Epidemiology Exam Flashcards
Two groups given different medications to see which medication performs the best…
Clinical Trial
Looking at people who have have been EXPOSED compared to those who have not been EXPOSED over the course of a few years to see if an outcome occurs and if it it’s linked to exposure or not…
Prospective Cohort
Looking BACK at people who have have been EXPOSED compared to those who have not been EXPOSED to see if an outcome occurred and if it was linked to exposure or not…
Retrospective Cohort
You want to know if two things are associated…so you call a bunch of people and ask about their habits to help you find out…
Cross Sectional
To calculate PREVALENCE…
The people WHO DO HAVE divided by the TOTAL NUMBER OF PEOPLE WHO COULD HAVE.
To calculate SENSITIVITY…
True positives / True positives + False negatives.
To calculate SPECIFICITY…
True negatives / True negatives + False positives
To calculate POSITIVE PREDICTIVE VALUE…
True positives / True positives + False positives
To calculate NEGATIVE PREDICTIVE VALUE…
True negatives / True negatives + False negatives
Why do we randomize?
Decrease the likelihood that observations are due to chance
If you give some people an actual drug and some people a fake drug but people in both groups improve…this is likely due to ____
Placebo effect
You are doing a study which requires you to follow up with patients, but some of them are morons and drop out…but you still include the results associated with the dropouts in your study. This is an example of____
Intention to treat
If it is a new drug…chances are a ____ was utilized to get it approved.
Randomized controlled clinical study
If you are trying to see if “one thing” affects another…the “one thing” is the ____
Independent variable
To calculate RISK of something for EXPOSED people…
Those exposed WITH the outcome / Total number of people exposed
To calculate RISK of something for UNEXPOSED people…
Those unexposed WITH the outcome / Total number of people unexposed
To calculate RISK RATIO…
Risk of EXPOSED / Risk of UNEXPOSED
Making sure you account for all possible factors that may be associated with outcome into your study prevents ____ from contaminating data
Confounding Factors
% of people with a characteristic/disease in a population at a certain point in time is ____
Prevalence
If two people perform two different tests to determine if a disease is present…the one performing tests with HIGHER SENSITIVITY (regardless of specificity) will likely correctly identify ____ people with the disease
MORE. You will have a higher chance of catching MORE people with the disease (letting fewer go).
You have a hypothesis you are trying to test. You perform the test and your p-value ends up being a very low number (0.001) which forces you to ____ “null hypothesis”.
REJECT. Your hypothesis was CORRECT and you could safely REJECT the null.
If an odds ratio associated with an exposure is < 1.0…then the exposure is…
PROTECTIVE (subjective). Whatever the exposure is…it is associated with LOWER ODDS of outcome.
If an odds ratio associated with an exposure is > 1.0…then the exposure is…
HARMFUL (subjective). Whatever the exposure is…it is associated with HIGHER ODDS of outcome.
If an odds ratio associated with an exposure is = 1.0…then the exposure…
DOES NOT AFFECT THE ODDS OF OUTCOME
If a 95% confidence interval range is given and DOES NOT include 1.0 (null value), then it is ____
Statistically significant
If results are found to be statistically significant at the 5% level…this means that if an association was made between two things…there is a ____% possibility that the association occurred by chance alone.
5%
If the results of a study can be applied to other populations…this study would have good ____
External validity
Survey studies sort of suck because they are limited in many ways…including…
The fact that a person volunteers may mean they are biased in the first place, survey answers may not be accurate, in order to be considered externally valid there must be standardizing of the survey, and other stuff can suck too…just answer all of the above
Rejecting the null hypothesis when its actually correct…
Type 1 error
Failing to reject the null hypothesis when it is actually incorrect…
Type 2 error
Hey…if you add more people to the study and increase sample size…the confidence interval gets ____
Narrower
T/F If the confidence interval INCLUDES the null value of 1.0, it is an awesome study and totally significant.
FALSE. Don’t be a numb skull.
If it says “at a 95% confidence interval between X and 2X” this means that we are ____% confident that the ____ proportion is between ____ and ____
95, true, X, 2X
T/F Only diseases with high prevalence are suitable for screening programs.
FALSE
____ studies are good at studying rare exposures
Cohort
____ studies are good at studying rare outcomes
Case-control
(Retrospective / Prospective) cohort studies include more complete and accurate data…
Prospective
How do you increase statistical power?
Increase the sample size!!!
The probability that a person with a negative result actually does NOT have the thing for which they were tested is called…
Negative predictive value
The probability that a person with a positive result actually DOES have the thing for which they were tested is called…
Positive predictive value
Width of a confidence interval ____ as the sample size ____
Decreases, increases
Incidence rates are usually acquired by calculation of ____
Person-time
If no statistical significance is found between one thing and its effects on another…then you (reject / fail to reject) the null hypothesis.
Fail to reject the null hypothesis (that there is no significant difference between the two)
Knowing what exists in a population at a given point in time is considered its ____
Prevalence
Seeing what happens to a population over a given time (new cases) is considered ____
Incidence
Null value for relative risk is always ____
1.0
A stronger association between two things is identifiable the (closer / further away) the relative risk is to the null value of 1.0.
Further away.
____ is the ability a study has to detect a true effect
Statistical power
A ____ variable is one that can be measured in intervals between whole numbers and can be in the form of any unit of measurement and can take on an infinite number of possible values. (time, distance, etc.)
Continuous
____ studies are associated with relative risk
Cohort
____ studies are associated with odds ratios
Case-control
Confidence intervals reflect the ____ of study results
Precision
A ____ variable is one that assigns a number to two categories that DO NOT have intrinsic ordering (male = 1, female = 2 or visa-versa)
Nominal
Finding a positive correlation between an exposure and an outcome in a population is a good way to begin ____ generation
Hypothesis
A group is said to be victims of an exposure that led to an outcome and a positive correlation between exposure and outcome is made. However, this group was NOT actually exposed but were labeled as being exposed because they were part of a larger group to which the exposure was attributed. This is an example of ____
Ecologic fallacy
of people exposed and a # of people not exposed…FOLLOWED over time to see if outcome occurs and if exposure is associated with outcome.
Prospective cohort
of people with outcome and # of people without outcome…assessed for presence of exposure to determine if outcome is associated with exposure.
Case-control
of people exposed and a # of people not exposed…identified…their histories obtained…to see if outcome occurred and if exposure was associated with outcome.
Retrospective cohort
If people are lost during a study…any associations identified will be (increased / decreased).
Decreased
Failing to account for factors that could affect results is failing to identify ____
Confounders
‘P’ in PICO is ____
Patient, problem, population
‘I’ in PICO is ____
Intervention, exposure
‘C’ in PICO is ____
Comparison group
‘O’ in PICO is ____
Outcome
Best study to use to help identify good medications is likely a…
Double-blind randomized clinical trial
The least statistically valid study is a ____ study
Case
T/F Broadening the selection criteria for enrollment in a study will help reduce bias.
FALSE
T/F You need to know disease prevalence in a population before you can estimate predictive values.
TRUE
Publication bias (weakens / strengthens) the validity of meta-analysis and systematic reviews.
Weakens
The greater the (sensitivity / specificity) the test, the more likely the test will detect persons who HAVE the disease in question.
Sensitivity. Remember, that sensitivity is best at catching EVERYONE WITH THE DISEASE (ruling out)…and specificity is good at identifying EVERYONE WITHOUT THE DISEASE (ruling in).
