Chapters 10-14 Flashcards
What is bias?
a systematic error in the design or conduct of study that leads to an erroneous association between the exposure and disease
What is confounding?
The mixing of effects between the exposure, the disease, and a third variable that is termed a confounder. Like bias, confounding distorts the relationship between an exposure and disease
What is random error?
The probability that the observed result is due to “chance,” an uncontrollable force that seems to have no assignable cause.
What are key factors of bias?
an alternative explanation for an association
Usually does not result from a “prejudiced” investigator
Can pull an association either toward or away from the null
Amount of bias can be small. moderate, or large
Is avoided when the study is carefully designed and conducted
What is selection bias?
An error resulting from subject selection and participation. Association among study subjects differs from association among eligible source population.
What are some examples of selection bias in a cohort study?
loss to follow up
the “healthy worker” effect
self selection bias
differential surveillance, diagnosis, or referral of subjects
What are key factors of selection bias?
It is more likely to occur in case control and retrospective cohort studies (both exposure and disease have occurred by the time of subject selection)
It can also occur in experimental studies and prospective studies due to loss to follow up
Little can be done to fix this bias once it occurs
It can pull an association either towards or away from the null
What is Information Bias?
a flaw in measuring exposure, covariate, or outcome variables that results in different quality (accuracy) of information between comparison groups.”
What is recall bias?
This occurs when there is a differential level of accuracy in the information provided by compared groups.
What is interviewer bias?
a systematic difference in soliciting, recording, or interpreting information that occurs in studies using in person or telephone interviews.
What is misclassification?
A measurement error
There is an error in the classification of the exposure or the disease (essentially a clerical error)
What are characteristics of information bias?
Results in participants who are incorrectly classified as either exposed or unexposed or as diseased or not diseased
Occurs after subjects have entered a study
Occurs in both retrospective and prospective studies
Specific types include recall bias, interviewer bias, and differential / nondifferential misclassification
Methods for prevention include masking / double masking (like blind and double blind), using multiple measurements, specific criteria, and detailed questionnaires / surveys=
What is differential / non-differential misclassification?
Non-differential misclassification occurs when the probability of individuals being misclassified is equal across all groups in the study. Differential misclassification occurs when the probability of being misclassified differs between groups in a study (Porta et al. 2014).
What is an example of a confounding variable?
The association between Diabetes and Dementia; those in a certain study were an older population. The RR of Diabetes leading to dementia was determined to be about 3.5
When another study was conducted with “control” of the confounding aspect (age,) risk of diabetes leading to dementia was revealed as closer to 2.
What criteria does a variable need to meet in order to be confounding?
It must be associated with the exposure in the population that produced cases (it most be more or less common in the exposed group than the control group). The variable must be an independent cause or predictor of the disease (association between confounder and disease is present among both exposed and unexposed individuals).
It also cannot be an intermediary step in the causal pathway between exposure and disease.
How do we calculate the magnitude of a confounding variable?
RR crude - RR adjusted / RR adjusted
What are methods for controlling confounding variables in design and analysis within a study?
Design stage: randomization, restriction, and matching
Analysis Stage: standardization, stratified analysis, matched analysis (needed for case control only), multivariate analysis
What is residual confounding?
a term for the confounding that remains even after many confounding variables have been controlled. Residual confounding can come from several sources (inability to collect needed data, categories too large (e.g. 10 year age groups with many differences within this category), misclassification, various uncontrollable factors
What is precision?
Lack of random error
Most epidemiological methods are not exact, leading to the need for the term “measurement error” (or standard error?)
What is statistical inference?
the process of making generalizations from a sample to the source or parent population (based on the mathematical theory of probability)
What is the null hypothesis versus the alternative hypothesis?
The null hypothesis is often stated as the assumption that there is no change, no difference between two groups, or no relationship between two variables. The alternative hypothesis, on the other hand, is the statement that there is a change, difference, or relationship
What are characteristics of p values?
Can be one or two sided (e.g. upper value is probability of observing RR greater than the expected value, the lower value is the probability of observing RR smaller than the expected value).
A continuous statistic ranging from 0 to 1; a small p value indicates a low degree of compatibility between the observed data and the null hypothesis (there is a very small chance this result would have generated, indicating a difference within the population )
How do the terms “(not) statistically significant” relate to the null hypothesis?
If results are not statistically significant, this means the data failed to provide sufficient evidence to doubt the null hypothesis.
What are alpha and beta errors?
Alpha errors refer to errors which would “prove” null hypothesis is not correct,
whereas Beta errors refer to errors which would “prove” alternative hypothesis is not correct
What is the strict statistical definition of a 95% confidence interval?
if a study were repeated 100 times and 100 point estimates and 100 confidence intervals were calculated, 95 out of 100 confidence intervals would contain the true measure of association
What are the 3 probability distributions most typically used in epidemiological research?
normal distribution
Poisson distribution
binomial distribution (including Bernoulli trials)
When is the Poisson Distribution used?
this is a discrete distribution commonly used to describe rare events
What are epidemiological methods used for hypothesis testing?
the T test
Chi Square
Z test
What are absolute versus relative measures of comparison?
Risk Difference (and its Confidence Intervals) is absolute
Relative Risk / Odds Ratio (and their Confidence Intervals) are relative
What are the two main sources of random error?
Measurement error
Sampling variability
What is internal validity?
The extent to which the results of a study accurately reflect the true situation in the study population
What is external validity?
The extent to which the results of the study are applicable to other population
True or False, If a study lacks internal validity, it can still have external validity.
FALSE (if a study lacks internal validity, external validity is not even examined)
What is effect measure modification?
the strength of the association between an exposure and disease differs according to the level of another variable (effect modifier)
What does homogeneity effect mean?
effect modification is absent
What does heterogeneity / effect modification mean?
change in association according to the 3rd variable
(dependent upon ratio measure or difference measure)
How do confounding variables and effect measure modification differ?
confounding variables are considered a nuisance
effect measure modification is an interesting phenomenon that epidemiologists aim to describe. Its presence depends on a particular measure of association
It is measured in detail via stratified analyses, through visual inspection, statistical tests, or evaluation of excess risks
What are synergy and antagonism in terms of effect modification?
synergy occurs when two factors work together to produce more disease than one would expect based on the action of either factor alone
antagonism occurs when one factor reduces the impact or even cancels out the effect of another factor
Describe criteria for effectively evaluating epidemiological studies
Describe the decision making process for evaluating epidemiological data (especially for confounding variables and effect modifiers)
