Lecture 8: Causation in Epidemiology Flashcards
What are other ways to refer exposures by?
Causes, risk factors, independent variables
What are other ways we think about outcomes?
effects, diseases, injuries, disabilities, deaths, dependent variables
Statistical association versus biological causation
For causation, we need statistical association (how much is this affection the population) and the biological mechanisms. The presence of statistical association alone does not necessarily imply a causal relationship.
Association (relationship)
Statistical dependence between two or more events, characteristics, or other variables.
Ex. as one increases/decreases, what happens to the other variable.
Causality (causation / cause-effect relationship)
relating causes to the effects they produce.
Cause
An event, condition, characteristic (or a combination of them) which plays an important role/predictable change in occurrence of the outcome.
Deterministic Causality
cause closely related to effects, as in “necessary”/”sufficient’ causes
Necessary cause
Has to be present in order to produce the outcome
Sufficient cause
inevitably initiates or produces an effect. Sufficient, is when the factor can produce the outcome by itself. Includes “component causes.”
Component Causes
Together they constitute a sufficient cause for the outcome in question
Component Causes
Together they constitute a sufficient cause for the outcome in question
Probabilistic Causality
Weak relationship, neither necessary nor sufficient.
Effect Measures/Impact Fractions
- effect measures and impact fractions are closely related to the strength of association
- The higher the effect measures and population attributable risk, the more exposure is predictive of the outcome in question
Necessary and Sufficient
- if the factor is present, the disease will always occur
- without the factor, the disease never develops
only factor A -> Disease
Necessary but Not Sufficient (Alone)
- each factor necessary but not in itself sufficient to cause the illness in itself, all are necessary to cause disease, but individually, none are sufficient to cause the disease.
- each risk factor alone cannot cause disease
- Thus multiple factors are required often in a specified temporal sequence
Sufficient but not necessary
- that factor alone can produce the outcome but it is not the only factor that produces the outcome
- that outcome can still occur if that factor is not present
Neither Sufficient Nor Necessary
- None of the risk factors are enough or alone to cause a disease
- There are multiude risk factors
Henle-Koch’s Postulates
Four postulates should be met before a causal relationship between a disease agent and disease.
1. The agent must be present in every case
2. Agent must not be found in cases of other disease
3. Once isolated, the agent must be capable of reproducing the disease in experimental animals.
4. The agent must be recovered from the experimental disease produced.
Bradford Hill’s Considerations for Causality
- Strength of association
- Consistency
- Specificity
- Temporality
- Dose-response relationship
- biological plausibility and coherence
- Experiment-randomized controlled trials
Strength of Association
- The larger an association between exposure and disease, more likely to be causal
- Percival Pott’s study: scrotal cancer is 200 times more likely to those exposed to chimney soots
Consistency
- multiple epidemiologic studies using various locations, populations, and methods to show consistent association
- over 100s studies that show smoking and lung cancer are associated
Specificity of Outcome
- more likely to be causal when they are specific
- the exposure causes only one disease
- asbestos causes lung cancer
Temporarilty
- Exposure must precede the onset of disease
- is low serum cholesterol a cause of colon cancer or is the early phase of colon cancer cause low cholesterol
Dose-response relationship
- increase the dose/exposure, increases the response
- those with alcohol addiction die faster than those who consume less or abstain
Biological plausibility and coherence
- align with pre-existing theory, biology, statistical support
- agree with biological or clinical data?
- Hepatitis B associated with liver cancer -> viral copies identified in cancerous liver
Experiment
- does the removal of the exposure affect the frequency of the outcome
Screening in Epidemiology
Process of identifying an unrecognized disease or defect by the application of tests, examinations, or other procedures.
Classifies asymptomatic people as likely or unlikely to have a disease or defect.
Purpose of screening in epidemiology
Delay onset of symptomatic or clinical disease. Improve survival.
Screening Criteria
- suitable disease
- suitable test
- suitable screening program
Suitable Disease
- has serious consequences
- is progressive
- disease treatment must be effective at an earlier stage
- prevalence of the detectable pre-clinical phase must be high
- ex. breast cancer, cervical cancer, hypertension
Suitable Test
- inexpensive
- easy to administer
- has minimal discomfort
- high level of validity and reliability
*validity is more important than reliability
Valid Test
- does what it’s supposed to do
- classify people with pre-clinical disease as positive
- people without pre-clinical disease as negative
Reliable Test
Gives you same results on repetition
How to measure test validity
sensitivity and specificity
valid test has high sensitivity and specificity
Sensitivity
enables you to pick up the cases of disease
a / a + c = those that test positive / all with disease
Specificity
enables you to pick out the no diseased people
d / b + d = those that test negative / all without disease
Suitable Screening Program
Application of a specific test in a specific population for a specific disease
Evaluation of Screening Program
Predictive Value of a Positive Test (PV+)
Predictive Value of a Negative Test (PV-)
Predictive Value of a Positive Test (PV+)
a / a + b = number who test positive with disease/ number with positive result
Predictive Value of a Negative Test (PV-)
d / c + d = number who test negative without disease / number with negative result
Bias with evaluating a screening program
volunteer bias
lead-time bias
length bias
Volunteer bias
- people who choose to participate in the screening program may be healthier or at a higher risk of developing the disease than those that don’t participate
- those with higher health seeking behaviours
motivated to maintain health and wellbeing - access to resources, opportunities, and networks
Lead-time bias
- the amount of time by which the diagnosis was advanced due to screening
- that survival may erroneously appear to be increased among screen-detected cases simply because the diagnosis was made earlier in the course of the disease
Length Bias
- less aggressive forms of a disease are more likely to be picked up in a screening program because they have a longer detectable pre-clinical phase
- less aggressive forms of disease usually have better survival
