Ph 150A - Epidemiological Studies

Front 1

Association

Back 1

the occurrence of an outcome increases/decreases with the level of exposure

* if associated, the relationship may/not be causal

Front 2

Ecological Study

Back 2

* use average group data to assess exposure and outcome in POPULATIONS

* Advantages:
- grouped data may reflect best exposure
- easier to detect exposure differences btw groups than within
- useful in identifying problems
- inexpensive b/c data already collected

* disadvantages:
- ecological fallacy: ppl with outcome may not be the ones w/ exposure
- groups usually differ in other ways beside exposure
- group data may vary in quality
- possible selection bias

Front 3

Observational Study

Back 3

describes relationship between exposure and outcome in pop'n at a single point in time

1) begin with defined study group
2) measure exposure and outcome simultaneously
3) classify and compare groups

* advantages:
- useful to describe healthy pop'ns
- associations found can be useful in hypothesizing
- efficient and quick

* disadvantages:
- can't determine cause and effect btw exposure & outcome
- can't sort out existing (prevalent) outcomes from new (incident) outcomes
- can have problems in selecting subjects & making measurements

Front 4

Case Control Study

Back 4

1) start with diseased cases (outcome)
- select controls
- controls can be matched/unmatched
2) go back in time to determine exposure in both groups
3) classify by exposure
4) compare exposure rates from cases and contols

* most case control studies use incident cases to describe factors associated with disease development

* advantages:
- efficient, cheap, quick
- only feasible method for rare outcomes
- can evaluate wide ranges of exposures

* disadvantages:
- difficult to select cases and controls
- retrospectively asseses exposure
- not able to assess cause-effect relationship
- can't tell you about development of disease

Front 5

Prospective Cohort Study

Back 5

describes the relationship between exposure and development of outcom over time

* advantages:
- study disease development over time
- investigator can control methods of data collection
- allows exposure to occur before disease
- provides best evidence for a cause-effect relationship

* diadvantages:
- can take a long time for results
- expensive
- not good for rare outcomes

Front 6

Prospective Cohort Study

Back 6

1) identify study group based on existing records
- data on exposure and outcomes were prospectively recorded in the past
2) classify exposure based on records
3) analyze subsequent outcome from records

* advantages:
- studies outcome development over time
- less bias in assessing exposure
- best observational design for cause-effect

* disadvantages:
- time consuming
- expensive
- need large numbers
- exposures can change over time
- no response, loss to follow-up, surveillance bias

Front 7

Retrospetive Cohort Study

Back 7

1) identify study group using existing records where data on exposure and outcome were prospectively recorded in the past
2) classify exposure based on records
3) analyze subsequent outcome from records

* advantages:
- because data already collected, less expensive, quicker
- other advantages of cohort study

* disadvantages:
- record data may not be of uniform quality/available

Front 8

Nested Case Control Study

Back 8

1) identify the cases and controls from a cohort where data on exposure and outcome were prospectively collected

2) classify exposure of the cases and controls based on the already-collectd cohort data

* case control study is nested within a cohort

* advantages:
- quick, less expensive, good for rare outomes
- exposure data collected and recorded before outcome occurred
- know underlying pop'n: cases & ctrls may be more representative
- depends on recorded/stored data

Front 9

Variations of Cohort Study

Back 9

1) prospective: enroll in present, follow in future

2) retrospective: same direction, but all data already collected for another purpose

3) nested case control: case control study performed using data collected for a prospective cohort study

Front 10

Randomized Control Trials

Back 10

* tests how well an intervention prevents an outcome

* exposure (treatment) is assigned by investigtors & applied to INDIVIDUALS

1) select study group
2) divide study group into treatment and control
3) apply intervention
4) assess and compare outcome in treated and control groups

* advantages:
- strongset evidence for causality
- if randomized, selection bias and confounding removed
- if blinded, less observer bias
- meta analysis: can pool results of several studies

* disadvantages:
- not real life
- high cost
- inappropriate/unethical for many investigations
- can have poor compliance, loss to follow-up