Week 1 Flashcards Preview

HSPH > Week 1 > Flashcards

Flashcards in Week 1 Deck (92):
1

What is a cohort study

Take a group of individuals and follows them forward in time.
Allows studying of risk factor for disease outcome

2

Name 5 advantages of a cohort study

1- Does not require random assignment
2- Provides clearer temporal sequence of exposure and disease
3- Gives opportunity to study multiple outcomes of related exposure
4- Allows calculation of relative and absolute risk
5- Enables study of rare exposure

3

Name 6 disadvantages of a cohort study

1- Requires a large sample size
2- Requires a long time to see outcome
3- There can be a loss of subjects
4- Not suited for study of rare diseases
5- Expensive to carry out large numbers of subjects needed
6- Exposure patterns may change during study

4

Selection criteria for cohort study (3)

- Study group must be representative of population studied
- Must be healthy individuals with no predisposing factors to outcome
- Must be randomly selected

5

What is a case controlled study

Compare characteristics of a groups of patients with a particular disease outcome to a group of individuals with disease outcome

6

Name 4 advantages of a case controlled study

1- Permit study of rare disease
2- Permit study of disease with long latency periods between exposure and manifestation
3- Relatively inexpensive
4- Can study multiple potential causes

7

Name 4 disadvantages of a case controlled study

1- Information exposure and past history subject to recall bias
2- Validation of information on exposure difficult
3- Concerned with only 1 disease
4- Choice of control group

8

Give an example of binary data

Alive or dead

9

Give an example of ordinal data

Pain score

10

Give an example of nominal data

Occupation

11

Give an example of continuous data

Weight

12

Define crude death rate

Total death in year/ mid age population

13

Define infant mortality rate

No of deaths in first year of life/ no of babies born alive or dead that year

14

How do you calculate standardised mortality rate

(Observed deaths/ expected deaths) * 100

15

Define age adjusted death rate

Statistical method allowing the comparison of communities with different age structures

16

Define numbers needed to treat

The number of people who need to take a treatment in order for one person to benefit from it

17

Define incidence

Number of new cases/ deaths of a disease per 100,000 people per year

18

Define prevalence

Amount of patients with a disease in a given time

19

How do you calculate risk of death in treatment group

NUmber of deaths in treatment group/ number of patients in treatment group

20

How do you calculate risk of death in control group

Number of deaths in control group/ number of patients in control group

21

How do you calculate risk of death in treatment group

Number of deaths in treatment group/ number of patients in treatment group

22

How do you calculate relative risk

Risk of death in treatment group/ risk of death in control group

23

What is standard error

The number of standard deviations from the mean of all potential samples of the same size

24

What calculation is used to work out confidence internvals

Standard error

25

In a normal distribution, what % of values are 1 SD above/ below the mean

68

26

In a normal distribution, what % of values are 1.64 SD above/ below the mean

90

27

In a normal distribution, what % of values are 1.96SD above/ below the mean

95

28

What is a p value

A value used to determine whether a difference observed is due to change

29

What does a 95% confidence interval=

mean +/- 1.96 * standard error

30

How does the size of the sample size affect the confidence interval

Larger sample size leads to narrower confidence interval

31

What is the relative risk if the treatment has no effect

1

32

Who gets treatment in an uncontrolled trial

Everybody

33

Who gets treatment in a controlled trial

1 group, 1 placebo

34

What is a randomised control trial

Allocation to group is determined by chance

35

What is meant by a geographical control

Treated at a different hospital where new intervention is not provided

36

What are 2 potential threats to validity of results

Chance and bias

37

Why/ how does chance threaten the validity of results

Chance occurs due to random variation
Can lead to imprecise results

38

How can you reduce the risk of chance impacting validity

Use large sample size
Calculate p values

39

What is selection bias and how can it be avoided

How subjects are recruited to treatment/ control group
Random allocation/ selection used to avoid this

40

What is measurement bias and how can it be avoided

Inaccurate measurement of outcomes such as inaccuracy in measuring instrument or bias in expectation
Can be avoided by binding participants and researchers

41

How is analysis bias avoided

Reduced by maximising follow up and carrying out intention to treat analysis

42

When does survival bias occur

When incidence is similar in all groups but not all groups survive long enough to be analysed

43

What is meant by single and double binding

Single- participant doesn't know treatment
Double- neither participant or researcher know

44

What a parallel group randomised control trial

Used when effect of treatment is irreversible

45

When are crossover groups used in randomised control trial

Used when effect of treatment is reversible

46

What are the advantages of crossover groups (3)

Each patient is their own control
Smaller sample size produces same number of observations as a larger one
Better for subjective measurement

47

What are the disadvantages of crossover groups (2)

