Bias

Question 1

Bias definition

Answer 1

A systematic error that results in a deviations from the true effect of an exposure and on the outcome

Question 2

Types of bias

Answer 2

Measurement bias = different equipment measuring differently
Observer bias = observes expectations influence reporting
Recall bias = past events not recalled correctly
Reporting bias = people don’t tell the truth because of shame/judgement
Selection bias = bias in recruiting for a study and some may be lost to follow up
Publication bias = trials with negative results less likely to be published

Question 3

Bradford-Hill Criteria: Causality

Answer 3

• Strength: Stronger association between exposure and outcome
• Consistency: Same result from various studies
• Dose response: Increased risk of outcome with increased exposure
• Temporality: Exposure must precede the outcome
• Plausibility: Reasonable biological mechanism
• Reversibility: Intervention to reduce outcome
• Coherence: Logical consistency with other info
• Analogy: Similarity with other established cause effect
• Specificity: Relationship specific to outcome of interest

Question 4

Con founders

Answer 4

A situation in which the estimate between an exposure and an outcome is distorted because of the association of the exposure with another factor (confounder) that is also independently associated with the outcome.

Question 5

Causes of association Important
- Why would you get results that suggest that exposure influences the outcome?

Answer 5

• Bias
• Confounding factors
• Chance
• Reverse causality
• TRUE ASSOCIATION = CONFIRMED BY MEETING THE BRADFORD-HILL CRITERIA

Question 6

Epidemiology definition Important

Answer 6

The study of the frequency, determinants and distribution of diseases and health related states in populations in order to prevent and control disease

Question 7

Incidence definition Important

Answer 7

Number of new cases over a certain time period

Question 8

Prevalence definition Important

Answer 8

The number of people with a disease at a certain point in time

Question 9

Person time definition Important

Answer 9

Measure of time at risk of all the patients in the study - therefore if 1000 patients where studied for 2.5 years, the study would have looked at 2500 person Yeats

Question 10

Risk definition Important

Answer 10

Number of new cases/ number of people at risk of the disease (within a given time frame)

Question 11

Relative risk definition Important

Answer 11

Risk among exposed group divided by risk in unexposed group - doesn’t take into account baseline risk

Question 12

Absolute/ Attributable risk

Answer 12

Works out how much of the event (e.g. disease) occurs BECAUSE of the exposure.
Substract the risk of the control group from the exposed group = gives you the excess risk caused by the exposure

Question 13

Number needed to treat

Answer 13

The number of patients you need to treat for one to benefit.
1 divided by absolute risk

Question 14

Relative risk and attributable risk calculations
*GO TO POWERPOINT FOR TABLE PAGE 68)