Interpreting epidemiological findings (1 of 3)

Question 1

Koch’s postulates for if a pathogen is causative of a disease? (old)

Answer 1

• microorganisms must be found in abundance in all organisms with the disease, but not in healthy organisms
• microorganisms must be isolated from a diseased organism and grown in pure culture
• the cultured microorganism should caused disease when introduced into a healthy organism
• microorganism must be reisolated from the inoculated, diseased experimental host and identified as being identical to the original specific

Question 2

what is the problematic postulate from koch?

Answer 2

• the cultured microorganism should caused disease when introduced into a healthy organism
• unethical to intentionally subject someone to an infection

Question 3

bradford-hill causative criteria (9)

Answer 3

• strength
• consistency
• specificity
• temporality
• biological gradient
• plausibility
• coherence
• experiment
• analogy

Question 4

strength aspect of bradford-hill causative criteria

Answer 4

when there is a stronger relationship between and exposure and outcome, it makes it more likely there’s some causative element going on

Question 5

consistency aspect of bradford-hill causative criteria

Answer 5

• consistent findings across setting tended to rule out error or fallacies that might befall one/two studies
• use carefully and with difficulty to determine causation

Question 6

specificity aspect of bradford-hill causative criteria

Answer 6

• most criticised of BH’s points
• he does assert that a specific disease arising among specific workers is valuable if supporting the argument for causality
• hard to apply when issue is multifactorial

Question 7

temporality aspect of bradford-hill causative criteria

Answer 7

it’s insufficient for exposure A and outcome B to coexist

• A must commonly precede B
• cross-sectional studies can’t always assess which came first
• longitudinal studies are useful here

Question 8

biological gradient aspect of bradford-hill causative criteria

Answer 8

idea of dosage

- e.g. smoking more causes more lung cancer

Question 9

plausibility aspect of bradford-hill causative criteria

Answer 9

• relationship shouldn’t be implausible

- where deficient understanding this may be hard to assess

Question 10

coherency aspect of bradford-hill causative criteria

Answer 10

association ought to be consistent with existing theory and knowledge
- overlaps to some extent with biological plausibility

Question 11

experiment aspect of bradford-hill causative criteria

Answer 11

• evidence from experimentation should be supportive of the proposed link
• though scientifically desirable, experimentation is often not ethical when dealing with public health issues

Question 12

analogy aspect of bradford-hill causative criteria

Answer 12

• drawing upon analogous findings, we may make inference on the relationship
• relies on prior knowledge that doesn’t always exist

Question 13

external validity means

Answer 13

findings from a study can be applied to other settings, not where the research was conducted

Question 14

bias

Answer 14

an inference is valid when there is no bias
- bias is any trend in the collection, analysis, interpretation, publication or review of data that can lead to conclusions that are systematically different from the truth

Question 15

which errors are allowable

Answer 15

random

• large sample cancels out errors
• systematic errors are the problem

Question 16

three types of bias

Answer 16

• selection bias
• information bias
• confounding

Question 17

External validity is synonymous with:

Answer 17

generalisability

Question 18

definition of internal validity.

Answer 18

The extent to which findings accurately describe the relationship between exposure and outcome in the context of the study

Question 19

A study, which is designed very well, and implemented equally well - but which is carried out a single teaching hospital in London can be said to have high:

Answer 19

Internal validity

Question 20

The findings of a study which is designed well but which loses more participants than expected are submitted for peer review. What might be called into question by the peer reviewers?

Answer 20

internal validity

Question 21

The findings of a study which is designed poorly but implemented exceptionally well, is submitted for peer review. What might be called into question by the peer reviewers?

Answer 21

internal validity

Question 22

An international multicentre randomised controlled trial with triple blinding is reported in the literature. On the basis of this sentence alone, what is likely?

Answer 22

Good external validity

Question 23

selection bias occurs when:

Answer 23

individuals or groups in a study differ systematically from the population of interest leading to a systematic error in an association or outcome.
- e.g. study on consumption of red meet causing cancer only advertised in vegetarian shops might impact cohort composition

Question 24

example of selection bias

Answer 24

Berkson’s bias

• hospital- based control study using hospitals patients
• healthy worker effect - active workers more likely to be healthy compared to those who have retired/ stopped working

Question 25

ways to reduced selection bias

Answer 25

• controls representative of target population
• minimise non-response
• compare respondent with on-respondents

Question 26

information bias

Answer 26

• misclassification of exposure or disease status or both
• any systematic difference from the truth that arises in the collection, recall, recording and handling of information in a study, including how missing data is dealt with

Question 27

interviewer bias is an issue for which type of bias?

Answer 27

information bias

- mitigated if interviewer doesn’t know the health status of participant

Question 28

recall bias is prevented how

Answer 28

• using non-subjective methods for assessment e.g. biomarkers and medical records

Question 29

non-differential misclassification does what to odds ratio?

Answer 29

• always biased towards the null

Question 30

response bias describes what?

Answer 30

the tendency of participants to respond to questions in a way that is more socially acceptable.
–> This not a type of selection bias.

Question 31

information bias can include

Answer 31

• recall bias
• response bias
• interviewer bias
• diagnostic bias

Question 32

Where cases and controls equally mis-report their exposures, this would more likely be described as:

Answer 32

Non differential misclassification

Question 33

Where cases and controls unequally mis-report their exposures AND in such a way that the overall consequence is an association that tends away from the null, this is best described as:

Answer 33

Information bias

Question 34

what is confounding?

Answer 34

the effect of an extraneous variable or wholly or partially accounts for the apparent effect of the study exposure, or that masks an underlying true association
e.g. birth order being associated with a baby having down syndrome is actually masking the increasing age of mothers when they give birth

Question 35

4 ways to identify confounders

Answer 35

1. Knowledge of the subject matter: you could undertake an evidence review and see what other people have proposed or found.
2. Test by the three criteria: associated with exposure in the source population; associated with the outcome in the absence of the exposure; not a consequence of the exposure.
3. Stratification: we can look at the difference of apparent effect within different population strata.
4. Compare the crude and adjust statistical estimates.

Question 36

what is effect modification?

Answer 36

when the strength of the association varies over different levels of a third variable

Question 37

should you control effect modification/

Answer 37

no - like confounding it’s naturally occurring

- way results reported should take it into account

Question 38

tests for effect modification

Answer 38

• Breslow-Day test
• Q test
• Interaction terms in regression models

Question 39

what are the two critical points for why differentiation between cofounding and effect modification are important?

Answer 39

1. Addressing a confounded relationship by addressing the exposure exclusively is very unlikely to yield a gain. In example (3), stopping a mother from smoking is unlikely to reverse the childhood mental health outcome.
2. Addressing an exposure where effect modification is apparent may be useful: it may be possible to target an intervention into a more homogeneous pool of participants where a greater impact will be yielded

Question 40

what is a highly effective way for us to identify possible confounding.

Answer 40

Multivariate regression