480 likes | 495 Views
Causality assessment Theoretical background. Prof E. P. van Puijenbroek, MD, PhD October 2018. Learning Objectives. Understand the concept of counterfactuals Apply the role of the Bradford Hill criteria in causality assessment. Outline. Counterfactual theory Hill criteria
E N D
Causality assessment Theoretical background Prof E. P. van Puijenbroek, MD, PhD October 2018
Learning Objectives • Understand the concept of counterfactuals • Apply the role of the Bradford Hill criteria in causality assessment
Outline Counterfactual theory Hill criteria Examples exam-questions
Question 1 It is important to annotate the strength of the causal relationship between drug and adverse drug reaction • Agree • Do not agree
Question 2 Do you think it is possible to determine the strength of the causal relationship between drug and adverse drug reaction? • Agree • Do not agree
Counterfactualtheory • When B occurs after X and B does not occur without X, than X is considered to be the cause of B- Someone entered his house (B) because the door was open (X)- Because he uses this drug (X), he got an Adverse Drug Reaction (B) • X causes B because the counterfactual “when X is not used, that B does not occur” is true.
Counterfactual theory Drug X Event B
Counterfactual theory Drug X Event B
Counterfactualtheory • Theory, that tries to answer the question “how do I know if something is a cause of a certain event”? • Is a theoretical approach but it does not provide a practical solution for the assessement of the causal relationship • How to solve this problem? Study setting: use control groups Assessement of individual cases: causality models
The epidemiological approach.. • Exchangeability of groups- Patients (using X) occurrence of B, given use of X- Patients (no X) occurrence of B, given no use of X • When both groups are “exchangeable”, and B differs between both groups, we stat the B has a causal relationship with X • Size of both groups should be very large when a small actual relationship exists
Conclusive epidemiological evidence? • Information about “chances” for events to occur in groups of patients may be available. • In medical practice, methods to assess the strength of the causal relationship are also needed.
Outline Counterfactual theory Hill criteria Examples exam-questions
Hill AB. The environment and disease: Association or causation? Proceedings of the Royal Society of Medicine 1965;58:295-300
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Strength of evidence … focus of epidemiology …
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Time to onset analysis Van Holle et al. PEDS 2012;21:603-10
Minutes… Days… Years…
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Specificity of the association Aronson & Hauben. BMJ 2006;333:1267-69
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Bradford Hill Criteria • Strength of the evidence • Order in time • Consistency • Plausibility • Specificity • Biological gradient • Coherence • Experiment • Analogy
Intrinsic Intrinsic / Extrinsic experiment Intrinsic Biological gradient Intrinsic
“None of these nine viewpoints can bring indisputable evidence for or against a cause and effect hypothesis… What they can do, with greater or less strength, is to help answer the fundamental question – is there any other way of explaining the set of facts before us, is there any other way to answer equally, or more, likely than cause and effect?” Hill AB. The environment anddisease: Association or causation? Proceedings of the Royal Society of Medicine 1965;58:295-300.
Summary • Causation may be difficult to prove in practice • The counterfactual theory helps to understand causality • Bradford Hill criteria play a role in the assessement of the causal relationship between drug and adverse event
Outline Counterfactual theory Hill criteria Examples exam-questions
Exam question –example 1 The idea of counterfactual theories of causation is that the causal relationship between these events can be expressed by so called “counterfactual conditionals”. An example is for instance: “When B occursafter X and B does notoccur without X, than X is consideredtobethecause of B” This statement is: • True • False
Exam question –example 2 One of the criteria used in causality model, originally formulated by the epidemiologist Bradford Hill, is the strength of the association. Where would you look for his type of information in daily practice? • In medical literature • In de medical history of the patient • In de medication history of the patient • None of the above
Learning Objectives • Understand the concept of counterfactuals • Apply the role of the Bradford Hill criteria in causality assessment
Literaturecausality assessment Hill, Austin Bradford (1965). "The Environment and Disease: Association or Causation?". Proceedings of the Royal Society of Medicine. 58 (5): 295–300.. PMID 14283879. Höfler M (2005). "The Bradford Hill considerations on causality: a counterfactual perspective?". Emerging themes in epidemiology. 2 (1): 11. doi:10.1186/1742-7622-2-11. PMID 16269083.