Public Health 2012: Test 1

Front 1

Circulating antibodies may mean what 4 things?

Back 1

1. Exposure without infection
2. Exposure and currently infected
3. Exposed, infected, and recovered
4. Recipient of antibodies via colostrum

Front 2

Prevalence of detectable antibodies depends on what 3 things?

Back 2

1. Rate of seroconversion
2. Half-life of antibodies
3. Length of time after seroconversion

Front 3

true vs. apparent prevalence

Back 3

* Apparent prevalence: probability that a randomly selected animal will have a positive test result

* True prevalence: the probability that a randomly selected animal has the disease

Front 4

Apparent prevalence

Back 4

* Apparent prevalence: probability that a randomly selected animal will have a positive test result

* AKA the proportion of individuals that test positive

* Will not be same as true prevalence if test is not perfect

Front 5

True Prevalence

Back 5

* True prevalence: the probability that a randomly selected animal has the disease

* AKA the proportion of individuals that have the condition

Front 6

Circulating antibodies may mean what 4 things?

Back 6

1. Exposure without infection
2. Exposure and currently infected
3. Exposed, infected, and recovered
4. Recipient of antibodies via colostrum

Front 7

Prevalence of detectable antibodies depends on what 3 things?

Back 7

1. Rate of seroconversion
2. Half-life of antibodies
3. Length of time after seroconversion

Front 8

true vs. apparent prevalence

Back 8

* Apparent prevalence: probability that a randomly selected animal will have a positive test result

* True prevalence: the probability that a randomly selected animal has the disease

Front 9

Apparent prevalence

Back 9

* Apparent prevalence: probability that a randomly selected animal will have a positive test result

* AKA the proportion of individuals that test positive

* Will not be same as true prevalence if test is not perfect

Front 10

True Prevalence

Back 10

* True prevalence: the probability that a randomly selected animal has the disease

* AKA the proportion of individuals that have the condition

Front 11

Sensitivity vs. Specificity

Back 11

* Sensitivity: the probability that an animal that has the disease will test positive

* Specificity: the probability that an animal that has the disease will test negative

Front 12

Sensitivity

Back 12

* Sensitivity: the probability that an animal that has the disease will test positive

Front 13

Specificity

Back 13

* Specificity: the probability that an animal that has the disease will test negative

Front 14

How do sensitivity and specificity vary if the cut off values are moved to the right (to a lower titer with a greater denominator)?

Back 14

* Specificity increases- higher true negatives AND lower false positives

* Sensitivity decreases- lower number of true positives AND higher number of false negatives

Front 15

How do sensitivity and specificity vary if the cut off values are moved to the left (a higher titer with a smaller denominator)?

Back 15

* Specificity decreases- lower true negatives AND higher false positives

* Sensitivity increases- higher true positives AND lower false negatives

Front 16

Positive predictive value

Back 16

probability an animal that tests positive has the disease

Front 17

Negative predictive value

Back 17

probability an animal that tests negative does not have the disease

Front 18

How does a HIGH prevalence formally influence predictive values?

Back 18

PPV is high and NPV is low. You are likely to believe a positive test result.

Front 19

How does a LOW prevalence formally influence predictive values?

Back 19

PPV is low and NPV is high. You are likely to believe a negative test result.

Front 20

4 factors affecting sensitivity estimates

Back 20

1. Immune status
2. Stage of disease or severity of disease
3. Duration of infection
4. Informal relationship between prevalence and above 3 factors (formal relationship PV and Prev)

Front 21

3 factors affecting specificity estimates

Back 21

1. Cross-reaction due to non-target components

2. Maternal antibodies and vaccine antibodies

3. May improve over time in control programs by removing test positives

Front 22

parallel vs. serial testing

Back 22

* Parallel: both tests applied up front

* Serial: one test applied, positive animals are given second test

Front 23

How do you interpret results from parallel testing?

Back 23

* Positive: if one test is positive

* Negative: if both tests are negative

Front 24

How do you interpret results from serial testing?

Back 24

* Positive: if both tests are positive

* Negative: if one test is negative

Front 25

Sensitivity and specificity of parallel testing?

Back 25

* High sensitivity
* Low specificity

** Additional tests can verify positive test results

Front 26

Sensitivity and specificity of serial testing?

