Evidence based medicine, association and causation

Question 1

What is EBM

Answer 1

The conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients

Methods to critically appraise clinical information and classify it according to the strength of evidence

Question 2

Why does EBM matter to clinicians?

Answer 2

1. Patient care
2. Patient safety
3. Medical Knowledge
4. Professionalism
   - Does not replace clinical decision making
   - Only a tool
   - Some of the evidence not generalizable to specific cases
   - So much evidence

Clinical findings:
- Gather and interpret findings from history and physical examination

Aetiology:
- How to identify causes for disease

Clinical manifestation of disease:
- Know how often and when disease cause manifestations

Differential diagnosis:
- How to select which causes are more common

Question 3

What are the applications of EBM

Answer 3

Diagnostic test: appropriate one (accuracy, acceptability, expense, safety)
Prognosis: estimate likely clinical course and anticipate complications
Therapy: select best treatment worth effort
Prevention reduce chance of disease by identifying and modifying risk factors

Question 4

What is an association?

Answer 4

statistical dependence between two variables, that is the degree to which the rate of disease in persons with a specific exposure is either higher or lower than the rate of disease without that exposure.

- Link
- Relationship 
- Correlation

Question 5

What is might cause association due to chance?

Answer 5

Sample size
Power calculations
P values and statistical significance

Question 6

What are confidence intervals?

Answer 6

• range within which true value should lie

- Certain percentage of intervals will contain true underlying population parameter

Question 7

What is a p value?

Answer 7

• probability result could be chance

- Threshold usually less than 0.05 (can be pretty sure it is not due to chance)

Question 8

What is bias?

Answer 8

• systematic error leading to incorrect estimate of effect on outcome of interest
• Observed effect either above or below true value
• Consequence of defects in design or execution
• Cannot be controlled in analysis or eliminated by sample size
• Can be either selection or measurement

Question 9

What are examples of biases in selection?

Answer 9

• systematic difference between characteristic of people selected for study and those who were not
• Non response bias
• Healthy entrant effect
• Loss of follow up (attrition bias)
• Use random selection or use of entire population

Question 10

What are examples of biases in measurement?

Answer 10

• measurements or classifications of disease or exposure inaccurate
• Recall bias
• Make sure equipment is calibrated same

Question 11

What is a counfounder?

Answer 11

• any factor believed to have real effect on disease risk under investigation but also related to risk factor under disease
• Factors that may have direct causal link with disease
• Factors that are good proxy measure of more direct unknown cause
• Account for confounding using matching, randomization, stratification, multivariate analysis

Question 12

What are common confounders?

Answer 12

• Age
• Sex
• Socio-economic status
• Geography

Question 13

What is the hierarchy of study design?

Answer 13

see notes

Question 14

What are possible faults in case control trials?

Answer 14

• Compare against placebo (instead of best available literature)
• Inappropriate control (too little or too much of current drug given for it to be effective)
• Missing trials (withhold information so cannot know true effect size)

Question 15

Define causation?

Answer 15

judgement based on chain of logic addressing two main areas

1. Observed association between exposure and a disease is valid
2. Totality of evidence taken from a number of sources supports judgement of causality

Question 16

What is the Bradford-Hill criteria

Answer 16

• Temporal relationship (essential)
• Plausibility - is exposure plausible
• Consistency with other investigations - particularly if use other methods
• Strength of the association - being strong gives you more reason to think relationship is causal
• Dose-response relationship
• Risk of disease associated with increased exposure to risk factor
• Specificity - specific cause
• Experimental evidence - pretty strong (trials on animals or people)
  ○ Can’t really do experiments on bad exposure
• Coherence
  ○ With current views of conditions or its causes
• Analogy

also consider reversibility (taking away exposure would decrease risk of disease)

Question 17

Demonstrate epidemiological principles applied to clinical decision making

Answer 17

see notes