4) Observational Studies Flashcards Preview

Research Methods > 4) Observational Studies > Flashcards

Flashcards in 4) Observational Studies Deck (21)
Loading flashcards...
1
Q

Why use an observational study

A
  • Randomisation unethical - harmful exposure
  • Randomisation impractical/impossible
  • rare diseases adverse events
  • LT outcomes
  • Descriptive questions
2
Q

Types of observational studies

A
  • cohort studies (Prospective or retrospective)
  • case control studies
  • cross sectional
    Ecological
3
Q

Prospective Cohort study

A
  • how the present affects the future

- start with a pop, follow up and measure outcome

4
Q

Prospective cohort study strength

A
  • highest level of evidence for causality after RCT
  • outcome measured after exposure to risk has occured
  • can look at multiple risk factor + disease
  • examine incidence and natural history or clinical course of a disease
5
Q

Prospective cohort study weaknesses

A
  • Expensive
  • time consuming
  • no randomisation - confounding
6
Q

Retrospective cohort study strength

A
  • less costly or time consuming as used already collected data
  • useful for diseases/outcomes that require a long time to develop
  • can involve large administrative surveillance/datasets which can be useful for detecting rare outcomes/adverse events
7
Q

Prospective study weaknesses

A
  • less control over what is collected
  • bias
  • may be inacurracies of data collection
8
Q

Selection bias

A
  • distortion in results because of how the participants were selected
  • control and comparison groups need to be as similar as possible
  • lack of randomisation
  • loss to follow yp
9
Q

How to reduce selection bias in cohort studies

A
  • sample selection: restriction and matching
  • Loss to follow up: exclude participants that are likely to drop out/ get contact detail/ maintain regular contact
  • Analysis stage: adjust for confounders, measurement error
  • Blinding
10
Q

Case control studies strength

A
  • Good for rare diseases - require fewer participants

- good for diseases that take a long time to develop

11
Q

Case control studies weaknesses

A
  • can’t estimate prevalence and incidence
  • can only study one disease
  • more susceptible to bias
12
Q

Biases

A
  • selection bias
  • measurement bias
  • Recall bias
13
Q

Reducing measurement bias

A
  • use previously collected data that was recorded before the outcome occured
  • structured questionnaires/ interview scripts
  • use of memory aids to aid recall of past exposures
  • blinding
14
Q

Cross sectional studies

A

Observational studies conducted at one poitn in time withouit a follow up period
- risk factor and outcome are measured in selected sample at one point in time

15
Q

Two main types of cross sectional stidies

A
  • Descriptive: how common something is

- Analytic: examines relationship between predictors and outcome

16
Q

Cross sectional studies strength

A
  • Useful for establishing the prevalence of a disease - sample must be representative
  • no follow up : not time consuming or expensive
17
Q

Cross sectional studies weaknesses

A
  • difficiult to establish a cause-effect relationship
  • hypothesis generation only
  • prevalence- incidence bias
18
Q

Ecological studies

A
  • looks at relationships between two things that are measured at a population level rather than an individual level
19
Q

Ecological fallacy

A
  • relationships observed in groups also hold for individuals

- confounding factors

20
Q

Ecological studies strength

A
  • often uses available data that doesnt have to be collected and doesnt need an ethics application
  • cheap
21
Q

Ecological studies weaknesses

A
  • possibility that many other things could be responsible for the observed difference
  • hypothesis generation only