Study Design and Sampling Flashcards
Descriptive studies
Describe characteristics of disease in the population without identifying associations or causal inferences
Observational studies
investigation of associations or causal inferences that are occurring in the population without manipulation or intervention from the researcher
Experimental Studies
Researcher intervenes to allocate individuals to treatment or control groups then measures the effect
3 types of observational studies
- cross-sectional
- cohort
- case-control
Cross-sectional study
analyzes data across as population at a single point in time using prevalence – sampling is representative of the population NOT based on exposure or outcome
a “snapshot”
Cross-sectional studies identify ____ NOT ____
cross-sectional studies identify association NOT causation
What is a critical aspect of cross-sectional studies?
which animals you select to sample (needs to represent target population)
Why is questionnaire data collection difficult to do well?
always introduces bias (lies, mistakes, etc.)
Advantages of cross-sectional studies
- relatively easy, quick, and inexpensive
- can study multiple exposures/diseases/outcomes
- no loss to follow up
- no case or control selection bias
- used to ID high groups
Limitations of cross-sectional studies
- not useful in establishing causal relationships
- prevalence can underestiamte disease risk if the sick die, recover quickly, or are removed
- impractical for rare disease
- relies on good sampling
Case-Control Studies
studies that start with selecting cases and controls, then look back to ID exposures (sampling is based on outcome)
Case-control study design (5)
- define hypotheses
- develop case definition
- develop control definition
- access and select cases and controls
- conduct study, analyze data, report results
How should controls be selected for case-control studies?
- should be similar to cases but without disease
- should have the opportunity for exposure and diagnosis
Controls in a case-control study provide the _____
estimate of exposure rate among non-diseased
What calculated value is used in case-control studies?
odds ratio (how many times greater are the odds of disease among exposed than among non-exposed)
this value will estimate relative risk
Why can’t you calculate incidence rates?
you started with identification of prevalent cases
Advantages of case-control studies (3)
- best for rare diseases
- inexpensive and fast
- can study several potential exposures at the same time
Limitations of case-control studies (5)
- control group hard to pick
- high potential for bias
- can’t calculate incidence, prevalence, relative risk, or attributable risk
- can’t assess rare exposures
- problems with temporal sequence of data
Cohort Study
selection of exposed vs. unexposed groups, then comparing the incidence of disease among them
Advantages of cohort study (5)
- controls bias
- best assessment of exposure
- can estimate multiple outcomes per exposure
- more conclusive than case control studies
- allows study of reare exposures
What disease characteristics limit the use of cohort studies? (3)
diseases that:
1. have mild signs
2. do not require medical attention
3. can only be diagnosed with a medical visit
Disadvantages of cohort studies (5)
- not practical for rare disease
- requires larger sample size
- expensive
- long time commitment for follow-up
- can lose participants (causes bias, may reduce validity)
Why do a clinical trial?
- provide data for drug approval
- compare treatments
and more!
What is the only factor that should differ among groups in a clinical trial?
treatment (all other factors should be the same)
Different types of negative controls for clinical trials (3)
- no treatment (hard to make this a “blind study”)
- placebo
- standard care (new treatment is added to standard in tx group)
Types of random allocation
- completely random
- by pairs
- within blocks
Types of non-random allocation
- systematic (gait splitting, every other animal, repeating tx order)
- historical (borrowing information from animals enrolled as controls in previous studies)
- clinicians disecretion (doctor decides)
What is an important aspect of data collection to consider in a clinical trial?
masking/blinding procedures to keep those involved from knowing treatment assignments
Advantages of clinical trials
- high degree of internal validity
- considered “gold standard” in medicine for establishing efficacy
- required for drug approval
- clinical trials minimize bias
- assumption of causality (treatment cause whatever effect/outcome is present)
Limitations of clinical trials (3)
- ethics of placebos are under debate
- expensive
- can take a long time