Epidemiology Flashcards
1
Q
What is Epidemiology?
A
the study of the distribution and determinants of health-related states/events in specified populations, and the application of this study to the control of health problems
2
Q
What is a punitive exposures in the context of population health?
A
The variable being tried to associate with a change in health status
3
Q
What is outcomes in the context of population health?
A
The associated change in health status
4
Q
What is a primary population level intervention?
A
Preventing disease through control of exposure to risk factors
5
Q
What is a secondary population level intervention?
A
The application of available measures to detect early departures from health and to introduce appropriate treatment and intervention (slows progression)
6
Q
What is a tertiary population level intervention?
A
The application of measures to reduce/eliminate long term impairments and disabilities, minimising suffering caused by existing departures from good health and to promote the patients adjustments to their condition
7
Q
Define a statistic
A
A fixed value, derived from a sample that estimates the value in the population
8
Q
Define a parameter
A
A fixed, often unknown value, which describes an entire population
9
Q
Define a point estimate
A
Estimate of a parameter
10
Q
Define a confidence interval
A
The range of values in which there is a 95% confidence that true value lies within these
11
Q
Define effect modification
A
Exists when the strength of an association varies over different levels of a third variable
12
Q
What should happen when effect modification is detected?
A
conduct stratified analysis
13
Q
How do you test for effect modification?
A
- Breslow-Day test
- Q test
- interaction terms in regression
14
Q
What is synergism?
A
Effect modifier potentiates exposure effect
15
Q
What is an antagonism?
A
Effect modifier diminishes exposure effect
16
Q
Define confounding variable
A
A third variable which leads to bias in the estimate of association between outcome and exposure
17
Q
Define confounding
A
The effect of an extraneous variable that wholly/partially accounts for the apparent effect of the study exposure or masks an underlying true association
18
Q
Define internal validity
A
When association truly exists within study participants
19
Q
Define external validity/generalisability
A
When association observed within a study can be extended to the wider population
20
Q
Define bias
A
Any trend in the collection /analysis/interpretation/publication/review of data that can lead to conclusions that are systemically different from the truth
21
Q
Define correlation
A
A linear relationship
22
Q
Define association
A
Exposure A and exposure B follow similar trends
23
Q
Define causation
A
Exposure A leads to outcome B
24
Q
What are the 2 different types of error?
A
- Type 1
- Type 2
25
What is a type 1 error?
False positive finding
26
What is a type 2 error?
False negative finding
27
Define selection bias
When participation in a study is associated with the exposure and the outcome
28
What are the variations of selection bias?
- Berkson’s bias
| - healthy worker effect
29
What is Berkson’s bias?
Hospital based case control study with controls selected among the hospitals patients
30
What is the healthy worker effect?
Active workers are more likely to be healthy than those who have stopped working
31
How can selection bias be minimised?
- controls representative of the target population
- minimise the non-response of patients
- compare respondents to non-respondants
32
Define information bias
- misclassification of the exposure and/or outcome
| - due to I’ll-defined variables/flaws in data collection
33
What are possible flaws in data collection?
- interviewer bias
| - recall bias
34
Define interviewer bias
Prevented by interviewer being blind/strict protocol when collecting data
35
Define recall bias
Bias when participants self report exposures between cases and controls
36
How can you prevent recall bias?
Use objective ways to assess exposure
37
What are the different types of misclassification?
- non-differential
| - differential
38
What is non-differential misclassification?
When misclassification occurs equally across all groups of study
39
What does a non-differential misclassification result in?
Bias towards the null hypothesis
40
What is a differential misclassification?
When misclassification occurs disproportionately towards cases or controls
41
What domes a differential misclassification result in?
Bias towards or away from the null
42
What is hazard ratio?
Using arbitrary time points to see if trial participants are alive - loss of statistical precision
43
What is an alternative to a hazards ratio?
Survival analysis
44
What is survival analysis?
Whenever a patient dies, their death will be recorded at that time and a survival time will be calculated (range of survival times)
45
What is a Kaplan Meier plot?
