Pithua Flashcards
Epidemiology is based on what 2 fundamental assumptions?
- disease does NOT occur at random, rather there are factors/determinants that can increase or decrease the likelihood of disease
- the factors or determinants can be identified by systematic investigation of populations or subgroups within populations
__________ is the study of the distribution and determinants of disease frequency in human/animal populations and the application of this study to control health problems.
epidemiology
when studying human or animal populations, the group of subjects have common …
characteristics
such as: geographic location, age, group membership, breed, parity, housing, production levels, etc.
how do you measure disease frequency?
counting the number of cases of a disease in a population over a specific period of time
What is meant by studying the “distribution” of disease frequency?
how the disease is distributed through the population
Ex: age, sex, socioeconomic status, breed, location, time
_________ are factors that cause some individuals to acquire disease and/or factors that prevent some individuals from getting disease. These are also referred to as “exposures”.
determinants
What is the first step of the epidemiologic approach?
Make observations – suspect that exposure influences disease occurence
What is the second step of the epidemiologic approach?
form specific hypotheses about exposure-disease link
what is the third step of the epidemiologic approach?
perform epidemiologic studies to measure the relationship between exposure and disease
What is the fourth step of the epidemiologic approach?
judge if the association is valid and casual – factor in: evidence, chance, bias, cofounding, limits of study design, strength of study design
what is the fifth/last step of the epidemiologic approach?
evaluate prevention and treatment
What are the 2 criteria for a GOOD hypothesis?
- project expected association between 2 or more measurable variables
- carry clear implications for testing stated relations
__________ are important for advancing epidemiologic research and guiding study design, variable selection, sample selection, data analyses, and interpretation of results.
hypotheses
“Smoking causes cancer” is an example of what type of hypothesis?
fundamental
“People who smoke cigarettes daily have a higher incidence of lung cancer over a 10-year period than people who do not smoke cigarettes” is an example of what type of hypothesis?
operational
T/F: we must know the specific cause of a disease in order to intervene, that is why hypotheses are so important to guide our experiments.
false – we do NOT need to know a specific cause to intervene.
This type of epidemiology is described as:
1. describes patterns of disease
2. monitor public health, evaluate success of disease intervention programs, generate hypotheses about causes of disease, and monitor herd health
3. identify and count cases of disease in populations and conduct simple studies
descriptive epidemiology
This type of epidemiology is described as:
1. search for disease cause and prevention
2. evaluate hypotheses about causes of disease and evaluate success of intervention programs
3. compare groups and systematically determine if there is an association present
analytical/scientific epidemiology
what are the 3 main uses of descriptive epidemiology in veterinary medicine and public health?
- ID problems, trends, and high-risk groups
- planning: where to spend resources
- generate hypotheses for analytical epidemiology
what is the limitation of descriptive epidemiology?
cannot identify cause of disease
what are the 4 components for measuring disease frequency?
- population
- cases of disease (numerator)
- size of population (denominator)
- time
A ______ population is a population whose membership is defined on the basis of some event and membership is permanent (ex. veteran of vietnam war or people born in 1982)
fixed
A ________ population is where membership is defined by being in a state or condition, but membership is transient. (ex. being a student at a certain school, being a parent of a teenager)
dynamic
What are some examples of methods of disease ascertainment?
clinical records
diagnostic tests
disease registries
surveillance programs
self-reports
What factor is necessary to compare disease across populations?
size of the population you sampled from. you cannot compare the number of cases alone
When measuring disease frequency, this type of measure is the division of one number by another and these numbers do NOT have to be related.
ratio
ex. observed # cases of rabies in bburg during aug / expected # cases of rabies in bburg during aug
if there were 40 cases but you expected 20, then the ratio is 40/20 or 2. This means that there were 2 times as many cases of rabies in bburg during august than was expected.
From june to august, you collected 1000 blood samples from dogs at your clinic. 120 of them were positive for Lyme disease. What is the proportion of patients infected with lyme?
12% of patients were infected with lyme.
120/1000 = 0.12
Proportion is the division of two RELATED numbers in which the numerator is a SUBSET of the denominator.
You analyze data and determine that there are 100 cases of rabies per 100,000 cats per year.