More time consuming
Carry over effect

48

What are clustered randomised trials and why are they used

Groups of subjects are randomised as opposed to individuals
Avoids contamination across individuals

49

What is meant by a factorial

2 or more experimental interventions are not only evaluated separated but also in combination and against a control

50

What is the aim of a factorial

Allows evaluation of the interaction that may exist between 2 treatments

51

What are the 4 boxes on a factorial square

- Data on patients receiving none of the treatments
- Patients who received A
- Patients who received B
- Patients who received both

52

What is a placebo

Substance that has no therapeutic effect

53

What is an intention to treat analysis

Comparison of all subjects based upon treatment group assigned, regardless of whether they complied

54

What are intention to treat analyses used for

Assessment of clinical effectiveness, because they mirror actual practice where people don't comply

55

What is on treatment analysis

Comparison of subjects who actually received treatment

56

What is meant by loss to follow up

Patients who at one point where actively participating, but have become lost at the point of follow up in the trial

57

What is meant by compliance

Degree of constancy and accuracy with which a patient follows a prescribed regemin

58

What is the consequence of poor compliance in intention to treat analysis

Reduces ability to detect differences in treatment

59

How to maximise compliance

Selection of patients
Double blind
Run in period where all get treatment

60

What are meta-analyses

Contrasting and combining results from different studies, in the hope of identifying pattern among study results, sources of disagreement, or any other interesting relationships

61

What are the disadvantages of meta-analyses

- Heterogeneity
- Publication bias

62

What is heterogenity

Different studies not undertaken in exact same way leading to varied results

63

What is publication bias

Published studies may not be truly representative of all studies undertaken or research is undergone by somebody with a financial incentive

64

What is meant by numbers needed to treat (NNT)

Average number of patients who need to be treated to prevent one additional bad outcome

65

Describe the structure of a cohort study

Individuals identified cohort--> measur exposure in each individualsa--> follow up to determine disease occurence

66

What is a concurrent cohort study

Record any previous exposures in cohort then follow them up (disease not already occurred)

67

What is a historical cohort study

Data concerning exposure and occurence of a disease collected after the event has taken place and subjects recruited according to existing health care records

68

What are 3 examples of bias sometimes seen in cohort study

Loss to follow up
Exposure usually measured at one point in time
Selection bias

69

What is absolute excess risk

Risk is exposed- risk in unexposed

70

Define attributable proportion

Measure of the proportion of the total risk which can be attributed to the risk factor in question= incidence in population due to exposure

71

How do you calculate attributable proportion where p= proportion exposed in population

p(relative risk-1)/ 1 + p(relative risk-1)

72

What are survival analysis

Looking at how long someone lives

73

What does a kaplan meier curve plot

proportion of people surviving over time

74

What does hazards ratio give

Risk of dying at any point in time in one group compared to the other

75

What is a confounder

A factor associated with both the exposure and the disease

76

What is an unadjusted results

Shows relationship between risk factor and an outcome. Useful for deciding if certain individuals at risk

77

What are adjusted results

Show the effect that would occur if we changed that risk factor but all others remained the same. Used to investigate causality

78

What is a nested case control trial

Case and controls are selected from individuals with an established cohort study

79

What 3 kinds of biases may be seen in case control trials

- Recall bias
- Reverse causality
- Selection bias

80

What is recall bias

Cases may remember more than controls

81

What is reverse causality

Effects precede the cause (don't know what has caused what)

82

What is the odds ratio equal to in rare diseases

Relative risks

83

What are cross sectional studies

Data collected at one point in time, relationship between characteristics are considered. Describes the characteristics of a population

84

What is the limitation of a cross sectional study

Can't establish cause
Not useful in rare diseases

85

What study types provide the strongest evidence for causality

clinical trials and cohort studies

86

What are confounders

Factor associated with both the exposure and the disease

87

What are some real reasons for association between 2 variables

Co-incidence
Reverse causality
Caudal
Confounding

88

What are some not real reasons for an association between 2 variables

Bias
Incorrect analysis

89

What is a causal relationship

One where the cause directly affects the outcome of interest

90

What is the name of the 9 criteria required to prove causality

Bradford Hill criteria

91

Name and explain the first 4 criteria in the Bradford Hill criteria for causality

- Strength of association between risk factor and outcome measured by OR or RR
- Consistency- have the findings be observed by numerous studies
- Reversibility- if you stop exposure does this revert risk to 1
- Temporality- does exposure occur a reasonable time before the disease

92

Name and explain the final 4 criteria in the Bradford Hill criteria for causality

- Plausibility: is the result biologically plausible
- Coherence- are results consistent with other study types
- Specificity- is the relationship with the exposure limited to specific diseases
- Dose response: would risk increase with increased exposure