Brainscape's Knowledge GenomeTM

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


See full index

Research Methods & Study Design > Study Design > Flashcards

Study Design Flashcards

1
Q

What is a key factor to remember when interpreting ecological studies?

A

Associations seen at a POPULATION level CANNOT be applied to the INDIVIDUAL level

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

What are cross-sectional studies?

A

These collect observations in representative sample of individuals in a population at ONE POINT IN TIME (= snapshot of health of population)

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

What are some of the design issues associated with cross-sectional studies?

A
  • Confounders: need to collect data for these so can control for them in the analysis
  • Reverse causation
  • Recall bias (past exposures)
  • Selection bias (non-response)
  • Prevalent disease NOT incident disease
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the advantages and disadvantages of cross-sectional studies?

A

Advantages = can examine prevalence and associations, can look at more than one outcome and exposure and relatively inexpensive/quick due to no follow up

Disadvantages = no temporal association, not suitable for rare outcomes or exposures, high possibility of recall or reporting bias

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

Name the 6 types of study design? All are observational studies except for one - which one and what is it?

A
  1. Case report/series
  2. Ecological
  3. Cross-sectional
  4. Case-control
  5. Cohort/longitudinal
  6. Clinical trials

Clinical trials is experimental NOT observational

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are case control studies?

A

Where subjects are selected on the presence/absence of disease

They are good for rare diseases and allow you to obtain data on last or current exposures

Assess association but can’t assess prevalence or incidence (use cross-sectional or cohort studies, respectively, to do this)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the four important things to remember when choosing your cases for a case-control study?

A
  1. What are the case definition criteria? (Clinical/lab etc)
  2. What are the eligibility criteria? (E.g. Women only)
  3. What is the source of the cases? (Hospital/community/population)
  4. Prevalent or incident cases?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are the three important things to remember when choosing your controls for a case-control study?

A
  • They must be free from disease
  • They must fulfil the same eligibility criteria as cases
  • To avoid selection bias, they should be drawn from the population that gave rise to the cases
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What measure of effect is used for case-control studies?

A

Odds ratio

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

When calculating measure of effect for case-control studies using SPSS, what issue arises and how can we combat it?

A

The OR is calculated the wrong way round, because it orders the groups based on their code (e.g. Controls = 0, Cases = 1).

We can combat this by either taking the reciprocal, or by simply recoding the data so that cases = 0 and controls = 1.

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

What are the advantages/disadvantages of case-control studies?

A

Advantages include no need for follow-up and can look at multiple exposures

Disadvantages include not being able to estimate prevalence, subject to bias and can only study a single outcome/disease

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

How do cohort studies work?

A

You pick your target population of people without the disease with the relevant exposure, and then monitor them over time to see who develops the disease

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

What’s are the pros and cons of cohort studies over case control?

A

Pros = higher study numbers, can study multiple outcomes, prospective measure of exposure, can measure changes in exposure, less bias and can measure incidence

Cons= much longer time to complete, higher cost

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

What is the measure of effect used for cohort studies and how is it calculated?

A

The rate ratio

Rate ratio = rate of disease amongst exposed/rate of disease amongst unexposed

A rate ratio of 0.13 would signify exposed group were 87% less likely to develop the disease then the unexposed

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

Explain the use of retrospective cohorts for a cohort study

A

Using a cohort where there is existing data on exposure and disease

This overcomes loss to follow-up and long latency periods, is cheaper and quicker to perform than prospective cohorts but needs good exposure data with low misclassification

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

What are the advantages if cohort studies?

A
  1. Can Measure Multiple disease outcomes
  2. Assesses temporal rship between exposure/disease
  3. useful for rare exposures
  4. Selection/recall bias could be very low
17
Q

What are the disadvantages of cohort studies?

A
  1. Time consuming and Expensive
  2. Results can be affected due to loss of follow-up
  3. Inefficient if effect of exposure has long latency period
18
Q

What are randomised controlled trials and how do they different from observational studies?

A

RCTs compare one intervention against another (e.g. Treatment vs. Placebo). Any difference in outcome is due to the intervention

They differ from observational studies because the researcher decides who is getting what exposure

19
Q

Why are Clinical trials seen as the gold standard for study design?

A
  • Chance is minimised by estimating study power (study enough people
  • Confounding is removed by randomisation
  • Bias is removed by blinding
20
Q

Despite being the gold standard, why would you use other study designs instead of an RCT?

