Association and Causality

Question 1

3 types of associations

Answer 1

artifactual
non causal
causal

Question 2

artifactual associations

Answer 2

arise from significant bias and/or extensive confounding

Question 3

non-causal associations (2 things here)

Answer 3

two different ways they can occur

1. the disease may cause the exposure, instead of
   other way around

are people less likely to have arthritis if they exercise, or is exercise the cause of arthritis? Which is it?

2. the disease and the exposure are both associated with a third variable (such as a confounder)
   example: down syndrome and birth order

Question 4

Koch’s postulates

Answer 4

4
causal relationship for infectious disease

1. disease must be present in every instance
2. must not be found in cases of other diease or healthy people
3. must be capable of isolation
4. must be recovered from experimentally induced animals

Question 5

Mill’s cannons

Answer 5

cause of any effect must consist of a constellation of components that act in concert, rather than just one dependent/independent variable relationship

Question 6

Sufficient cause

Answer 6

set of minimal conditions that have to be met

cause must precede disease, and the disease must always occur

rare

Question 7

Sufficient causes can have

Answer 7

required “components” that collectively act to induce disease

Question 8

Necessary cause

Answer 8

cause precedes disease

cause must be present for the diease to occur, yet the cause may also be present without the disease occuring

TB is an example

Question 9

RF: risk factor, also known as

Answer 9

component clause

characteristic that, if present and active, increases the probability of a particular disease
some patients may be susceptible before component cause causes disease

Question 10

Multiple causation

Answer 10

most disease have a variety of causes, to understand them mathematically they have to be controlled for

restricting/matching/stratification

Question 11

Induction….what rules?

Answer 11

Hill’s guidelines

In what circumstances can we pass from observed association to a verdict of causation?

Hill didn’t like hard and fast rules, didn’t think they could generate a way to judge the likelihood of an event via caustation

Question 12

Hill’s Criteria

Answer 12

1. Strength
2. consistency
3. temporality
4. biologic gradient
5. plausibility

“RR/OR/HR”

Question 13

Hill’s guidelines #1

Answer 13

Strength refers to the size of the association

RR/OR/HR

the greater the association, the more convincing it is.

Question 14

Why is a strong association neither necessary nor sufficient for causality?

Answer 14

because causality is multifactorial, the product of many occurrences and conditions, and these may be distorted by confounders like biases, mistakes in measurement, or traits of the pathogen.

Question 15

Hill’s GL #2

Answer 15

Consistency

reproducibility

the repeated observation of association in different populations

consistency may still obscure the truth

Question 16

Hills GL # 3

Answer 16

Temporality

is the necessity that the cause precede the effect/outcome in time

induction and latent period times for instance

Question 17

Example of #3

Answer 17

a number of studies have demonstrated higher rates of cancer in former smokers in their first year of cessation than those who continue to smoke.

Question 18

Hill’s # 4

Answer 18

Biologic Gradient refer’s to the observation of a gradient of risk

light smokers = 5 times more likely to develop lung cancer than non-smokers

heavy smokers = 15 times more likely to develop lung cancer than non smokers

gradient

Question 19

“dose response”

Answer 19

is a gradient of risk associated with degree of exposure

Question 20

Hill’s #5

Answer 20

Plausibility

is it feasible? diagnosis may be made on prior beliefs which may be flawed

Question 21

Problems of causal research

Answer 21

bias–confounding–effect modification

synergism

Question 22

synergism

Answer 22

the interaction of 2 or more presumably causal variables so that the combined effect is clearly greater than the sum of the individual effects

example: interaction of health compromising behaviors on preterm births