Epidemiology and Biostats Pt. 5 Flashcards
Parallel vs. Serial testing
Parallel: 2 or more different tests run on a patient or herd at the same time, and only ONE of the test results must be positive for a positive diagnosis, or only animals that have negative results on all tests are considered free of disease
Serial: 3 or more different tests run on a patient or herd, but all test results must be positive for a positive diagnosis to be made
Pros/cons of parallel testing
Pros:
-useful when you have relatively insensitive tests and want to boost the net effect of a more sensitive diagnostic strategy by running multiple tests at once
-best used on individual animals
Cons:
-number of tests increases
-risk false positive increases
-not best for groups of animals
Pros/cons of serial testing
Pros:
-maximizes specificity and PPV
-more confident in positive test results
Cons:
-lowers sensitivity and NPV
-increases risk that disease will be missed
Likelihood ratio
the likelihood that a given finding on the hx, PE, or lab exam would occur in an animal with the condition of interest
likelihood ratio for a positive test = sensitivity/(1-specificity)
likelihood ratio for a negative test = (1-sensitivity)/specificity
Results:
<1: result is associated with the absence of disease
>1: result is associated with disease
closer to 0: negative test
closer to 1: test conveys no information
closer to infinity: positive test
ROC (Receiver Operating Characteristic) Curve
compares the true positive rate (sensitivity) on the vertical axis to the false positive rate (1-specificity) on the horizontal axis
-can be used to select cutoffs or to compare diagnostic tests.
-tests that discriminate well approach the upper left corner of the ROC curve
-the area under the curve measures overall test performance
3 basic types of questions
1) close-ended questions (easier to analyze)
2) open-ended questions (harder to analyze)
3) precoded, open-ended questions (participants can answer unprompted but the interviewer selects from precoded response categories)
active vs. passive surveillance
active: project staff are recruited to carry out a surveillance program
passive: available data on reportable diseases is used; used for routine notifiable diseases. Limited by variability and incompleteness in reporting
Surveillance systems at the state and local levels
-communicable disease surveillance reported by health care providers
-dz registers surveillance (ie. cancer)
-sentinel surveillance (ie. influenza)
-periodic population surveys
-syndromic surveillance
-vital registration (ie. infant mortality)
-administrative data on health
steps in planning a surveillance system
- establish objectives
- develop case definitions
- determine data source or data collection mechanism
- develop data collection instruments
- field-test methods
- develop and test analytic approach
- develop dissemination mechanism, so that decision makers can easily understand the data and implications
- ensure use of analysis and interpretation
The CDC describes 7 steps in a foodborne outbreak investigation
1) detect a possible outbreak (surveillance, PulseNet, clinicians)
2) develop case definition and find cases
3) generate hypotheses about likely sources
4) test the hypothesis (analytic epi studies, food testing)
5) find the point of contamination (traceback, interview food prep workers, etc.)
6) control the outbreak (ie. clean/disinfect food facilities, temporarily close, recall, education, etc.)
7) decide when the outbreak is over (when number of new cases drops back to normal)
on an epi curve, a sudden rise in the number of cases suggests:
a sudden exposure to a common source once incubation period earlier
in a point-source epidemic, all cases occur within on incubation period
T
How to identify the period of exposure from an epidemic curve of an apparent point source epidemic
1) look up the average and minimum incubation periods of the disease
2) identify the peak of the outbreak or the median case and count back on the x-axis one average incubation period
3) start at the earliest case of the epidemic and count back the minimum incubation period.
Ideally, these two dates will be similar and represent the probable period of exposure
possible sequela to STEC
hemolytic uremic syndrome (HUS)
molecular/genetic epidemiology
The study of distribution and determinants of disease using molecular biology
techniques. These techniques can be used in surveillance, outbreak
investigation, investigating modes of transmission, and identifying risk factors.
The clustering at a particular value is known as the central location or central tendency of a frequency distribution.
T
Two measures of spread commonly used in epidemiology
-range
-standard deviation
a normal distribution is characterized by:
1) it is symmetrical about the mean (μ), 2) the mean, median and mode are all equal, 3) the area under the curve is symmetrical, right to left, 4) the area under the curve for a distance of +/- one standard deviation from the mean is about 68% of the total area; +/- 2 standard deviations about 95% of the total area
binomial distribution
used as a model for outcomes that can be only one of two possibilities: yes/no, success/failure, sick/well, etc. The probability of “success” (p) remains constant from trial to trial. The trials are independent, where the outcome of any particular trial is not affected by the outcome of any other trial.
The 2 distribution parameters are the number of independent trials (n) and the probability of success (p). The Binomial Table is used to find the cumulative probability that a value (x) is less than or equal to some specified value
Poisson distribution
contains a single parameter lamda (λ) which is the average number of occurrences of a random event in an interval of time or region of space.
The probability of observing k events in an interval of a defined length given a known rate of occurrence (λ) can be determined using the formula p(k) = where e = the natural number (2.7182818…).
the center of gravity of a frequency distribution is the
mean
geometric mean
the mean or average of a set of data measured on a logarithmic scale.
-used when the logarithms of the observations are distributed normally (symmetrically) rather than the observations themselves
-is always smaller than the arithmetic mean
-less sensitive to extreme values
how to find the median
-middle position of all the values
middle position = (n+1)/2
if falls between 2 observations, is the average of the 2 values
One standard deviation includes 68% of the values in a sample population and two standard deviations include 95% of the values.
T