What is the rate of rabies in cats per year?
Rate is the division of one number by another in which TIME is an intrinsic part of the denominator.
Therefore, rate is the # of NEW cases of rabies among 100,000 cats over a one year period.
So, 100 cases/100,000 cats per year.
_________ measures the presence of EXISTING cases of disease in a population during a specific time period.
The denominator will include the TOTAL population (healthy, sick, etc.) and it is the proportion of the population of HAVE the disease.
Prevalence
___________ is the proportion of a population that has the disease at a single point in time.
point prevalence
What is the prevalence of bordetella in Blacksburg if 89 dogs test positive for bordetella on January 1st? Note, the total population of dogs in bburg is 40,000 on Jan 1st.
0.002
You have a clinic with 25 canine patients. On monday, 5 of them have upper respiratory signs. By thursday, 3 have recovered. On saturday, they have all recovered. What is the prevalence of upper respiratory infection on thursday? what about saturday?
thursday – 2/25
saturday – 0
_________ measures the occurrence of NEW cases of disease in a population during a specific time period. This involves TRANSITION from one state to another (healthy –> ill, alive –> dead). The denominator only includes the population AT RISK (excludes those who are already measured and/or immune).
incidence
What are 2 reasons for measuring prevalence?
- assess burden of disease (plan & allocate resources)
- used for research when incidence is difficult to measure (uncertain onset)
What are 3 reasons for measuring incidence?
- etiologic research – find the cause of the dz
- evaluate preventions – are they working?
- evaluate treatment – is it working to improve survival or QOL?
___________ incidence is when time is described in words that go along with the number. This measures the occurrence of NEW cases of disease in a population during a specified time period, assuming you have a fixed population/ no loss to follow-up.
cumulative
___________ measures SPEED at which NEW cases of disease occur within a population. Time is an intrinsic part of the denominator. It does NOT assume the group being observed has been followed the entire time period.
incidence rate
You have a clinic with 25 canine patients. On monday, 5 of them have upper respiratory signs. By thursday, 3 have recovered. On saturday, they have all recovered. What is the cumulative incidence of upper respiratory infection throughout the week?
5/25 = 20%
__________ is a special type of cumulative incidence that measures the proportion of individuals exposed to an infectious agent who become infected over a certain time period.
attack rate
________- is a special type of cumulative incidence that measures the proportion of individuals with a disease who die of that disease.
case fatality rate
What are the biggest limitations of measuring cumulative incidence?
- not a perfect measure in a dynamic population or in a fixed population that loses members over time.
- doesnt consider time of occurrence
What is a person-time?
the amount of time each person AT RISK is under observation. this is the unit of measure for incident rate.
You followed 30,330 dogs from 1999-2004. There were 2,314 cases of bordetella during follow-up. What is the incidence rate?
Incidence rate is: # new cases during this time period / total person-time of observation in population at risk
IR = 2,314 / 104,574 = 0.022/PY = 2,215/100,000 PY
during 1999-2004, the incidence rate of bordetella was 2,215/100,000 PY.
T/F: to be a prevalent case, an individual has to have been an incident case at some point and must still have the disease so that duration is involved
true
What formula can you use to estimate the average duration of a disease?
D = P / IR
P is prevalence
IR is incidence rate (PY)
What would happen to the incidence of rabies if you introduced primary prevention efforts to reduce exposure to rabies?
(increase, or decrease, or no effect)
incidence would decrease
What would happen to the prevalence of rabies if you introduced primary prevention efforts to reduce exposure to rabies?
(increase, or decrease, or no effect)
prevalence would decrease
What would happen to the incidence of FIP if you introduced new treatment that prolongs the life of patients with FIP?
(increase, or decrease, or no effect)
there would be NO effect on incidence because the cases are going to occur regardless of treatment
What would happen to the prevalence of FIP if you introduced new treatment that prolongs the life of patients with FIP?
(increase, or decrease, or no effect)
the prevalence would increase because you are not preventing cases from occurring, you are only treating them and prolonging their lifespan with disease.
Which type of disease frequency is the following:
The % of dogs that contract influenza during the course of an outbreak (June 1-15, 2022).