Displays survival analysis
46
What does a Kaplan Meier chart show?
The proportion of participants alive at any particular time
47
Evaluate primary data
+ collected for the purpose
- time consuming
- expensive
48
Evaluate secondary data
+ faster
+ cheaper
- have to make assumptions as data wasn’t collected for the new purpose
49
What is data linkage?
Joining 2 or more data sets together
50
Evaluate data linkage
+ find out more than from each data set individually
- technical issues
- privacy concerns
51
What are examples of non-routinely collected data?
- surveys
| - bespoke data sets
52
Evaluate non-routinely collected data
- limited use
- expensive
- time consuming
53
What are examples of routinely collected data?
- census
- electoral register
- first language of reception children
- hospital episode statistics
- prescribing data
54
What is a narrative review?
Puts published literature into a single article
| Literature review/scoping review/non-systematic review
55
Evaluate narrative reviews
+ easier
+ faster
+ good when starting research
+ good when there is a lot of variation in research
- bias, works may be unbalanced/evidence could be missed by chance
56
What is a systematic review?
Highly structured approach to searching + including + summarising the literature
- basis for meta-analysis
57
What is the process involved in a systematic review?
- research question
- structured search
- indices and registries
- screening (PRISMA flow diagram)
- reporting
- writing
- submission
58
Evaluate systematic reviews
+ aims to collate all available evidence
+ specific protocol and inclusion criteria allows for reproducibility
- only as good as the method/indices/evidence used
- very quickly out of date as time consuming
59
Define endpoint
An outcome that is usually clinically meaningful
60
Define efficacy
How well a therapy works in achieving a desired outcome
61
Define primary endpoint
The endpoint for which the study has been powered
62
Define secondary endpoint
A slightly different endpoint to the primary that is examined as well (may be proven even if primary is not)
63
Define safety
How well a therapy works in not causing adverse effects
64
What must be taken into account when assessing safety?
Judge whether the safety profile is offset by the efficacy
65
Define phase composite endpoint
- Multiple potential endpoints added together
- when the outcome is uncommon
- eg: MI + ischaemic stroke = cardiovascular event
66
Define descriptive epidemiology
Describes the problem (often aggregate level) and can be used to inform analytic research
67
What are some examples of descriptive epidemiology?
- case report
- case series
- cross-sectional
- longitudinal
- ecological
68
What is a case report?
Used for new diseases/presentations/findings
69
What is a case series?
Multiple case reports
70
What is a cross-sectional study?
Describes the prevalence of an exposure/outcome across a population at a single point in time
(No follow up and only prevalence)
71
What is a longitudinal study?
describes the prevalence/incidence of an exposure/outcome over time (can be aggregate/person-level)
72
What is an ecological study?
compare groups rather than individuals (can be longitudinal/corss-sectional)
73
Define analytic epidemiology
- deploy and test hypotheses
| - allows to measure association and infer causation
74
What are the different types of analytic epidemiology?
- observational
| - experimental
75
What is an example of experimental analytic epidemiology?
randomised controlled trials
76
What is an example of an observational analytic epidemiology?
- case control studies
| - prospective cohort studies
77
Define detection bias
differences between groups in how outcomes are determined
78
Define performance bias
differences between tested groups in either provided care or exposure to factors not being investigated
79
What are the advantages of double blind trials?
prevents:
+ performance bias
+ detection bias
80
What does including blinding in a study prevent?
prevents or reduces withdrawal from study
81
When is blinding not possible?
- ethically
- if surgery is involved
- if the drug needs titrating
82
What do randomised control trials do?
evaluate the impact of an intervention on an outcome
83
What is included in a randomised control trial?
- tests 2 or more treatments to see which is better
- always need controls (existing treatment/placebo)
- include all patients in analysis irrelevant if they followed treatment
84
In the context of odds, what is the output number above which it becomes more likely an event takes place than does not take place?
1.0
85
Define prevalence
the proportion of individuals in a population who have the attribute at a specific timepoint
86
What is the epidemiological definition of cumulative incidence?
