Scribd Logo
Upload

Welcome to Scribd!

  • 33 views

    PDF Wk4 - 04 Bradford Hill Criteria

    Uploaded by

    jackbane69
    The document discusses the Bradford Hill criteria for causation. The Bradford Hill criteria are guidelines used to evaluate whether an observed association is causal. There are 9 criteria: temporality, biological plausibility, strength of association, consistency, specificity, dose-response relationship, reversibility, and residual confounding. The criteria help epidemiologists and public health experts determine if a risk factor could cause a disease or injury.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    PDF Wk4 - 04 Bradford Hill Criteria

    Uploaded by

    jackbane69
    33 views5 pages
    The document discusses the Bradford Hill criteria for causation. The Bradford Hill criteria are guidelines used to evaluate whether an observed association is causal. There are 9 criteria: temporality, biological plausibility, strength of association, consistency, specificity, dose-response relationship, reversibility, and residual confounding. The criteria help epidemiologists and public health experts determine if a risk factor could cause a disease or injury.

    Original Title

    PDF Wk4_04 Bradford Hill Criteria

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses the Bradford Hill criteria for causation. The Bradford Hill criteria are guidelines used to evaluate whether an observed association is causal. There are 9 criteria: temporality, biological plausibility, strength of association, consistency, specificity, dose-response relationship, reversibility, and residual confounding. The criteria help epidemiologists and public health experts determine if a risk factor could cause a disease or injury.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    33 views5 pages

    PDF Wk4 - 04 Bradford Hill Criteria

    Uploaded by

    jackbane69
    The document discusses the Bradford Hill criteria for causation. The Bradford Hill criteria are guidelines used to evaluate whether an observed association is causal. There are 9 criteria: temporality, biological plausibility, strength of association, consistency, specificity, dose-response relationship, reversibility, and residual confounding. The criteria help epidemiologists and public health experts determine if a risk factor could cause a disease or injury.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 5

    Causation

    The Bradford Hill criteria

    Narrated by Prof Brendan Girdler-Brown

    Introduction
    In this video we will explain and learn about the Bradford Hill criteria.

    The process of applying these criteria for causation is called causal inference.

    By the end of the video you should understand the different criteria and be able to
    critique and apply them.

    Bradford Hill 2020/2023

    Causal Inference
    “Causal inference” is……

    • The process of deciding if observed associations are likely to be causal; and


    • Use of guidelines and making of judgements about causality.

    The Bradford Hill criteria are a popular aid to deciding about causation

    However, only temporality is a necessary condition for causation


    (the putative cause must come before the effect)

    Bradford Hill 2020/2023

    1
    Criteria for Causal Inference
    Modified Bradford Hill criteria for causation, 1965:

    • Temporality;
    • Biological plausibility;
    • Strength of association;
    • Dose-response;
    • Consistency;
    • Reversibility;
    • Study design; and
    • Judging the evidence

    Bradford Hill 2020/2023

    The Bradford Hill Criteria


    The Blue Dog Ate the Red Cat
    (a mnemonic to help you recall the long list! Pron> “NEMONIK”)
    T Temporal relationship?
    B Biological/ pharmacological plausibility?
    D Dose-response OR cumulative exposure?
    A Association: recognized with product class?
    R Re-challenge/ recurrence after re-exposure?
    C Causes: Other causes?
    • Pre-existing conditions
    • Concurrent conditions
    • Other factors (e.g. family history, environment)
    • Other interventions (e.g. medications)

    Bradford Hill 2020/2023

    Bradford Hill Criteria


    Temporality

    The exposure must precede the disease/ outcome/ effect

    For disease with long latency the exposure must exceed the latency, e.g.
    Asbestos exposure and mesothelioma;
    HPV infection and cancer of the cervix

    If exposure is on-going and disease is chronic with no clear onset time, difficult to apply

    Bradford Hill 2020/2023

    2
    Biological Plausibility
    Causation should be plausible given what we know about disease pathogenesis

    • Support from laboratory evidence or theory using biological reasoning


    - e.g. the effect of HIV on the immune system
    • Note that lack of plausibility may reflect lack of medical/ biological knowledge

    • Seemingly implausible causal routes may later be shown to be causal


    - e.g. the effect of lead exposure on the CNS, even at low levels

    Bradford Hill 2020/2023

    Bradford Hill Criteria


    Consistency is …..

    • Demonstrated if several studies yield the same result;


    • Especially if many different study designs were used in different settings and times.

    • A lack of consistency does not exclude a causal association; since


    • Different exposure levels may reduce the impact of the exposure in some studies

    Bradford Hill 2020/2023

    Strength of Association
    OR or RR that is far from = 1

    Stronger associations are less likely to be due to bias or sampling error

    Relative risks or odds ratios > 2 or < 0.5 are considered “strong” associations

    However weak associations do not preclude causation

    Bradford Hill 2020/2023

    3
    Dose-Response Relationship

    If increased levels of exposure are associated with increased disease rates, even if only

    “Low” “medium” and “high” This is quite convincing

    Watch out for where there are threshold values of an exposure, however.

    Bradford Hill 2020/2023

    10

    Reversibility

    If the exposure is removed and the disease recovers, this suggests causation;

    e.g. cessation of smoking results in reduced risk of lung cancer going forward

    But if the disease or effect is not reversible, this reversibility may not be evident.

    Bradford Hill 2020/2023

    11

    Judging the evidence

    There are no completely reliable criteria for causation

    Causal inference is usually tentative only

    Temporality is essential and …..


    Plausibility,
    Consistency, and
    Dose-response relationships

    Are ? the best indicators after temporality

    Bradford Hill 2020/2023

    12

    4
    Thank You

    13

    You might also like