A. prevalence
B. incidence rate
C. cumulative incidence
C. cumulative incidence
Which type of disease frequency is the following:
The % of potential breeding dogs rejected from service dog breeding program because of poor hip conformation.
A. prevalence
B. incidence rate
C. cumulative incidence
prevalence
Which type of disease frequency is the following:
The number of UTIs diagnosed in 1,000 person-years within a veterinary hospital.
A. prevalence
B. incidence rate
C. cumulative incidence
incidence rate
Which type of disease frequency is the following:
The % of live-born foals with flexural deformities among 100,000 live births.
A. prevalence
B. incidence rate
C. cumulative incidence
prevalence
T/F: the total population contributes to the denominator of cumulative incidence
false – only population at risk
T/F: when you calculate incidence rate of disease, you are required to follow subjects for the same amount of time
false
_____________ is a device or procedure with the ability to differentiate a diseased individual from a non-diseased individual
diagnostic test
Diagnostic tests are used for what type of patients in order to confirm a diagnosis, assess the severity of a disease, determine a prognosis, and assess the progress/response to treatment?
individual sick patients
____________ are used in healthy populations or individuals to detect carriers or subclinical disease. They are commonly used in disease control and prevention, detection and monitoring, as well as eradication programs.
screening tests
A test is considered useful if the result …
has the potential to alter your diagnosis or course of treatment/action.
Necropsy, bacteria culture, exploratory sx, and biopsies are considered “gold standard” tests that are accurate in determining presence or absence of disease. The downside to these tests are …
- complex
- costly
- risky
- time-consuming
T/F: when evaluating a new diagnostic test, we gather data comparing the results of the gold standard test against the results of the new test. The sample population (group that the test was evaluated with) should reflect the target population (group that you will be using test in real life).
true
_________ is the proportion of all animals sampled that TRULY HAVE the disease.
In the 2x2 table, this is a+c / (a+b+c+d)
(test pos + dz pos / total tested)
true prevalence.
_________ is the proportion of all animals sampled that appear to be infected based on a positive result from the test kit.
In the 2x2 table, this is *a+b / (a+b+c+d)
(total test pos / total)
apparent prevalence
When comparing a new diagnostic test and the gold standard test, if the apparent prevalence is LESS than true prevalence, what conclusion can you draw?
the new test is underestimating prevalence of disease.
__________ is the true positive rate or the proportion of truly diseased animals that test positive using the test kit.
In the 2x2 table, this is a / (a+c)
(test pos and dz pos / total that have dz)
sensitivity
________ is the true negative rate or the proportion of truly uninfected animals that test negative using the test kit.
In the 2x2 table, this is d / (b+d)
(test neg and dz neg / total that do not have dz)
specificity
_______ is measuring how close the test results are to the gold standard.
test validity
_________ is the proportion of all animals (diseased and non diseased) tested that were CORRECTLY classified by the test kit.
In the 2x2 table, this is (a+d) / (a+b+c+d)
(test pos with dz + test neg without dz) / total)
test accuracy
_____________ is assessing the consistency or repeatability of a test. If we repeat a test on the same sample, how often would we get the same answer?
Test precision
T/F: if a test is NOT precise, it CANNOT be accurate
true
T/F: A test can be very precise, but still inaccurate
true
__________ is the probability that a test result reflects the TRUE disease status. I.e. if we get a positive result, how confident are we that this animal is TRULY infected/diseased (PPV) and vice versa (NPV)
predictive value
What is the equation to calculate positive predictive value based on the 2x2 table?
a / (a+b)
in other words, the individuals who have disease AND tested positive divided by the total individuals that tested positive.
What is the equation to calculate negative predictive value based on the 2x2 table?
d / (c+d)
in other words, the individuals who do NOT have disease AND tested negative divided by the total individuals that tested negative.
Positive predictive values are affected by sensitivity and specificity. Choose the correct answer for the following scenarios:
sensitivity: if you have fewer false negatives –> (higher or lower NPV)
specificity: if you have fewer false positives –> (higher or lower PPV)
sensitivity: if you have fewer false negatives –> higher NPV
specificity: if you have fewer false positives –> higher PPV
Which of the folllowing can change dramatically and have an impact on predictive values?
a. sensitivity
b. specificity
c. prevalence
c. prevalence
If the prevalence of a disease is HIGH, what effect will that have on the positive and negative predictive values?
