L7 - Descriptive Reports and Observational Study Designs Flashcards
Experimental
Investigators control who receives the intervention and how
Helps to determine Cause and effect
Can use randomization and controls to minimize bias and confounding
Study Types
Experimental
Observational
Descriptive
Observational
Does not involve investigator intervention
Compares outcomes from naturally occurring differences between populations
Can test for associations, but not causality
Descriptive
Allows associations to be considered
Forms initial hypothesis
Better to have big sample
Exposure
In epidemiology, indicates potential risk factor
In studies involving drug therapy often means treatment
Outcome
In epidemiology, may mean whether disease developed
In studies involving drug therapy, often means whether disease was successfully treated
Prospective
Normal
Data are collected forward in time from the date of study initiation
Retrospective
Study that analyzes historical (previously collected) data
-Medical records, insurance info, etc
Q: Which is more robust, Prospective or Retrospective?
Prospective
-More in control
-In charge of patient recruitment, how the exposure is given, how the data is collected, etc.
-Retrospective this is already done for you - no wiggle room
Bias
Systematic deviation from the truth that distorts the results of research
NEVER random
Most common: Selection bias
Types of Selection Bias
Referral Bias
Self selection bias
Prevalent bias
-Prevalence: existing cases in a given period
Protopathic
-Effect is treated as a cause
Information and Misclassification BIas
Happens when you are collecting data
Misclassification
-When one group of patients is incorrectly classified as the other (i.e. controlled as experimental and vice versa)
Recall bias
-Asking patients about a medication they took last year
-If you had a bad reaction, you are more likely to remember that
Detection bias:
-When investigators investigate a study
-Subconsciously you know who is the treatment group and who is the controlled
-You will focus more on the treatment
Confounding
Relationship between treatment and response (or exposure and disease) is actually attributable to another variable (the “confounder”)
Confounder is independently related to BOTH the exposure AND the outcome
If variable is exposed to just one – potential confounder
Descriptive Reports
Purpose:
-Document and communicate experience, share ideas, or describe novel treatments and unusual events
Case reports and case series
Ex.
Possible adverse drug reactions
Unusual complications of illness
Surprising presentations of disease
Descriptions of rare treatments
CANNOT test a hypothesis in descriptive reports
-Can only form hypothesis
Advantages and Disadvantages of Descriptive Reports
Advantages
-Allow hypotheses to be formed
-Often the first step in exciting discoveries
-Can guide “salvage therapy”
Disadvantages
-Do not always provide detailed explanations
-Do not establish causality or association
