Biostatistics Flashcards
What will you look at if you’re wondering the frequency of a disease in a population? 3 things
prevalence
incidence
attach rate
What should you look at when asking how well a test differentiates sick from healthy?
sensitivity
specificity
What measure will answer this question: of those in a population who test as sick or healthy, how many are truly sick or healthy?
predictive values
What concepts should you look at when asking what is the impact of a treatment?
risk reduction
number to treat or harm
What are the three general types of prevalence?
point prevalence (at one time) period prevalence (during a period of time) lifetime prevalence (over a course of a lifetime)
What is the equation for prevalence?
number of people with a disease divided by number of people at risk for the illness at that point in time.
can change depending on how you define the at risk
What is indicence?
the number of NEW cases of a disease during a time period over the number of people at risk for the illness during that time period
True or false: incidence excludes those that already had it.
true - because you’re not at risk for being a new case
What are two subtypes of incidence?
cumulative incidence - total number reported over time
attack rate - refers to outbreaks
What is attack rate?
It’s a type of incidence used when the nature of the disease is acute and the population is observed for a short period of time
so outbreaks
What is the equation for attack rate
number of new cases / number of exposed
exposure is what’s really important
WHat is secondary attack rate?
It’s the number of new cases over the number of exposed minus the primary cases in the primary attack rate
so it’s a measure of disease frequency among contacts of those affected by the first exposure
What will duration of illness do to prevalence?
longer illness, higher prevalence
What will migration do to prevalence?
if you have lots of sick people moving into the area, the prevalence will increase
if healthy people move out of an area, prevalence will increase
What will prevention do to prevalence or incidence?
should decrease incidence
What cause the huge jump in AIDs incidence in 1993?
a change in diagnostic criteria
What is the equation relationship between prevalence and incidence?
prevalence = incidence x (average) duration
When will prevalence roughly equal to the incidence?
when the illness is super acute like the flu
What is sensitivity?
the probability that a diseased person will be identified correctly by a diagnostic screening test
What are other terms for sensitivity?
true positive probability or true positive rate
What’s the equation for sensitivity?
true positives/total # of ill people (true positives and false negatives)
What is specificity?
the probability that a well person will be identified correctly by a diagnostic screening test
aka true-negative probability
What’s the equation for specificity?
true negatives / total # of well people (true negatives and false positives)
What does it mean to have a highly sensitive, but not highly specific test?
It helps you catch all cases if you really want to be cautious and catch something early - if letting something go would be fatal
errs on the side of over-diagnosing
good screening test
What does it mean for a test to have high specificity but low sensitivity?
Good for making sure that the person really does have the disease - good if the treatment has lots of risks
errs on the side of under-diagnosing
good for confirmation test
What is the predictive value?
probability that a test will give a correct diagnosis
WHat are the two types of predictive value?
positive predictive value
negative predictive value
What are the three thigns that a predictive value will depend on?
- test sensitivity
- test specificity
- prevalence of the disease in the population being tested
What is positive predictive value?
probability that a person who tests positive actually has the disease
true positives/all positives (true and false positives)
What is negative predictive value?
probability that a person who tests negative actually is healthy
true negatives/all negatives (true and false negatives)
How does higher disease prevalence affect predictive values?
it makes the positive test more likely to be actually positive = increases positive predictive value
decreases negative predictive value
How does lower disease prevalence affect predictive values?
it makes the negatives more likely to actually be negative
increases negative predictive value
decreases positive predictive value
In a randomized control trial, what is the control event rate
proportion of the control group who have a bad outcome after “treatment”
If 10 of 30 control group participants become sicker,
CER = 10/30 = 0.33 = 33% had adverse outcomes
What is the experimental event rate
proportion of treatment group participants who have a bad outcome after treatment
For example…
If 4 of 30 treatment group participants become sicker,
EER = 4/30 = 0.13 = 13% had adverse outcomes
What is absolute risk?
the probability of developing a disease or undesired outcome
How do you determine absolute risk reduction?
when the control event rate is higher than experimental evenr ate
CER - EER > 0
How do you determine absolute risk increase?
when the control event rate is lower than experimental event rate
CER - EER < 0
How do you determine the number needed to treat? (# of patients treated in order to save one person)
inverse of absolute risk reduction
NNT = 1/ARR
How do you determine number needed to harm? (# of people affected in order to harm 1 person)
inverse of absolute risk increase
NNH = 1/ARI