Back 26

* Low sensitivity
* High specificity


** Additional tests can verify negative test results

Front 27

Veterinary public health

Back 27

the sum of all contribution to the physical, mental, and social well-being of humans through an understanding and application of veterinary science

Front 28

Epidemiology

Back 28

the study of the distribution and determinants of health related states or events in specified populations and the application of this study to control health problems

Front 29

List the main criteria for causation

Back 29

* Exposure must precede outcome
* Dose response (most cases)
* Reasonably big effect (e.g. odds ratio or risk ratio)
* Statistically significant
* Results consistent with other work
* Statistical association is NOT causation
* One study is NOT causation
* Usually a complicated combination of factors needed to cause an outcome

Front 30

Differentiate between component, necessary, and sufficient cause

Back 30

* Component: needed in some cases

* Necessary: a component causes that is a member of every sufficient cause

* Sufficient: a group of causes that makes disease inevitable

Front 31

Explain what the agent-host-environment triad represents.

Back 31

o Shows the interaction and interdependence of agent, host, environment, and time as used in the investigation of diseases and epidemics

Front 32

Differentiate between a continuous and categorical variable.

Back 32

* Continuous: can have any value (weight)

* Categorical: can have a limited number of separate categories or groups (breed)

Front 33

Differentiate between a nominal and ordinal categorical variable.

Back 33

* Nominal: data that has no inherent relationship between the categories (breed, sex, etc)

* Ordinal: data that has an ordering or ranking for the categories (BCS, severity of heart murmurs, etc

Front 34

paired vs unpaired data

Back 34

* Paired data can be: the same subject at two time points, different legs/ears/etc of the same subject


* Unpaired data: observations are independent of each other (i.e. diabetic vs. non-diabetic subjects)

Front 35

P-value

Back 35

gives statistical significance, determines how likely it is for results to be due to chance

*** P≤0.05 is considered statistically significant

Front 36

Power

Back 36

the ability to detect a difference between groups when there really is a difference, want a high power to detect a difference between groups (if there is one)

Front 37

survival analysis

Back 37

* A type of statistical analysis used for data that measures time to some event


* The “event” can be death, onset of clinical signs, discharge from the hospital, etc

Front 38

Exposure (independent) variable VS. Outcome (dependent) variable

Back 38

* Exposure: the factor of interest (the characteristic being observed or measured, x axis)

* Outcome: an event or end point of interest (y axis)

Front 39

what 3 things make a good clinical trial?

Back 39

1. There are at least two groups
( A treatment group and a control group)
2. It is randomized
3. It is blinded

Front 40

three main types of analytical observational studies

Back 40

1. Cohort studies
2. Case-control studies
3. Cross-sectional studies

Front 41

Cohort Studies

Back 41

* Exposure status is observed and the subjects are followed until the outcome status occurs

* Enrollment is based on exposure status

* Always proceeding forward from exposure to outcome (even if retrospective)

* Assess if a statistical association exists with Risk Ratio or Odds Ratio

Front 42

Case-control studies

Back 42

* The outcome status is observed and the prior exposure status is determined

* Enrollment based on outcome status

* Always proceed backward from outcome to exposure (even if prospective)

* Assess if a statistical association exists with Odds Ratio (NOT risk ratio)

Front 43

Cross-sectional studies

Back 43

* The outcome and exposure statuses are determined at the same time

* Outcome and exposure status are unknown at time of enrollment

***NOT prospective or retrospective!

Front 44

Prevalence

Back 44

The number of instances of a disease in a known population, at a designated point in time

Front 45

Incidence Risk

Back 45

The proportion of *new* cases that occur in a population over a specified period of time


** New animals cannot be added to the population under observation

* If animals are removed, the denominator is subtracted by half the number of animals removed

Front 46

Incidence Rate

Back 46

The rapidity with which new cases occur over time

Front 47

Calculate Prevalence

Back 47

(# of animals with disease at time)
_________________________
(total # of animals in population at time)

Front 48

Calculate Incidence Risk

Back 48

(# of animals diseased over time period)
_________________________
(# healthy animals in pop at beginning of time)

Front 49

Calculate Incidence Rate

Back 49

(# of new cases over time)
_________________________
(sum, across all individuals, of length of time at risk developing disease)

Front 50

Calculate Mortality Risk

Back 50

(# of animals that die from disease over time period)
_________________________
(# of animals in pop at beginning of time)

