Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Epidemiology > Introduction to epidemiology > Flashcards

Introduction to epidemiology Flashcards

1
Q

What are the three elements involved in a research question?

A

Determinant, domain, outcome

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Describe the formula of the occurrence relation

A

Outcome = f(Determinant(s) | confounding)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the DEPTH model?

A

What’s wrong with patient?: Diagnostic knowledge

Why is this patient ill?: Etiologic knowledge

What’s will happen if I don’t intervene?: Prognostic knowledge

What will be the effect of an intervention? Therapeutic (prognostic + etiologic) knowledge

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is changed in the present epidemiology compared to the history of epidemiology?

A
  • Strong effects between single determinants and disease occurrence (smoking and lung cancer etc.) are largely established
  • More attention to chronic diseases
  • Shift towards research into combined causes
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What are the frequency of measures?

A
  • Prevalence
  • Cumulative incidence
  • Incidence density
  • Odds
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are the association measures?

A
  • Relative risk
  • Relative rate
  • Odds ratio
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the impact measures?

A
  • Attributable risk
  • Population attributable risk
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is the definition of prevalence?

A

Proportion of a population affected by the disease/health status at a given point in time, expressed as a percentage.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How do you calculate the prevalence? Give the formula.

A

(Number of cases of disease OR health status at a given time ) / (Population (at risk)at that given time) * 100%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

With which measure of frequency do you describe the “disease burden”?

A

Prevalence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is the definition of cumulative incidence?

A

Probability (0 to 1) of getting the outcome in a population at risk in a set period of time.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

How do you calculate the cumulative incidence? Give the formula.

A

(Number of new cases of disease OR health status in a set period of time) / (In a population at risk in that set period of time ) * 100%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Why are cumulative risks a crude (naïve) estimation?

A

It is an underestimation, you don’t take competing risks into account.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

True or false? Risks increase with the length of the risk period.

A

True

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

True or false? You don’t need to know the length of time over which the risk applies to interpret a risk.

A

False! The only way to interpret a risk is to know the length of time over which the risk applies.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Give de definition of incidence density.

A

The number of occasions (between 0 and infinity) of disease onset in the population divided by the time period of observation in new cases per person years of population at risk. Time specification is essential!

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

How do you calculate the incidence density? Give the formula.

A

Incidence rate/density = A/Time = (Number of subjects developing a disease) / (Total time experience for the subjects followed) = A/person time = 1/A * person time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Give de definition of incidence

A

Number of new cases of disease/health status in a set period of time.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is more easy / intuitively to use? Cumulative incidence or incidence density?

A

Cumulative incidence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

True or false? Over a small period of time CI = ID

A

True, over small time periods incidence is always low

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is more ideal for long follow-up times? Incidence density or cumulative incidence?

A

Incidence density, because of competing risks.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What is odds?

A

(Probability that an event will happen) / (Probability that an event will not happen)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

When to use odds ratio’s?

A
  • Sometimes used instead of real risks because statistical models (Logistic regression)
  • Exposure odds used in designs in which risks cannot be directly estimated (case control study)
  • But, use real risks where you can…..!
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

How do you make a 2x2 table?

A

Outcome in horizontal line with yes (outcome) on the left. Determinant on vertical line and yes (interest/determinant) on top.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

How do you calculate the risk ratio / relative risk?

A

using Cumulative Incidence (CI) = (a/(a+b))/(c/(c+d)) (95% CI: xx-xx)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is the 95% Confidence interval?

A

The 95% confidence interval (95% CI) show the spread of 95% of all observations. It is more narrow the confidence interval is more precisely.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

How to calculate the rate ratio?

A

Rate Ratio = (a/(person years exposed)) / (c/(person years unexposed))

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

Risk ratio or rate ratio? In short follow-up, long follow-up, etiological research?

A
  • For short follow-up, easy interpretation of risk ratio
  • For long follow-up: risk ratio suffers from limitations of CI estimation
  • For etiological research rate ratio is often preferred
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

What is the risk difference? What is the formulate to calculate it?

A

Risk difference: additional cases per xx persons compared to healthy/not exposed persons.
= a/(a+b)-c/(c+d) = xx additional cases per xx subjects

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

What is the rate difference? What is the formula to calculate it?

A

additional cases during xx time of follow-up compared to healthy/not exposed persons.
= a/(person years exposed) - c(person years exposed) = xx additional cases during xx years of follow-up

31
Q

What is the attributable risk? What is the formula?

A

relative measure of risk difference, fraction of risk in exposed that is attributable to the exposure itself.

AR = (Risk exposed - Risk unexposed) / (Risk exposed) = (a / (a+b) - c/ (c+d)) / (a / (a+b))

32
Q

What is the population attributable risk? What is the formula?

A

how much disease in the total population is attributable to exposure. This could be helpful to calculate in how many people you prevent the outcome when take away the determinant.