PPV will be higher – if you get a positive test, you are MORE confident that it is correct due to the disease being more prevalent.
NPV will be lower – if you get a negative test, you are LESS confident that it is correct due to the disease being so prevalent
T/F: we can improve positive predictive value by testing only individuals from high risk / high prevalence populations
true
because as prevalence increases, PPV increases
T/F: we can improve predictive value by altering sensitivity and specificity through changing the conditions of the test or how it is interpreted (change cut points, etc.)
true
because this will change the threshold for determining whether an animal tests positive or negative
T/F: We can improve predictive value of a test by using individual diagnostic tests rather than multiple to make the results more clear.
false
we should do sequential testing or concurrent testing –
do the sensitive test first (to avoid false negatives)
then do a specific test only on the individuals that tested positive on the first test (avoids false positives)
T/F: outbreaks are sudden events that occur spectacularly
false – they can be sudden, but they can also be insidious (silent outbreaks)
T/F: if you see multiple clinical signs in an individual animal/person, there tends to be a single etiology. And, a single syndrome in a population tends to have multiple etiologies.
true
_____________ problems are a combination of errors over time.
additive or cyclic
_____________ problems are long-acting errors that require the passage of time before the consequences become of sufficient magnitude to be recognized.
chronic
When attempting to determine the problem within an outbreak, you aim to characterize cases in the form of a ____________ which will evolve into a case definition.
working definition
In order to verfiy the outbreak, what needs to be confirmed?
is there a true excess of disease – obtain objective data to verify the magnitude of the problem. then, compare the actual number of cases to the expected number to determine if the frequency is “excessive”
T/F: if you cannot make a definitive diagnosis, it is not possible to create a case definition
false – you can!
if there is a def dg, it looks like this: pasteurella pneumonia in recently weaned calves
if there is not a def dg, it looks like this: acute respiratory disease in recently weaned calves
T/F: a case definition is good because it usually establishes the pathologic or etiologic diagnosis, but it does not solve the producers problems.
true
it does not solve anything because the real question is WHY the problem is occurring.
How do you calculate an attack rate?
individuals with dz / # individuals exposed
How do you calculate a risk ratio?
incidence risk in exposed / incidence risk of unexposed
How do you calculate attributable fraction?
AF = Re - Ro / Re
Re is a/a+b
Ro is c/c+d
why do we describe an outbreak according to place?
Because different groups or pens of animals often have different levels of exposure. we want to figure out the affected versus unaffected animals and their location to see if it gives us clues
What 2 things do we want to consider when we are describing an outbreak according to time?
- when (in calendar time) did the problem begin
- what is the pattern of performance over time
What are pseudoepidemics?
onset of producer awareness of a more chronic problem
Describe a common point source epidemic curve
many cases during during/around one time period pointing towards a common source being the issue.
Describe a continuous source epidemic curve
cases are not just appearing one time, they are continually appearing indicating that continual exposure to the source.
Describe a propagated epidemic curve
you see several case peaks over time with a common source but resulting from person to person spread, not as a result of the source just being present.
T/F: you should generate your hypothesis about the nature of exposure BEFORE you develop quantitative information (who, when, where, etc.)
false – quantify first, then generate a hypothesis
how can you implement interventions as a part of your outbreak investigation?
- give hands-on practical advice on how to decrease the impact of this problem
- monitor progress to make sure the intervention is working
- write written reports of your findings
if you want to compute a measure of effect, what measure of disease frequency should you use?
you should use an absolute comparison: risk/rate difference
If you calculate a relative comparison such as a risk ratio, rate ratio, and/or an odds ratio, what is this comparison called?
hint: a measure of __________.
measure of association
___________ measures such as risk ratio or odds ratio describe the strength of a causal relationship between exposure and the disease.
relative measures
A study is considered “valid” ONLY when what 3 alternative explanations have been eliminated?
