Public health science USMLE step 1
What is current recommendation of giving a meningococcal vaccine - ?All children age 11-12
y\nBooster at 16\nCan be given to children as young as 2 years if high risk
What type of rates should be associated with Incidence? - Attack rate---you shrink the
denominator as per the exposure.
Create a difference 'attack rate' for every food that was eaten in a restaurant, saw, and the one with the
highest rate is the culprit. (Acute conditions)
What are most disease rates in the US for most all diseases? - Less than 50 per 100,000.... Almost
ALL are below 100 per 100,000
What number is in the denominator for the RATE of infant mortality? - Per 1000 live births
What is the typical RATE given in epidemiology for most cases? - Per 100,000.
What is an INCIDENCE rate? - The rate of NEW cases in a population divided by people at risk to
be a case
What is a PREVALENCE rate? - The rate of ALL cases in a population divided by people at risk to be
a case.
How do you remember was an incidence rate is? - IN-cidence is the rate of new IN-coming cases
What two prevalences are associated with prevalence rate? - Point and Period prevalence
What is point prevalence? - A particular INSTANCE in time....Who has the disease over who is at
risk.
,What is period prevalence? - Period is for a larger period of time, who has it over this period over
who is at risk
Mortality data is by definition what type of date? - Incidence data....the number of newly dead
What is the general relationship between prevalence and incidence? - Prevalence = incidence x
duration....
If either incidence or duration increase, so does prevalence.
What is a CRUDE rate? - A rate that is generalized to an entire population, ie death rate in the US
in 2000
What is a SPECIFIC rate? - One that has been modified to have some sort of qualifier, ie male
death rate in 2000... This is GENDER SPECIFIC
What is a CASE fatality rate? - The number of deaths from a specific cause (heart disease) divided
by the number of patients with that disease.
What is the Proportionate mortality rate? - Deaths from a cause (heart disease) divided by all
deaths.
Sensitivity detects what?? - DISEASE! The calculation deals with diseased people. They actually
HAVE the disease.... So it is true positives divided by the rest, True positives over (TP and the false
negatives)
Specificity detects what?? - HEALTHY people!! True negatives over all people without the disease
(TN+ FP)
What does PPV detect? - The value of a positive test result and of course is given by the TRUE's
divided by everything in the row. True positives divided by all positives, TP AND FP
,What does NPV detect - The value of a negative test result, given by the TRUE negatives divided
by all the negatigves, TN and FN
How do you calculate the ACCURACY of a test? - TP + TN over EVERYTHING, TP FP TN and FN
As prevalence increases what happens to positive and negative predictive values? - PPV goes up.
More people have the disease, thus a + test result is more believable.
NPV goes down. More people have the disease, fewer people are without it so a negative is less likely.
There are NO direct changes on a screening test based on increased or decreased incidence.... -
Dur.
What is the relationship between pretest probabilities of sensitivity and sensitivity WRT prevalence? -
NONE...Pretest values are constant and are a measure of the test itself, and will never change.
What point of the double hump diagram gives you the FEWEST FALSES or the highest accuracy for a test?
- The point where the two curves cross...To move either direction gives you a few fewer falses in
one category but a whole bunch more falses in another direction.
What is the FIRST thing you need to do when constructing a 2x2 table for the USMLE? - Pick a
total population number and anchor is with PREVALENCE before plugging in numbers. Prevalence is the
number of people in the population with disease, so for 100 people and prevalence of 20 % is 20 with
disease 80 without....Then plug in the corresponding Sensitivities and Specificities.
What are the four study design questions asked on USMLE questions? - What type of bias is there,
if any?
How would you fix the bias, if any?
What type of study design is this?
How would you analyze data from this study?
Reliability is equal to - Precision... same result every time
CAN have no validity tho.
, Validity is equal to - Accuracy....
Validity implies.... - Reliability... but the opposite is not always true.
What is a selection bias? - Where there is a nonrandom assignment to the study group... ie to
study the fitness of people in NY you go to 10 local gyms in NY and take measurements....
What is selection bias AKA? - Sampling bias
Berkson's bias (using hospital records to estimate population parameters)
Non-respondent bias--ALWAYS a problem, because this is based on people who VOLUNTEER, which is
mandatory for study groups.
(a convenience sample)
What is the way to fix selection bias? - Take a random sample or selection and weight the data to
make sure it matches the population if there are any discrepancies
What is measurement bias? - Asking leading questions that suggest in word voice tone what the
bias of the asker of the questions is.
What is the Hawthorne effect? - The fact of observation ITSELF changes the activity of those
observed...To meet the expectations of the researcher
How do you FIX the Hawthorne effect? - Have a control group!---
What is the expectancy effect AKA? - The Pygmalion effect
What is the Pygmalion effect? - Occurs when a researcher's beliefs in the efficacy of treatment
changes the outcome of that treatment.
