Disease Causation Flashcards
Importance of cause
-so that we can intervene and prevent disease
-Cause is any factor that produces a change in severity or frequency of outcome
**do not need all causal factors
Inductive reasoning
-the process of making generalized inferences about causation based on repeated observations
Inductivism and logical fallacies
-After this, therefore because of this
Koch’s postulates
Causal if:
-present in all cases of disease
-it does not occur in another disease as a fortuitous and non pathogenic parasite
-it is isolated in pure culture and induces the same disease in other animals
Issues with Koch’s postulates
-ignores environmental factors
-not applicable to non infectious diseases
Epidemiology vs lab
-hard to recreate in lab (ethical, don’t know how, cost)
-complex issues occurring in natural world
-discussions of causation are usually limited to observational research rather than experimental research
Experimental studies
-traditionally, the Gold standard
-randomize individuals to receive a factor and some to receive nothing
*factor precedes disease and other variables accounted for by randomization
-compare outcomes of tx and control groups
-Assume if groups had been switched we would have got the answer
Observational studies
-Estimate the outcome differences between individuals that happen to vary in their exposure status
-use matching and restriction to minimize differences between groups
-Measure association between changes in exposure and outcome
Limits to experimental studies
-often difficult to duplicate realistic dose, exposure pathway, complete set of typical cofactors
-difficult to carry out experiments that resemble real world conditions
Observational studies
-Have environmental exposure AND disease or outcome
-must have complete and careful description of the referent group
But need to be able to determine if this is a cause-effect relationship
Trials of treatedd vs untreated
Not really allowed now. Usually have to give one medication to one and the second best treatment to the other
Cohort studies
-Classify groups based one exposure
-follow these groups forward in time
What is cohort studies reported as?
Report as relative risk
*compare attack rates and then get relative risk
*can be used to look at more than one disease resulting from a specific exposure
*Closest observational study to randomized control trial
Case-control studies
-define groups of diseased and healthy animals
-assess whether the animals in the two groups have differences in past exposure to different risk factors
**hard because looking back in time= RECALL BIAS
**had to determine whether the exposure came before the disease actually began (eg. cancer)
How are Case-control studies reported?
Calculate the odds ratio
-estimates the relative risk provided that the incidence of disease is low and cases/controls are random
**good for rare disease
*can assess more than one exposure in same study
Statistical significance and causation
Statisitical significance does not equal cause
**to prove causal association we need to describe a chain of events from cause to effect at the molecular level
Confounding
The effect of an extraneous variable that can wholly or partly account for an apparent association between variables in an investigation
*confounding can mask a real association
eg. large herd size and aprons= leptospirosis
Confounder components
- be associated with response variable
- Be associated with the risk factor of interest
- Not be an intervening or intermediate step between the risk factor and response
Component model of causation
-all disease is multifactorial
-cause is sufficient if it produces effect
-a cause almost always comprises a number of component causes
**A particular disease may be produced by different sufficient causes
Necessary cause
-A risk factor that is a component of every sufficient cause
Components of a sufficient cause
-Factors may be present concomitantly or may follow one another in a chain of events
Causal web
A number of chains with one or more factors in common
*includes indirect causes activating direct causes
Indirect vs direct causes
Indirect= the effects of exposure are mediated through one or more intervening variables
direct= often the proximal causes emphasized in therapy
Causal mechanism and strength of an association
Mechanism remains constant
Strength between an exposure of interest and the outcome will vary
*depends on distribution of risk factors
Interaction among causes
Two or more component causes acting in the same sufficient causes interact causally to produce disease
Hills criteria for causality
-Temporality
-strength of association
-biological gradient or dose response
-coherence or plausibility
Others:
-consistency
-specificity
-analogy
-experimental evidence
Temporality
-cause must always precede effect in time
-but the same factor could occur again after disease in some individuals
-often difficult to establish time sequence especially with surrogate measutes
Strength of Association
-a strong statistically significant association between a factor and disease increases the likelihood that the factor is causal
-assumes less likely for residual confounding to explain results
Issue with strength of association
Depends on distribution of other components of sufficient cause
eg. weak associations (tobacco smoke and lung canceR) can be considered causal
Biological gradient
-a dose response relationship between a factor and disease increases the plausibility of a factor being causal
Consistency
-repeated observations of an association in different populations under different circumstances
**role of systemic reviews and meta analysis
Coherence/plausibility
-compatibility with existing knowledge
-it is more reasonable to infer that a factor causes a disease if a plausible biological mechanism has been identified than if mechanism is unknown
Specificity
-A cause leads to a single effect or an effect has one cause
**not necessary but can support when logical deduction from causal hypothesis
Analogy
-too subjective and open
-possible source of new hypothesis
Experimental evidence
-clinical trials, animal lab experiments or both
-uncertainty in extrapolating across species and outside tested dose ranges
Does using hills criteria balance out interpretations?
No, there are lots of different perspectives and even using criteria, epidemiologists only agreed 68% of the time
Causal inference
-Need a mathematical association between the exposure and the hypothesized effect or outcome
**the outcome have a monotonic association with the increasing exposure
-besides temporality, there is no criteria that is needed or sufficient
Potential errors that can lead to observed association
- Chance
-assessed through correct application of statistical analysis - Systemic error in the design of the study or data
-sampling bias
-misclassification of exposure or disease status - Confounding or mixing of effects
-results from a 3rd unaccounted risk factor associated with both exposure and disease
What does consistency across different study populations mean?
Means that information is supportive
