Bias, Sample Collection & Confounding Flashcards
Any systematic error in an epidemiology study that results in an incorrect estimate of the association btwn exposure & risk of dz.
Bias
List 2 broad types of bias
- Selection bias
- Information bias
List the different types of selection bias.
(3)
- Surveillance bias
- Non-response bias
- Inapproriate comparison group
When does surveillance bias occur?
animals are more closely followed by clinicians b/c are these animals are predisposed to a particular dz.
What happens in inappropriate comparison group?
Control does not appropriately represent the population from which cases arose
How can we avoid Selection Bias?
- randomly select subjects
- high response rates
- low withdraw rates
List the 5 different types of Information Bias.
- Interviewer bias
- Recall bias
- Incomplete medical record use
- Misclassifcation bias
- Observer bias
What happens in interviewer bias?
Interviewer asks questions in such a manner that affects subjects’ response
(leading questions, etc)
What happens in Recall Bias?
- Subjects don’t accurately remember exposure
- Family members provide info on subject
What happens in Misclassifaction bias?
- Inaccurate data collection leads exposure or disease to be misclassified.
- Case as control, control as case
When does Interobserver variation of Observer Bias occur?
- >2 observers
- variation btwn those reading or interpreting results
When does Intraobservation variation of Observer Bias occur?
- 1 observer
- variation btwn 2 observations made by 1 person
(think about interpretation of radiographs, experienced vs. novice)
How can we control bias?
- Careful study design
- Good data collection
- blindness
- well trained personnel
- multiple sources of data
Best method of sampling?
Random sampling
Method characteristics:
- each subject in a population has equal chance of being selected
- computer programs
simple random sampling