What is current recommendation of giving a meningococcal vaccine - ?All children age 11-12
y\nBooster at 16\nCan be given to children as young as 2 years if high risk
What type of rates should be associated with Incidence? - Attack rate---you shrink the
denominator as per the exposure.
Create a difference 'attack rate' for every food that was eaten in a restaurant, saw, and the one with the
highest rate is the culprit. (Acute conditions)
What are most disease rates in the US for most all diseases? - Less than 50 per 100,000.... Almost
ALL are below 100 per 100,000
What number is in the denominator for the RATE of infant mortality? - Per 1000 live births
What is the typical RATE given in epidemiology for most cases? - Per 100,000.
What is an INCIDENCE rate? - The rate of NEW cases in a population divided by people at risk to
be a case
What is a PREVALENCE rate? - The rate of ALL cases in a population divided by people at risk to be
a case.
How do you remember was an incidence rate is? - IN-cidence is the rate of new IN-coming cases
What two prevalences are associated with prevalence rate? - Point and Period prevalence
What is point prevalence? - A particular INSTANCE in time....Who has the disease over who is at
risk.
,What is period prevalence? - Period is for a larger period of time, who has it over this period over
who is at risk
Mortality data is by definition what type of date? - Incidence data....the number of newly dead
What is the general relationship between prevalence and incidence? - Prevalence = incidence x
duration....
If either incidence or duration increase, so does prevalence.
What is a CRUDE rate? - A rate that is generalized to an entire population, ie death rate in the US
in 2000
What is a SPECIFIC rate? - One that has been modified to have some sort of qualifier, ie male
death rate in 2000... This is GENDER SPECIFIC
What is a CASE fatality rate? - The number of deaths from a specific cause (heart disease) divided
by the number of patients with that disease.
What is the Proportionate mortality rate? - Deaths from a cause (heart disease) divided by all
deaths.
Sensitivity detects what?? - DISEASE! The calculation deals with diseased people. They actually
HAVE the disease.... So it is true positives divided by the rest, True positives over (TP and the false
negatives)
Specificity detects what?? - HEALTHY people!! True negatives over all people without the disease
(TN+ FP)
What does PPV detect? - The value of a positive test result and of course is given by the TRUE's
divided by everything in the row. True positives divided by all positives, TP AND FP
,What does NPV detect - The value of a negative test result, given by the TRUE negatives divided
by all the negatigves, TN and FN
How do you calculate the ACCURACY of a test? - TP + TN over EVERYTHING, TP FP TN and FN
As prevalence increases what happens to positive and negative predictive values? - PPV goes up.
More people have the disease, thus a + test result is more believable.
NPV goes down. More people have the disease, fewer people are without it so a negative is less likely.
There are NO direct changes on a screening test based on increased or decreased incidence.... -
Dur.
What is the relationship between pretest probabilities of sensitivity and sensitivity WRT prevalence? -
NONE...Pretest values are constant and are a measure of the test itself, and will never change.
What point of the double hump diagram gives you the FEWEST FALSES or the highest accuracy for a test?
- The point where the two curves cross...To move either direction gives you a few fewer falses in
one category but a whole bunch more falses in another direction.
What is the FIRST thing you need to do when constructing a 2x2 table for the USMLE? - Pick a
total population number and anchor is with PREVALENCE before plugging in numbers. Prevalence is the
number of people in the population with disease, so for 100 people and prevalence of 20 % is 20 with
disease 80 without....Then plug in the corresponding Sensitivities and Specificities.
What are the four study design questions asked on USMLE questions? - What type of bias is there,
if any?
How would you fix the bias, if any?
What type of study design is this?
How would you analyze data from this study?
Reliability is equal to - Precision... same result every time
CAN have no validity tho.
, Validity is equal to - Accuracy....
Validity implies.... - Reliability... but the opposite is not always true.
What is a selection bias? - Where there is a nonrandom assignment to the study group... ie to
study the fitness of people in NY you go to 10 local gyms in NY and take measurements....
What is selection bias AKA? - Sampling bias
Berkson's bias (using hospital records to estimate population parameters)
Non-respondent bias--ALWAYS a problem, because this is based on people who VOLUNTEER, which is
mandatory for study groups.
(a convenience sample)
What is the way to fix selection bias? - Take a random sample or selection and weight the data to
make sure it matches the population if there are any discrepancies
What is measurement bias? - Asking leading questions that suggest in word voice tone what the
bias of the asker of the questions is.
What is the Hawthorne effect? - The fact of observation ITSELF changes the activity of those
observed...To meet the expectations of the researcher
How do you FIX the Hawthorne effect? - Have a control group!---
What is the expectancy effect AKA? - The Pygmalion effect
What is the Pygmalion effect? - Occurs when a researcher's beliefs in the efficacy of treatment
changes the outcome of that treatment.