- bias
- confounding
- random error
________ is a systematic error in the design or conduct of a study that results in an erroneous association between the exposure and the disease or obscures an association that actually exists.
bias
_________ is mixing of effects between exposure, disease, a confounder. it distorts the relationship between an exposure and disease.
confounding
__________ is the probability that an observed study result is due to chance (no assignable cause).
random error
T/F: is bias, confounding, and random error are ruled out as alternative explanations of an observed exposure-disease association, then the association is true and the study is “internally valid”
true
T/F: if a study has demonstrated an association between an exposure and a disease, then you can assess whether the exposure caused the disease.
false – only after the observed study is considered VALID is it appropriate to assess whether the exposure has caused the disease.
you have to establish internal validity before you could apply to a generalized population.
__________ is when the observed results accurately reflect the true association.
internal validity
__________ is error due to a recognizable source.
systematic error
T/F: it is possible to have both random and systematic error occurring at once.
true
If the direction of a bias is TOWARDS the null
(RR=1 or RD=0), then it means the bias is …
masking an association that is truly present
aka it is underestimating the true measure of association
_________ is when the observed effect is HIGHER than the true value.
positive bias
________ is when the observed effect is LOWER than the true value.
negative bias
If the direction of a bias is AWAY from the null, then it means the bias is …
overestimating the true measure of association.
ex. overestimating the effectiveness of a vaccine or overestimating the risk of drinking alcohol on liver failure
If an exposure RR is greater than the null (1), what does this indicate about exposure?
exposure increases the risk of the outcome
ex. drinking alcohol increases risk of liver failure
If an exposure RR is less than 1, what does this indicate about the exposure?
exposure decreases the risk of the outcome
ex. vaccines (are protective against dz)
T/F: if you detect bias in your study, you can correct it before you continue
false – little can be done to correct or remove bias once it has occurred so it should be carefully avoided when designing and conducting a study
___________ is a design method to deal with confounding that limits the study to people who are WITHIN one category of the confounder
(ex. only selecting individuals of the same sex or age)
Restriction
T/F: restriction is a good way to deal with confounding, however it does limit the sample size, which limits generalizability, and it is possible that incomplete control occurred and residual confounding is present.
true
________ is a design method to deal with confounding that selects study subjects so that confounders are distributed identically among exposed and unexposed or cases and controls. For example, if your potential confounding variables are age and sex, then your cases are 35yo males, then your controls should also be 35yo males.
Matching
_________ is a design method to deal with confounding in which participants are allocated to treatment groups by a random method that gives EQUAL chance of being in any treatment group. This ensures baseline comparability of groups and controls for unknown and known confounders (with no limit).
randomization
Which type of bias can be (to some extent) adjusted or controlled for in the analysis?
confounding
_______ bias has to do with WHO gets into your study and results from the procedures used to select subjects. Essentially, it arises when/if people who agree to take part in the study are different from the source population you want to study, and therefore the study population will NOT represent the source population.
selection bias
in what studies is selection bias MOST likely to occur in?
- case control studies
- retrospective studies
because outcomes and exposures have already occurred when you are selecting the subjects. - experimental cohort studies
- prospective cohort studies
due to differential loss to follow-up
If you select participants of a study in a manner related to developing the outcome of interest OR participation of individuals if differs for exposure and disease, what will occur in your study?
selection bias
you will get an Risk Ratio estimate that is NOT consistent with the truth as a consequence of how you selected individuals to investigate
What is the healthy worker effect?
it is when you select the general population as a comparison group in an occupational cohort study
__________ bias occurs if controls are MORE or LESS likely to be selected if they are exposed or unexposed. It occurs when controls fail to represent the exposure distribution in the source population from which the cases arose because controls do not accurately represent the same source population as the cases.
For example, if you are attempting to determine if SES is assoc. with prostate cancer – if your exposed pop is men with prostate cancer, but your control group is individuals from a door-to-door survey during a workday, then you have control-selection bias because unemployed or retired men are mostly at home during those hours (not representative of a true control sample).
control-selection
What is the equation to calculate odds ratio?
OR = AD / BC
this is comparing the exposed to the unexposed
If your research question aims to determine if women who took antidepressants during pregnancy increases their risk of having babies with birth defects, what could a potential differential participation/self-selection bias be?
that mothers who took antidepressants during pregnancy are MORE willing to participate in this study vs other mothers.
