Download presentation
Presentation is loading. Please wait.
Published byMarshall Roman Modified over 10 years ago
1
WHO IS IN CHARGE OF CAUSATION? Olaf Dekkers/20-11-2014 Dept Clinical Epidemiology Aarhus and Leiden
2
Le Malade Imaginere (1673), Act III
3
Basic problem I We don’t usually see that x causes y Causal judgement is based on inference
4
Basic problem II Inferences can be wrong Statistics do not differentiate between association and cuasation
5
Basic problem III We are taught, over and over, to be sceptical Why most research findings are false. Ioannidis 2005 Am J Epidemiology does not allow the use of the word ‘effect(s)’ to denote association(s) It may promise more than an observational study can deliver Only studies in which there is an intervention to change exposure can study causes It misleads the uncritical reader
6
So… Making causal inferences in clinical research is not straightforward (old news) But we should not refrain from searching for causes, as clinical research aims to improve health
7
Where do we stand today? Theories of causation -Regularity theories (Hume) -Probabilistic theories -Counterfactual theories -Interventionism
8
Counterfactual theory of causation
10
Counterfactual theory cf Paul LA Counterfactual theories Intuitive interpretation: If it is the case that Y occurs after X and that Y does not occur in the absence of X, than X is a cause of Y More formally: “X causes Y because the counterfactual ‘if not X then not Y’ is true.”
11
“X causes Y” Counterfactual thinking is intuitively appealing Italy lost the game (Y), because the referee was not objective (X) She was late (Y), because of the snow (X) She/he fell deeply in love with him/her (Y), because of his beautiful eyes (X) Attacking Pearl Harbor (X), caused the USA to enter WWII (Y)
12
Why do we need causation theories? What is a philosophical theory of causation giving an account for? The theory facilitates causal inference, and gives an answer to the question how we can know whether X is a cause of Y (epistemic notion) In epidemiology the epistemic notion is more relevant The theory might facilitate our understanding of what causation (in general) stands for (semantic/ontologic notion)
13
Epistemic notion of causation Epistemic notion Obesity (x) is a cause of mortality (y) because the counterfactual ‘if not X then not Y’ is true (in a least one person) (Counterfactual theory) Obesity (x) is a cause of mortality (y) because being obese increases mortality probability (Probability theory)
14
Ontologic notion of causation Semantic/ontologic notion There are no working forces, there are only probabilities (Probability theory) This notion is in a counterfactual framework difficult as the theory refers to non-occurring situations (or even worse: non-existing worlds)
15
Counterfactuals: philosophical problems Pre-emption (see next slide) Counterfactual dependence is not a necessary condition for causation Prevented causes Think of a new extremely powerful weapon (X) causing world-destruction (Y). Causation cannot be judged in a counterfactual framework Many irrelevant causes Counterfactual dependence is not a sufficient condition for causation
16
Pre-emption Smoking Hypertension AMI
17
An unstable christmas tree Counterfactual theory needs many auxillary theories to keep it upright
18
Counterfactuals in epidemiology What makes the theory so appealing in epidemiology? Counterfactuals point towards non-existing situations: ‘If not X, etc’ In medical research counterfactuals do not exist Even not in cross-over studies But: we are used to control groups Control group as approximation for the counterfactual Counterfactual thinking resembles the way we do studies
19
Counterfactual in epidemiology BUT No practical decision rule to infer causality Refers to non-occurring situations There is however a set of basic assumptions for causal inference in the counterfactual framework (cf Hernan): ‘Exchangeability’ ‘Positivity’ X=1 and X=0 (exposure contrast) should be well-defined
20
Counterfactuals and exchangeability Patients (X=1) Outcome Y=y|X=1 Patients (X=0) Outcome Y=y|X=0
21
Counterfactuals and exchangeability Patients (X=1) Outcome Y=y|X=1 Patients (X=0) Outcome Y=y|X=0 Counterfactual definition of causation: If it is the case that Y occurs after X and that Y does not occur in the absence of X, than X is a cause of Y This can be inferred from a comparison under exchangeability: Two groups would have had an identical outcome (Y=y) if the exposure (X=1 or X=0) would have been the same
22
Exchangeability Two groups would have had an identical outcome (Y=y) if the exposure (X=1 or X=0) would be the same If the baseline prognosis is the same, then a difference in outcome between groups has a causal interpretation This is threatened by confounding in observational studies
24
Confounding and selection-bias Causal question Z in an observational study Epidemiological answer: bias and confounding unlikely Then we are (philosophically speaking) referring to a counterfactual theory of causation
25
RCTs: a counterfactual paradigm
26
A RCT
28
DREAM
29
RCTs The randomized design fits the counterfactual framework The control group approaches the counterfactual situation by design (randomization) It is deprived from theory
30
RCTs The randomized design fits the counterfactual framework It gives decision rule for causation: if a treatment x is randomized, then a difference in outcome Y between two groups has logically a causal interpretation because the counterfactual statement holds
31
RCT: a counterfactual paradigm The counterfactual model and the RCT fit well Stop here: don’t move philosophically further
32
Epistemic reduction? The counterfactual model and the RCT fit well That does not mean that we can only have causal inference in case of randomization Reason: the counterfactual model is (as we have seen) neither sufficient nor necessary to infer causality
33
Ontologic reduction? The counterfactual model and the RCT fit well A reason for some to accept only interventions as true causes (interventionism) This does not mean that there is only causation in case of a (randomized) intervention Things you can hardly intervene like SES
34
Interventionism Makes causation human-like Many phenomenon difficult to capture in an interventionist framework Big Bang Volcano Eruptions Postulation of hypothetical interventions (which do not differ from counterfactuals)
35
Who is in charge of causation?
36
Where do we stand today?
37
What do we believe? Option 1: Roziglitazone reduces blood glucose level because the counterfactual (if not X then not Y) is true? Option 2: Roziglitazone reduces blood glucose level because of an intrinsic drug capacity to do so?
38
Counterfactual theory presupposes causation If we adopt option 2: The counterfactual notion is true because x is a cause of y, and not the other way round Counterfactual thinking can still be used to judge causality as it is be derived from causality But: it can not be the only argument for causation
39
How foolish are we today? This was not shown in a RCT We have no ultimate counterfactual proof
40
How foolish are we today? JAMA 1982 RR roken – myocardinfarct = 2.0 Risico normaliseert niet na weghalen risicofactor
41
We (researchers) have to do more There are reasons to think that risk factors and drugs have some inherent effects But we have to do more for causal inference Experiments (thinking along the lines of counterfactual theories is helpful) Explaining why x causes y Stating the conditions under which x causes y This means that the philosopher is not in charge of causation
42
In summary Counterfactual thinking fits epidemiology, probably because we use preferably a controlled study design Counterfactuals can be helpful when thinking how to design/perform/analyze a study Causality can not be reduced to counterfactual dependence, and causal inference requires more than counterfactual dependence
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.