A
  • Can only test beneficial interventions (e.g. would be unethical to make people smoke)
  • Expensive and can take a long time
  • May not be generalisable due to recruitment restriction (e.g. Older ppl, pregnant women etc)
21
Q

What are the two main types of RCT and what do they measure?

A
  1. Explanatory trials: these measure the efficacy of a particular intervention
  2. Pragmatic trials: these measure effectiveness by assessing the intervention in practice
22
Q

Results from two similar studies showed identical risk ratios but one rejected the null hypothesis but the other did not, why?

A

To reject the null hypothesis you need a significant p-value. Risk ratio doesn’t take into account sample size. Therefore the study that obtained a statistically significant result with the same risk ratio will have had a greater sample size - more study power

23
Q

How can you increase your generalisability in an RCT?

A

Minimise the exclusion criteria

24
Q

Give four methods for randomisation

A
  1. Tossing a coin
  2. Random numbers
  3. Permuted blocks
  4. Minimisation
25
Q

What is allocation concealment?

A

This is where the investigator and subject are blinded to the randomisation sequence and therefore do not know who is in which group

Methods include sequentially sealed opaque envelopes or use of a central allocation office

26
Q

What are the 5 main types of basic RCT design?

A
  1. Parallel group
  2. Crossover
  3. Within patient
  4. Equivalence
  5. Cluster
27
Q

What is parallel group RCT design?

A

This is the most simple whereby patients are randomly allocated to the active drug or placebo groups

28
Q

What is the crossover RCT design?

A

This is a more complex version of parallel groups whereby patients are randomly allocated to one of two groups but at a specific point in the study they undergo a ‘washout period’ where they return to baseline and they then swap to the other group

Both groups end up getting both treatments

29
Q

What is ‘within patient’ design for RCTs?

A

Similar to a crossover design but here the subject gets both treatments at the SAME time (e.g. One treatment on each side of the scalp)

A smaller sample size is needed and no washout period is required

30
Q

What is an equivalence RCT design?

A

Here you are comparing whether drug A is as effective as drug B

A parallel group design is used but you assess whether new or old treatment are better than the placebo

A non-significant p-value here may not always mean they are equivalent, it may be that you have a lack of study power and need to increase sample size

31
Q

What is a cluster design for RCTs?

A

Similar to a parallel group study but here groups rather than individuals are randomised to an intervention

Allow a wider population to be examined but need to be wary of confounders such as relating to geographical distribution

32
Q

How is loss-to-follow-up dealt with in RCTs?

A

Using the Intention to Treat principle

For pragmatic trials comparing two treatments, you analyse based on the treatment group the subjects were initially randomised to (regardless of if they discontinued or swapped over)

For per protocol (efficacy) analysis you would analyse based on the final treatment they received, not necessarily the initial group to which they were randomised

33
Q

What are the advantages if RCTs?

A
  • Rigorous evaluation of a single exposure or treatment
  • Prospective design so can assess causation
  • potentially eradicates bias and confounding
34
Q

What are the disadvantages of RCTs?

A
  • Expensive and time consuming
  • sample size can be too small
  • risk of hidden bias (due to poor blinding)
  • limited generalisability due to exclusion of high risk groups (elderly, pregnant women etc)
35
Q

What is the incidence rate for cohort studies? How is it calculated and interpreted?

A

This is when you use open cohorts

Incidence rate = number of new cases of disease/total persontime at risk

An incidence rate of 7 per 100 person years could mean:

  • 7 new cases of disease in 100 ppl followed for 1 year
  • 7 new cases of disease in 10 ppl followed for 10 years
36
Q

What are the best study designs for:

  1. Investigation of a rare disease
  2. Investigation of a rare cause
  3. Examining multiple outcomes
  4. Studying multiple exposures
  5. Measure of time relationship
  6. Direct measurement of incidence
  7. Investigation of long latent periods
A
  1. Ecological or case control
  2. Cohort
  3. Cohort
  4. Case control
  5. Cohort
  6. Cohort
  7. Case control or cohort (if historical)
37
Q

Incidence can be used to calculate prevalence - how?

A

Prevalence = incidence X average duration of study

P=IxD

38
Q

What studies can be used to calculate prevalence and incidence.

A

Prevalence: cross-sectional studies

Incidence: cohort and longitudinal studies

Research Methods & Study Design (10 decks)

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