How can you prevent differential participation/self-selection bias in your case-control studies?
have a high participation rate for ALL groups
If your controls have an unusually LOW prevalence of the exposure, then “b” in in 2x2 table will tend to be smaller. This will bias the odds ratio (toward or away) from 1?
AWAY
it will OVER-estimate the odds ratio
If your controls have an unusually HIGH prevalence of the exposure, then “b” in in 2x2 table will tend to be large. This will bias the odds ratio (toward or away) from 1?
TOWARD
it will UNDER-estimate the odds ratio
___________ occur when study participants exit the study for reasons related to both exposure and disease
differential loss-to-follow up
_________ is bias that can occur if study participants are placed into the wrong exposure or disease category
measurement/misclassification error
Non-differential/ random measurement/misclassification errors are about the same in both groups and tend to have what effect on the true difference between the groups
dilute/mask/obscure
A ____________ measurement/misclassification error is when information is better in one group than another and the association may be over or under estimated.
differential/non-random
____________ bias is error that can occur if the information you collect from or about study participants is erroneous
information
__________ bias is a type of information bias that is error occurring when people with disease remember/recall/report their exposure differently (more or less likely).
recall
How can you resolve/prevent information recall bias?
use controls who are also diseases to promote comparable recall.
use standardized, close-ended questionnaries to promote consistency and specificity
examine pre-existing data
_________ bias is a type of information bias that occurs if there is a systematic difference in soliciting, recording, or interpreting information or control status. In other words, the person recording/interpreting information is influenced by the participants treatment or exposure status.
interviewer
How can you prevent interviewer information bias?
use blinding/masking to prevent interviewers from seeing the hypothesis or knowing someones exposure status
use high qual standardized closed ended questionnaire to promote consistency and specificity
examine pre-existing data
If your research question wants to determine an etiology, which study designs are BEST to carry out?
- cohort
- case-control
If your research question wants to determine the effect of a treatment, what study design is BEST to use?
randomized controlled trial
if your research question wants to determine the prognosis of a exposure, what study design is BEST to use?
cohort
if your research question wants to determine the harm/effect of an exposure, what study designs are BEST to use?
- cohort
- case-control
if your research question wants to determine a diagnosis, what study designs are BEST to use?
- cross-sectional
- case-control
A well-built clinical research question has what 4 architectural components?
hint: the abbreviation is PICO
- patient or population (specific group of interest)
- intervention/exposure
- comparison intervention/exposure (control group)
- outcome
If the investigator assigns exposures, what kind of study are your performing – experimental or observational?
experimental
if the investigator is performing an experimental study and chooses to do random allocation, what is this study design called?
randomized controlled trial (RCT)
If the investigator does NOT assign exposures, this is an observational study. If within this observational study, the investigator includes a comparison group, what type of study does this become?
analytical study
(as opposed to a descriptive study that would NOT have a comparison group)
What analytical study looks at the exposure before looking at the outcome?
cohort study
What analytical study looks at the outcome first then the exposure?
case-control study
What analytical study looks at the exposure and outcome at the same time?
cross-sectional study
T/F: a descriptive study investigates and describes the distribution of disease within a population by using routinely collected data (person, place, and time) and attempts to explain what is driving the outcome.
false – all is true except the explanation of the outcome. this study simply outlines the data.
__________ studies are analytical studies that describe disease frequency in different populations in relation to exposures within those populations.
ecological
A _____________ study gathers information from individuals to measure prevalence at one point in time of the individuals exposed vs the unexposed individuals.
descriptive cross sectional study
what is the difference between a descriptive cross sectional study and an analytical cross sectional study?
analytical goes one step further to calculate a prevalence ratio between the 2 prevalences.
So, it calculates the prevalence of the exposed / prevalence of unexposed.
what is the starting point for a case-control study?
what about a cohort study?
case-control studies starts with a group of individuals WITH the disease/outcome AND a control-group.
cohort studies define a group of people by their exposure status and then follow them over time to see which individuals develop the disease to compare the incidence of disease.
which study design is best for studying multiple exposures?
case-control
which study design is best for measuring time relationship between exposure and outcome?
cohort
which study design is best for the direct measurement of incidence?
cohort
which study design is best for investigations with long latent periods?
case-control