Epidemiology: Interpreting Findings

Question 1

Define narrative review

Answer 1

Brings together published literature into a single article allowing reader to rapidly understand the topic. Also known as critical review, scoping review or non-systematic review.

Question 2

Define systematic review

Answer 2

Sets out a highly structured approach to searching, sifting, including and summarising the literature. Underpins meta-analysis.

Question 3

What are the advantages and disadvantages of a narrative review?

Answer 3

Adv: Agile so easier and faster to write and hence may be more up to date, allows with rapid understanding of a topic so may be suited to those with limited research or greater variation in research methods, also useful when work from different disciplines is brought together to answer a less-easily answered question.

Disadv: Room for potential bias as authors are free to select included works although robust peer review should prevent overly-speculative or unbalanced reviews,

Question 4

What are the advantages and disadvantages of a systematic review?

Answer 4

Adv: Collates all available evidence for a focused research question, implements highly specified protocol ensuring reproducibility, only includes evidence based on an inclusion criteria

Disadv: Only as good as method employed (if search not comprehensive, results will not be either), only as good as indices used, becomes out of date quickly due to the sheer amount of time process takes (18-24 months)

Question 5

What are indices and registries?

Answer 5

Indices publish articles based on research that has already been conducted while registries contain information on research that is ongoing or hasn’t begun yet. Researchers register their work in registries to prevent duplication of research and publication bias which is underreporting of negative results and overreporting of positive results.

Question 6

What is a PRISMA diagram?

Answer 6

Shows how many research papers initially identified then how many duplicates were removed, how many papers were screened and excluded through abstracts and finally full text analysis. Shows number of papers at each step.

Question 7

What is grey literature?

Answer 7

Reports eg from governments about implementations and programmes. However, these are not always peer reviewed and hence have varying levels of accuracy.

Question 8

Define meta-analysis

Answer 8

Meta-analysis is a quantitative, formal, epidemiological study design used to systematically assess previous research studies to derive conclusions about that body of research.

Question 9

What are the sources of heterogeneity in meta analysis?

Answer 9

1. Clinical: different patient groups, populations and selection criteria used by studies
2. Methodological: different study designs (blinding, collecting info, recording info)
3. Statistical: reporting differences
   There are statistical ways of evaluating this heterogeneity between studies

Question 10

What are other considerations when weighing meta analysis?

Answer 10

1. Weighting of study usually depends on study size but we can build in assumptions to account for other similarities and differences. Whether a fixed effects model is used or a random effects model is used for these assumptions will affect the overall weighting of the study.
2. Publication bias - using a publication funnel plot can help show balance of evidence assuming an overall effect size

Question 11

Define endpoint

Answer 11

An endpoint is an outcome that usually describes a clinically meaningful outcome.

Question 12

What are the 2 clinical trial outcomes?

Answer 12

Efficacy – how well a therapy works in achieving a desired outcome.
Safety – how well a therapy works in not causing adverse events.

Question 13

What are the 2 types of efficacy endpoints?

Answer 13

Primary endpoint – this is the endpoint for which the study has been powered; that is to say that the number of trial participants (sample size) will have been recruited on the basis of the pre-specified power and difference.
Secondary endpoint – it is common that a study will want to examine a slightly different endpoint in addition to the primary endpoint. For example, while a study seeks to examine survival (i.e. alive or dead) another – often ‘softer’ - measure such as recurrence of disease or hospital admission might also be measured. If the secondary endpoint is proven but the primary endpoint is not, then the findings of the study may still contribute to the understanding of disease.

Question 14

What is survival analysis?

Answer 14

Arbitrary timepoints lead to loss of statistical precision as effects observed may not be the same at different timepoints. Hence, a survival analysis is used which tracks participants across a time period and records when certain events take place.

Question 15

What is the Bradford Hill criteria?

Answer 15

Enables us to infer causation from both observational and interventional methods. Includes:
Strength: stronger association increases confidence that an exposure caused an outcome
Specificity: Informative when present but its absence doesn’t rule out causality
Temporality: exposure must preceed outcome
Biological response: clear response effect is an indicator of causality
Should be biologically plausible
Coherence: association ought to be consistent with existing theories and knowledge
Evidence from experimentation should support association
Analogy

Question 16

Define internal validity

Answer 16

Where findings are valid and applicable to the study population

Question 17

Define external validity

Answer 17

Where findings of a study can be applicable to others apart from the source population. Also known as generalisability. Internal validity is a prerequisite.

Question 18

Define bias

Answer 18

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 19

Define selection bias

Answer 19

Selection bias occurs when participation is related to both the exposure and the outcome, alterring the magnitude of the association between the exposure and the outcome.

Question 20

What are 3 methods of reducing selection bias?

Answer 20

1. Make sure to select controls representative of the target population
2. Minimise non-response - the more people that drop out, the greater the chance bias is introduced
3. Compare respondents with non-respondents and see if there are any systematic differences between the

Question 21

What is misclassification?

Answer 21

A form of information bias where participants are wrongly assigned to either exposure or outcome. More likely when study variables are not properly defined or data collection done poorly.

Question 22

What is recall bias?

Answer 22

Those suffering from a disease are more likely to recollect events that could have influenced their disease as compared to those without the disease who may have forgotten these details. Can be prevented by using objective methods of assessment.

Question 23

What is non-differential misclassification?

Answer 23

Where there are errors with classification among both exposed and unexposed populations or those with a particular outcome and those without. This results in bias towards the null hypothesis.