PAR = (Risk total population - Risk unexposed)/(Risk total population)

=

((a+c)/(b+d)-c/((c+d) ))/((a+c)/(b+d))

33
Q

What do you have to assume if you interpret the attributable fraction?

A

To interpret the attributable faction you have to assume that there is a causal relation between low antibody levels and diarrhea and the effect measures are unbiased.

34
Q

True or false? Both attributable risk and population attributable risk are also called etiologic fraction

A

True

35
Q

True or false? Incidence proportion = Cumulative incidence | Incidence rate = Incidence density

A

True

36
Q

What is bias? Which kind of bias does excist?

A

Bias is any systematic error resulting in an incorrect estimate of the association between exposure and the outcome, i.e the estimated measure of association (risk ratio, rate ratio, odds ratio, difference in means, etc.) deviates from the true measure of association. There two kinds of biases; confounding and other biases (information and selection).

37
Q

What are selection and information bias, and how do you deal with it?

A
  • Selection bias: there is an issue in how samples are selected. The (control) sample doesn’t reflect the source population.
  • Information bias: information is not gathered in the same way for every patient or group.
    To be dealt with by valid data collection.
38
Q

What is confounding, what are the criteria,

A
  • A form of bias, intrinsic in nature
  • Be associated with exposure (Without being the consequence of exposure) AND Be associated with outcome (Independently of exposure) AND Not causal pathway
39
Q

After adjustment for confounding, how much does the RR need to change to be a confounder?

A

> 10% difference comparing to crude RR

40
Q

What is a nominal scale?

A

Some measurements are categories that have no numeric value and no logical order
Example: sex , genotype

41
Q

What is an ordinal scale?

A

Some measurements are categories that have a rank order (e.g. 1 st , 2 nd , 3 rd , 4 th etc ), but there is no meaningful difference between values.
Example: Socioeconomic status (poor, middle class, rich) OR likert Scale (strongly disagree, disagree, neutral, agree, strongly agree)

42
Q

What is an interval scale?

A

Some measurements have values with equal intervals, and the differences between values are
meaningful. (continuous variable)
Example: Intelligence quotient, body temperature, blood pressure, age, body weight, body height

43
Q

What is a disadvantage of database from a healthcare providers?

A

There is always a reason why the data is collected. GP’s only collect BMI in patients who have a BMI which is too low or too high. Or collecting smoking when a patient is heavily smoking and has some medical conditions.

44
Q

Measurement can be continuous or in categories. What is preferred?

A

In etiologic epi, always use all information (c.q. variation) available as a ground rule!

45
Q

What are the essential descriptors of data collection?

A

Time, population, analysis, experimental/non-experimental

46
Q

Describe the essential descriptor of data collection “time”

A
  • t=0, no time lapse between measurement of determinant and outcome
    o Problematic if you don’t know what came first
  • t>0 time lapse between measurement of determinant and outcome
    o You know what came first in the time interval.
47
Q

Describe the essential descriptor “population”

A
  • Closed cohort:
    o Once member always a member
    o No further members can enter (after formal stop of entry) and start of follow-up
    o Member only leave by death or end of follow-up
    o Closed cohort gets smaller over time
  • Open cohort:
    o Dynamic: members an leave the cohort (e.g. move) or join over follow-up, can leave the cohort by death
    o Open cohort can vary in size over time
48
Q

Describe the essential descriptor “analysis”

A
  • Census: analysis on (experience of) all participants
    o Everyone who meets your criteria is included (RCT)
  • Sampling: analysis (experience of) sample of participants
    o Analysis on (experience of) sample of participants
    o You don’t use every single person from the interested population
    o Case-control study or sometime in cohort study’s e.g. when you sample a part of you cohort for a subanalyses
49
Q

Describe the essential descriptor “experimental or not”

A
  • Experimental
    o Researcher induces change of determinant in order to learn about any effects on incidence of outcome
  • Non-experimental:
    o “observational”, researcher does not induce change of determinant
50
Q

How to deal with confounding?

A

Design study:
- Randomization (RCT)
- Restriction –> impacts domain/generalizability
- Matching
- Propensity scores

Dealing with it in your data analysis (if confounding is measured correctly)
- Stratification
- Multivariable statistical modelling

51
Q

Effect modification

A
  1. If the strength of the association varies over different categories of a third variable, this is called effect modification. The third variable is changing the effect of the exposure.
  2. There is no adjustment for effect modification. Once it is detected, stratified analysis can be used to obtain stratum specific odds ratios.
52
Q

Give the descriptors of the cohort study

A

Time: t>0
Population: Closed or open
Analysis: On all participants (census)
Exposure (determinant): Non-experimental* (observational)
• If a cohort study is experimental, it is called a trial

53
Q

Which measure of frequency is used in a cohort study?

A

Cumulative incidence, incidence density/rate

54
Q

Principle of cohort study

A

Followed up for specified period of time and you look if they develop the outcome or not.

55
Q

What is the purpose of a cohort study?