The proportion of the population with a new event during a given time period.
87
How do you calculate cumulative incidence?
Cumulative incidence = number of those with the incident / those susceptible to the incident
88
How is incidence rate calculated?
The count of new cases during the follow-up period, divided by the total person-time.
89
What are the limitations of a crude rate?
It is adjusted for the size of the population (count divided by persons), but not for the age or sex variation within that population – if it were, a standardisation approach would have been used. Crude is sometimes also termed ‘raw’ or ‘unadjusted’.
90
What is a confidence interval?
The range of values within which we are 95% confident the true value lies.
91
Describe the relationship between the size of the confidence interval and the number of observations.
inversely proportional
92
What characterises a historical cohort study?
both the exposures and outcomes are known at the start of the study
93
What is a major limitation of observational study designs?
The observed groups may differ in characteristics other than the variable of interest
94
What is an intention to treat analysis?
The process of statistically analysing patients’ outcomes using the original groups to which they were allocated, irrespective of whether they took the medicine or not
95
Describe a single blind trial?
A trial where the participants (the patients) are not aware which arm of the trial they are in (intervention vs. control).
96
Describe a double blind trial?
A trial where the participants AND attending clinicians are not aware which arm of the trial participants are in.
97
Describe a triple blind trial?
A trial where the participants / patients AND attending clinicians AND analysis team are not aware which arm of the trial participants are in.
98
What is the intention of allocation concealment?
The step in randomisation that attempts to prevent persons involved in the trial from knowing the allocation of participants to study arms.
99
Define the statistical power of the study
The ability to detect a difference if a difference exists.
100
What are the 3 main factors that impact sample size?
- power of at least 80%
- low alpha (0.01)
- the difference in arms (bigger = smaller sample)
101
Define internal validity
The extent to which findings accurately describe the relationship between exposure and outcome in the context of the study
102
Define information bias
Misclassification of the exposure, outcome or both.
103
What are different forms of information bias?
- recall bias
- interviewer bias
- response bias
- diagnostic bias
104
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:
information bias
105
What is an unwarranted variation?
variation from the norm or the expected, without any explanation - RED FLAG!!
106
What is an explained variation?
a variation from the norm that can be explained
107
What is a statistical artefact?
when the variation is due to differences in recording or reporting
108
Define ecological fallacy/aggregation bias?
the assumption that associations between groups hold between individuals
109
Evaluate ecological studies
+ useful when there is limited variability in the group (more confidence asssociations)
+ first step and can be used to explore hypotheses
- uses secondary data
- unsure if exposure preceded outcome
110
What is the implication of a difference of interest?
the smaller the difference of interest, the bigger the sample size
111
Define power in epidemiology?
the ability to detect an effect/association if one exists (minimum 80, ideally 90)
112
What can cause an increase in power?
an increase in sample size
113
What does a 90% power mean?
10% chance that the study won't detect an effect if it exists
114
Define alpha
how much chance can be ruled out to cause a positive finding
115
What is alpha?
specifying p-value and connect to 1 and 2 tail testing
116
What is the impact of decreasing alpha from 0.05 to 0.01?
increasing sample size
117
What are the disadvantages of using p values?
- if multiple analyses are run, one is likely to come back as less than 0.05 despite being false
- mistaken for clinical significance
118
What is a meta-analysis?
combines the quantitive findings of separate studies into a pooled estimate of association
119
What is a forest plot?
the way that meta-analysis is presented
120
Define publication bias?
studies with more positive findings are more likely to be submitted/published
121
What can be used to assess for publication bias?
publication funnel plot
122
What are the differences in studies that can be combined in a meta-analysis?
- patient/selection criteria
- study design/blinding/intervention approach
- reporting differences
- whether the effects are random or fixed
123
How could a population with an upward pointing triangle shaped population be described?
High levels of infant mortality and infectious diseases
124
What is the most appropriate method to compare age-specific data between 2 different populations?
Direct Standardisation
125
How does the UN see global health?
As critical to ensuring international order, peace, stability and security