Front 51

Calculate Case Fatality Risk

Back 51

(# of animals die from disease over time period)
_________________________
(# of animals with disease over time)

Front 52

Calculate Risk Ratio

Back 52

[a/(a+b)] / [c/(c+d)]

Front 53

Calculate Odds Ratio

Back 53

(a x d) / (b x c)

Front 54

interpret confidence intervals for Risk Ratios

Back 54

* If the CI for a RR is less than 1, then the RR is significant

* If the CI for a RR is greater than1, then the RR is not significant

Front 55

Absolute Risk Difference (AR)

Back 55

 Difference in incidence of disease in exposed animals and incidence in unexposed animals

Front 56

Calculate/interpret Absolute Risk Difference (AR)

Back 56

AR= a/(a+b) – c/(c+d)


Ranges from [-1,+1]
• AR < 0 exposure is protective
• AR = 0 exposure has no effect
• AR > 0 exposure is positively associated with disease

Front 57

Attributable Fraction (AF)

Back 57

the % of outcomes in the exposed group that are attributable to the exposure

Front 58

Population Absolute Risk Difference (PAR)

Back 58

PAR is the % of population that develop the disease due to the exposure

Front 59

Population Attributable Fraction (PAF)

Back 59

PAF is the % of disease in the population could be prevented if the exposure was removed

Front 60

Number Needed to Treat (NNT)

Back 60

Number of animals needed to be treated (vaccinated) to prevent one case

Front 61

Number Needed to Harm (NNH)

Back 61

Number of animals needed to be exposed to cause harm in one animal

Front 62

3 main types of bias

Back 62

1. Selection bias
2. Information bias
3. Confounding bias

Front 63

Selection bias

Back 63

usually results from comparative groups not coming from the same study base and/or not being representative of the populations they come from


* Clear definition of the study population
* Explicit case and control definitions
* Cases and controls taken from the same population

Front 64

Information bias

Back 64

occurs when the method of gathering information is inappropriate and yields systematic errors in measurement of exposures or outcomes

* Blinding prevents investigators/interviewers knowing status of subjects
* Questionnaires use multiple questions that ask the same information
* Accuracy through double-checking records, gathering data from multiple sources
* Use multiple controls

Front 65

Confounding bias

Back 65

a third factor which is related to both exposure and outcome, and which accounts for some/all of the observed relationship between the two.

* Randomization tries to evenly distribute potential (unknown) confounders in study groups

* Matching ensures equal representation of subjects with known confounders in study groups

Front 66

What is considered a herd level test?

Back 66

An evaluation of a sample of (or all) animals from a herd and the application of decision rules that classify the herd as positive or negative based on the test results from individual animals

Front 67

3 types of random sampling strategies

Back 67

1. Simple
2. Systematic
3. Stratified

Front 68

Simple random sampling strategy

Back 68

• No omissions of animals
• No duplicates
• Each animal uniquely identified
• A sampling frame is necessary: a list of all animals to be samples must be available
• Can use a random number table or computer random number generator

Front 69

Systematic random sampling strategy

Back 69

Only the first animal in an array of animals is selected truly at random, afterwards each nth animal is sampled

Front 70

Stratified random sampling strategy

Back 70

• Different production groups should be sampled proportionately to their size if they are likely to affect the outcome

Front 71

Haphazard (judgment) sampling

Back 71

No apparent plan, but done in the belief that it mimics probability

Front 72

Convenience sampling

Back 72

The animals that are the most expedient/opportune are

Front 73

Targeted sampling

Back 73

Selection of animals meeting some predetermined characteristic

Front 74

Minimum inhibitory concentration (MIC)

Back 74

 Minimum concentration of drug required to inhibit growth of a bacterial isolate under standardized conditions

Front 75

Susceptible

Back 75

Implies that an infection due to the isolate may be appropriately treated with the dosage regimen of an antimicrobial drug – it does NOT mean the drug will be effective in all cases!

Front 76

Phenotypic resistance

Back 76

A strain that grows at an MIC at or higher than the determined breakpoint

Front 77

Genotypic resistance

Back 77

A strain that is classified resistant due to results of PCR, gene sequencing, etc

Front 78

What type of categorical data has no inherent relationship between the categories?

Back 78

Nominal

Front 79

Power is defined as

Back 79

the ability to detect a difference between groups when there really is a difference

Front 80

What type of categorical data has an ordering or ranking got the catrgories?

Back 80

ordinal