with the disease (i.e. must be
associated with the exposure among the non-diseased)
Slide 29
29 H. pylori
Slide 30
30 Properties of a True Confounder Refined Properties:
Association with Disease A confounding variable must be associated
with the disease. It need not be a cause of disease; it may merely
be a marker for (i.e. associated with) a cause Confounder D D
Slide 31
31 Early Mortality Hep B and C virus infection IDU ? ? C causes
D
Slide 32
32 C as a marker for D Unknown biologic factor(s)
Slide 33
33 C as a marker for D Unknown biologic factor(s) Maternal Age
Birth Order Down Syndrome
Slide 34
34 Properties of a True Confounder Refined Properties:
Association with Disease A confounding variable must be associated
with the disease independent of its association with the exposure
in question i.e. must be associated with the disease among the
unexposed
Slide 35
35 Lung Cancer Smoking Matches
Slide 36
36 What is NOT a Confounder? A variable that is an intermediate
step in the causal path between the exposure in question and
disease under study is not a confounding variable. E E D D factor I
This factor is not a confounder
Slide 37
37 CCR5 and HIV Disease Progression CCR5 (chemokine receptor)
defect AIDS CCR5 is the receptor for HIV Is a defect in CCR5
associated with slower time to progression to AIDS? HIV viral load
is a powerful predictor of time-to-AIDS and is associated with CCR5
status How should HIV viral load be handled? CCR5 is the receptor
for HIV Is a defect in CCR5 associated with slower time to
progression to AIDS? HIV viral load is a powerful predictor of
time-to-AIDS and is associated with CCR5 status How should HIV
viral load be handled? ? ?
Slide 38
38 CCR5 and HIV Disease Progression Ioannidis et al. Ann Int
Med 2001 Effect of CCR5 defect on development of AIDS Stratified
(adjusted) for HIV plasma viral load Relative hazard: 0.82 (95% CI:
0.6 to 1.2) Is adjusting for HIV viral load the right thing?
Slide 39
39 CCR5 and HIV Disease Progression A variable that is an
intermediate step in the causal path between the exposure in
question and disease under study is not a confounding variable.
CCR5 (chemokine receptor) defect AIDS HIV viral load HIV viral load
is clearly an intermediary factor Crude (unadjusted) association:
Relative hazard: 0.71 (95% CI: 0.5 to 0.9) HIV viral load is
clearly an intermediary factor Crude (unadjusted) association:
Relative hazard: 0.71 (95% CI: 0.5 to 0.9)
Slide 40
40 Exercise and CAD Is HDL associated with exercise? Is HDL
associated with CAD? When evaluating the relationship between
exercise and CAD, is HDL a confounder or an intermediary? Exercise
CAD HDL
Slide 41
41 It depends on the pathway under investigation If interest is
in a pathway other than through HDL, then HDL is a confounder Here,
HDL is extraneous to pathway under study Confounding factors are
extraneous factors Exercise CAD not yet specified mechanism HDL ?
?
Slide 42
42 Exercise and CAD If there is only one pathway in question,
then HDL is an intermediary variable. e..g., Does exercise
influence CAD risk in a newly studied population (elderly Asians)?
Hence, classification of HDL as confounder or intermediary depends
upon the biological pathway under investigation Exercise CAD HDL
HDL is not a confounder here
Slide 43
43 What is NOT a Confounder? Variables that are the RESULT of
the disease, regardless of their association with the exposure are
NOT confounders Confounder D D
Slide 44
44 Lung CA Smoking Cough ? ? Cough is not a confounder. Do not
adjust for it! Cough is not a confounder. Do not adjust for
it!
Slide 45
45 When Planning a Study, Which Factors Should be Considered as
Potential Confounders? Any factor for which prior evidence
indicates it is a confounder and In newer research areas: factors
known to be associated with the disease and which may be associated
with exposure When in doubt, plan on measuring ALL factors
associated with the disease i.e. If you dont, you may regret it
later
Slide 46
46 Seeking cause of high Marin cancer rates Activists canvass
residents to search for trends Sunday, November 10, 2002 Thousands
of volunteers scattered across Marin County under baleful skies
Saturday in an unprecedented grassroots campaign against the
region's soaring cancer rate. Armed with surveys, some 2,000
volunteers went door to door in every neighborhood in the county,
asking people whether they or anyone in their household has ever
been diagnosed with cancer in Marin. The volunteers hope to collect
enough money to hire an epidemiologist to analyze the data for use
in future studies.
Slide 47
47 Methods to Reduce Confounding during study design:
Randomization Restriction Matching during study analysis:
Stratified analysis
Slide 48
48 Methods to Prevent or Manage Confounding D D D D or
Slide 49
49 Methods to Prevent or Manage Confounding By prohibiting at
least one arm of the exposure- confounder - disease structure,
confounding is precluded
Slide 50
50 Randomization to Reduce Confounding Definition: random
assignment of subjects to exposure (or treatment) categories All
subjects Randomize One of the most important inventions of the 20th
Century! Exposed Unexposed
Slide 51
51 Randomization to Reduce Confounding D D
Slide 52
52 Randomization to Reduce Confounding All subjects Randomize
Applicable only for intervention (experimental) studies Special
strength of randomization is its ability to control the effect of
confounding variables about which the investigator is unaware Does
not, however, eliminate confounding! Exposed Unexposed