Research Design and Methods Flashcards
What is the purpose of Research design?
refers to the research plan used to answer the question under consideration
-The research design is chosen based on the question
What defines the accuracy of study results?
Validity
The more a design refers to the question of the study the more valid the study is
Internal Validity
The extent to which a clinical outcome (dependent variable) is caused by the treatment (independent variable)
How valid are the methods to answer the question
What are the dependent and independent variables?
Dependent: Often the outcome (outcome measure) -> depends on the independent variable
Independent: We are trying to determine the outcome of the independent variable: disease or treatment
-> It influences the dependent variable and drives the outcome
External Validity
Application of the results to the broad population -> Can the study be replicated in other situations?
What factors may reduce internal validity?
-Confounding variables: a third variable that has an effect on the two I am interested in and messes up the results - f.e. LDL indicated heart disease -> confounding variable would be smoking, hypertension (it increases heart disease as an independent variable)
-Chance/coincident findings
How to determine if a study has external validity
Is the study repeatable in other populations?
How might a researcher maximize
external validity?
Choose participants from different populations in different countries with similar lifestyles, and behaviors, … f.e. Western countries
What is Correlation?
Two variables are related or correlated -> it does NOT mean that X causes Y
What is an association?
-If X looks like it is related to Y = Association, Correlation
-BUT Association is not equal to Causation
What are the criteria to prove a cause-and-effect relationship = Causality?
Criteria by Sir Austin Bradford Hill (1965)
-Temporality (time-related f.e. taking antihypertensive and feeling deezy an hour later due to low BP)
-Strength (higher dose - higher effect - similar increase in all curves on the graph)
-Biologic Gradient (similar to strength)
-Consistency (effect happens over and over again)
-Specificity (happens with a specific drug)
-Plausibility (antihypertensive drug lowers BP)
-Coherent (same effect in animals?)
-Analogy (similar drug class with similar effects)
-Experiment (experimental evidence)
What are spurious relationships?
a mathematical relationship where two variables have no causal connection but they look as if they have -> due to coincidence or a third factor (confounding factor or variable)
What is the difference between descriptive and analytical studies?
What is the purpose? Just describing or testing a hypothesis
Descriptive: just describing information about a disease or event without intervening
Analytical: trying to understand the relationship or causal mechanism between 2 or more variables
What are factors that help to describe effects in a descriptive study?
Who?
What?
Where?
When?
Why?
What are study designs based on Time Orientation?
-Prospective (takes longer, more expensive)
-Retrospective (often medical records, faster, cheaper)
What are study designs based on Experimental Setting?
-Randomized Controlled Trial (required by FDA, gold standard for safety and efficacy, ALWAYS prospective)
Experimental Design: The researcher controls the treatment (independent variable) through randomization
-Observational (observing a relationship between two variables, not randomized)
What are the Observational studies?
-Case Report - Single Case
-Case Series - Study of multiple similar cases
-> very weak evidence
->used to uncover relationships that can be further explored
-Cross-sectional studies
-Case-control studies
-Cohort studies
What is a Cross-sectional study?
-Examine population characteristics at one point in
time (a cross-section)
-presents a snapshot of the exposure status/outcome
Why do Cross-section studies cant explain causation?
Because exposure and outcome are measured at the same time
Which study is used to determine prevalence (proportion of population with a given characteristic in a time period)?
-Cross-sectional studies
What are Case-Control studies?
Comparing two groups - one with the disease/outcome and one without and tracking back to look for the exposure (retrospective)
-Case group AND Control group share same characteristics EXCEPT of the disease
What types of diseases are investigated with Case-control studies?
Rare diseases, where it takes time between exposure and the outcome
What are cohort studies?
-Two groups (exposed and unexposed - both don’t have the disease) are followed for a period of time until the development of the disease
What are cohort studies used for?
-Determine the incidence (number of cases) in the exposed vs. unexposed group
-Exposure causes outcome -> cohort studies can be used to study CAUSATION
What is a way to prove causality in an experimental study?
-Randomization
-varying dose
-Blindfolding
Are RCTs prospective or retrospective?
ALWAYS Prospective
What is the disadvantage of not having a randomized study?
variables and participants are self-selected -> opens room for confounding factor