A
  • Study if an exposure is associated with outcome(s)
  • Compare exposure to comparable non-exposure
  • Estimate risk of outcome in exposed and unexposed (parts of) cohort(s)
56
Q

Which forms of bias can be present in a cohort study?

A
  • Confounding
  • Information/observation bias (if observations/measurements of subjects in one group (e.g. exposed) are systematically different from observations/measurements of the other group )
  • Selection bias (reason of drop-out has a relation with the outcome OR reason of inclusion or no inclusion has a relation with the outcome)
57
Q

What are strengths of a cohort study?

A
  • Can directly measure
    o Incidence in exposed and unexposed groups
    o True relative risk
  • Well suited for rare exposure
  • Temporal relationship exposure-outcome is clear
  • Less subject to biases
    o Outcome not known (prospective)
58
Q

What are limitations of a cohort study?

A
  • Latency period
  • Lost to follow
  • Large sample size (because of costs and infrastructure)
  • Exposure change
  • Ethical considerations
  • Cost
  • Time consuming
59
Q

What is the principle of a cross-sectional study?

A

You measure the determinant and outcome at the same time. Works well to relate a patient profile to a disease  diagnostic research.

60
Q

What are the descriptors of a cross-sectional study?

A

Time: t=0  (problem: what caused what?)
Population: Open
Analysis: On all participants (census)
Exposure (determinant): Non-experimental (observational)

61
Q

What are the strengths and limitations of a cross-sectional study?

A

Strengths
- Quick and cheap

Limitations
- Selection bias
- Consequences of t=0
o Causal inference? Exposure precede outcome?
o Reverse causation?
o Exposure assessed during etiologically relevant time window?

62
Q

Principle of a randomized trial

A

You compare a new determinant with an old determinant/standard treatment. Because of randomization known and unknown confounders don’t play role (when study size is large enough).

63
Q

What are the descriptors of a randomized trial?

A

Time: t>0
Population: Closed
Analysis: On all participants (census)
Exposure (determinant): Experimental*
* Experimental: subjects intervened on with the purpose of learning about intervention effects

64
Q

Principle of randomization

A

Random allocation to exposures (e.g. drug vs no drug).
- Yields groups with comparable (~equal) mean levels of known and unknown determinants of outcome (comparable prognosis with respect to outcome)
- … thus, differences in incidence are due to intervention, not incomparability of prognosis
- So maximizes the probability of comparable confounder distributions across experimental groups
- Validity threats from selection bias (non-differential loss to follow-up) and information bias (blinding, etc.) are basically similar to those in non-experimental cohorts.
- Analysis of trial data is usually simple comparison (no need for adjustment

65
Q

Limitations of a randomized trial

A
  • You cannot always random the exposure (e.g. gender)
  • Unethical, unfeasible, or unaffordable to perform experiments on people?  “observational study”
66
Q

Principle of case control study

A

Provide an estimation of exposure distribution in the source population. If the frequency of exposure is higher among cases than controls then the incidence rate will probably be higher among exposed than non-exposed.

You identify all cases within your population and compare them with randomly chosen controls. 1:1 ratio is enough for case-control studies, but a higher ratio (up to 1:4) can help to narrow your confidence interval. It is not helpful to have a ratio of 1:5 or more.

67
Q

What are the descriptors of a case control study?

A

Time: t>0
Population: Open (sometimes closed/nested)
Analysis: On sample of participants
Exposure (determinant): Non-experimental (observational)

68
Q

Which measure of association do you use in case control studies?

A

Odds-ratio

In the case-control study you cannot calculate the prevalence, cumulative incidence or incidence rate ratio, you can only calculate odds ratios!!!

69
Q

Cohort study VS case control study

A

In a cohort study you compare exposed vs non-exposed. In a case-control study you compare cases vs controls, because you established the amount of exposed and not exposed yourself and therefore it is not informative at all!.

70
Q

What is an odds ratio, give the formula.

A

Exposure odds ratio(OR)= (odds of exposure in cases)/(odds of exposure in controls)= (a/c)/(b/d)=(ad)/(bc)

71
Q

What are the forms of bias in case control studies?

A
  • Confounding
  • Selection bias (if cases that were exposed are more likely to be identified than cases who were not exposed OR people in the control group refuse to participates with the reason related to the determinant)
  • Information bias (Cases know they have the outcome and e.g. over-/underestimate their exposure because they know that they have the outcome)
72
Q

What are strengths of a case control study?

A
  • Rare disease (in a cohort-study you need a huge cohort)
  • Several exposures (because it is smaller and therefore cheaper), but you have 1 outcome
  • Long latency (you can look back a long time), but is based on recall (BIAS!)
  • Rapidity
  • Low cost
  • Small sample size
  • Few ethical issues
73
Q

What are limitations of case control studies?

A
  • Usually no absolute rates/risks  only odds ratio
  • Less suitable for rare exposures
  • More prone to bias
  • Often performed “quick and dirty”

Epidemiology (2 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions