Epi Flashcards
Reproductive number
R0=Bcd
probability of transmission
Avg number of contacts per unit time
Avg duration of infectiousness
Rt=bcd*proportion susceptible
Infectious disease epidemiologic triad
Agent, environment, host
Characteristics:
virulence, pathogenicity, infectivity
Weather, geography, crowding, food
Age, sex, nutrition
Epidemic curve
Distribution of disease by time of onset
incubation period
Time of infection to time of onset of symptoms
Latency period
Time of infection to time of infectiousness
Herd immunity
Resistance of a group to infection
Type I error
Finding an association when there is none ALPHA
Type II error
Finding no association when there is BETA
Power
1 - Beta, Correctly finding an association when there is an association, usually 80%
Error of concern when confidence level is close to null value
Random error or lack of precision
What does wide or narrow CI tell you?
Sample size
95% Confidence Interval
When a study has been repeated 100 times, 95% of the time, the CI would contain the real population measure of interest
Epidemiology
The study of the distribution and determinants of disease frequency in the human population and the application of the study to control health problems
Types of variables
Categorical/nominal, ordinal, continuous
Decreasing prevalence means…
More people are cured
More people died
Morbidity vs Mortality monitoring
- early id of problem
- but more events to track, spectrum of disease, difficult to ID and count, Less complete
- Fewer events to track, ID impt health problems, very complete
- but Far downstream in the disease process, tip of iceberg, cause of disease misclassification
Case Fatality Rate
Number of people who die from disease within the people who have the disease, expression as per 100,000 population
Annual Mortality Rate
number of deaths in the year over population at mid year, expressed as percentage
Case Reports and Case series: impt and limitations
Importance: Early detection of unusual occurrence, preliminary source of information for potential
Limitations: No comparison group, Small sample size, no exposure collected
Observational vs Experimental Studies
- Exposure is assigned in experimental, assignment is randomized, maximize control, expensive and maybe unethical
Descriptive vs analytic studies
Descriptive: burden and patterns, no direct inference to causality, first impt clues about causes
Analytic: hypothesis stated and tested, associations betwee E and D to understand cause
Name some descriptive studies
cross sectional, case series, case report
Cross-sectional studies
snapshot, planning, but no temporal order, NHANES
Purpose of surveillance
est magnitude of problem, geographic distribution, detect epidemics, evaluate control measures, monitor infectious agents, detect changes in health practice, facilitate planning
Pearson’s correlation
parametric (outcome is continuous and normally distributed)
Spearman r
non-parametric,discrete, ordered
Instrumental Variable, mendelian randomization studies
predictors of exposure like genes
Attack rate
sick / # exposure to agent
3 Biases
Selection Bias, Information bias, confounder
How to adjust for measurement error?
blinding, triangulate, validated measures
How to adjust for confounder?
randomization, instrumental variable, matching, restriction at design stage
weighted average, propensity score, multivariate analysis at analysis stage
Selection bias
selecting participants not independent of exposure OR loss to f/u at different rates
Confounding definition
Unequal distribution of another cause of disease between exposed and unexposed
Koch’s POstulates
microorganism must be found in all suffering
must be isolated and grown in pure culture
should cause disease when introduce to healthy organism
reisolated from the inoculated, and ID as being identical to the original agent
Hill and Doll
Cigarette smoking
Hill’s causal viewpoints
- strength of assc
- consistency
- specificity
- time sequence
- biological gradient
- plausibility
- coherence
- experimental
- analogy
dose response, biological gradient
dose response, threshold effect, ceiling effect
Association and causation
assoc - being able to predict one based on the other
causation - without one, the disease may not happen or will happen at a later time