Midterm 1 Flashcards
What is epidemiology
The science of understanding the distribution and causes of population health so that we may intervene to prevent disease and promote health.
Quantitative
unique vocabulary
interdisciplinary
Descriptive vs analytical studies
Descriptive focuses on distribution
Analytical focuses on determinants and relationships between them
Two core functions of epidemiology
- identifying causes of health
- So that we may intervene (intervention)
What is a cause
cause is something that makes a differences or produces change
What are some challenges for epidemiology
Chronic diseases
Current conceptual movements
Most epidemiological studies are ___________
observational
What is confounding
Confounding is a distortion of an exposure-outcome association brought about by the association of another factor(s) with both outcome and exposure
The effect of the exposure is mixed together with the effect of another variable leading to bias
What is selection bias
Distortions that result from procedures used to select subjects and from factors that influence participation in the study
What is information/Measurement bias? Defining feature?
Distortion in the measure of effect caused by a lack of accurate measurements of exposure or disease status
Defining feature: occurs at data collection stage, misclassification of exposure is the main source of error
objectives and major dimensions of descriptive epidemiology
Objectives: Permit evaluation of trends in health and disease and comparisons among countries and sub groups, evaluation of health services, hypothesis generation
Major Dimensions: Assumption that diseases do not occur at random
Three standard questions are typically posed to characterize disease distribution
What are the three standard questions for descriptive epidemiology
Who gets the disease (person)?
Where does the disease occur (place)?
When does the disease occur (time)?
what is a population
A population is a collection of individuals, at moments in time, defined by at least one organizing characteristic
What are the measures of disease frequency that should be taken into account?
Number of individuals affected
Size of population
Time/Time period
What is prevalence
Measures existing cases of a disease at a particular point in time or over a period of time
The porbability that a member of the population has the disease
Point prevalence
Proportion of people who possess a certain attribute at a certain point in time
of existing cases at a given point in time/ total population at a given point in time
what factors can increase prevalence
Longer duration
Increased incidence
In-migration
Out-migration of healthy
better diagnosing
What is incidence? 3 key concepts?
- Quantifies number of new cases of disease that develop in a population at risk during a specified time period. Can be measured as a rate or a proportion
3 key concepts:
New disease events, or first occurrence
Population at risk can’t have disease already, should have relevant organs
Time must pass for a person to move from health to disease
Cumulative incidence
Number of new cases of a disease in a given time/ Total population at risk
Cumulative incidence is the proportion of an initially disease free group of individuals who develop the disease within a specified period of observation
Other words for cumulative incidence
Cumulative incidence = incidence proportion = Risk
Artifactual influences on changes in rates over time
Changes in the ability to recognize the disease
Changes in the efforts to recognize disease
Changes in the definition of the disease
Limitations of cumulative incidence/incidence proportion (2)
- Cumulative incidence calculation assumes that you have followed the entire population for the entire follow-up period
- cumulative incidence doesn’t explicitly account for the passage of time
Incidence proportions can only be directly calculated in _______ populations
closed/stationary
Stationary vs dynamic populations
Dynamic: allows for movement in and out of the population
Stationary: does not allow for movement in or out of a population. Population remains same never losing nor adding others
Censored observations
Measurements on those subjects who do not complete the entire study period for reasons other than developing the study outcome
What is incidence density (incidence rate)
Describe how rapidly health events are occurring in a population of interest
True rate because it directly integrates time into the denominator
Does not make assumption to complete follow-up
Incidence density = _______ _______
Incidence rate
ID equation
Total person time= the sum of every persons time at risk
We add up the period of time each person was at risk
what is person-time?
when is prevalence or incidence more important
In general, Incidence is more important when thinking of etiology of the disorder, prevalence when thinking of societal burden of the disorder including the costs and resources consumed as a result of the disorder
When is CI or ID more useful
CI is most useful if interest centers on the probability than an individual will become ill over a specified period of time.
ID is preferred if interest centers on how fast the new cases are occurring in the population
Summary (not real card)
CI and ID numerator and denominator
Morbidity measures vs mortality measures vs natality measures
Morbidity measures pertain to the sickness, disease or disability within specific populations.
Mortality measures describe the frequency of death in populations.
Natality measures measure the frequency and the probability of births within specific populations
Crude vs stratum-specific estimate
◼ Crude estimate – a measure of disease occurrence for an entire population
◼ Stratum-specific estimate – a measure of disease occurrence for a population subgroup (e.g, age, sex)
Crude annual death rate = _______
total number of deaths during calendar year
case-fatality
Number of deaths due to the disease in a specified period of time/Number of cases of the disease in the same period of time
proportionate mortality
of deaths due to a specific cause/ total # of deaths
What is inference
Process of gaining information about a population based on data collected from a sample
Target population
Target population is the subject of inference: population whose parameters are estimated through sampling
sample vs source vs target population
what are different estimates due to
sampling variability
What is the difference between a true value and a sample-based estimate
random error
how do you reduce random error
Larger sample size
During data analysis
What are the two ways random error can be addressed in data analysis
Confidence intervals
Statistical tests
What are confidence intervals
◼ Measures of disease frequency such as prevalence are point estimates of the population parameter.
◼ How well the point estimate estimates the parameter (i.e., its precision) depends on sampling (random) error.
◼ One way of estimating the precision of a point estimate is to calculate a confidence interval
95% confidence interval formula for CI or point prevalence
95%CI= p+/- 1.96√p(1-p)/(n+4)
where p = (x + 2)/(n + 4)
p refers to the adjusted sample proportion (either cumulative incidence or prevalence stated as a decimal fraction, for example, .10)
n is the appropriate denominator
x is the appropriate numerator
95% CI for incidence density
What is measurement
Measurement is the assignment of numbers to aspects of objects or events according to one or another rule or convention.
what do we measure with
Measurement instrument: a procedure or set of procedures designed to measure one or several variables of interest
4 types of data
Nominal
Ordinal
Interval
Ratio
Nominal and ordinal data
Qualitative/ categorical
Nominal= classification only
Ordinal= Classification + logical order
Interval and ratio data
Quantitative/ continuous
Interval = Classification + Logical Order + Equal Intervals
Ratio= Classification + Logical Order + Equal Intervals + Absolute Zero
what is the goal of our measurement
provide valid estimates of true disease prevalence/incidence
What is validity
Does the instrument measure what it is intended to
What is misclassification error
Measurement tool classify individuals into the wrong categories
two statistics which are indicators of validity
sensitivity and specificity
What are validity studies
◼ In these studies, a group of individuals are administered two tests: 1) first, a gold standard test (ie., criterion standard) in which there are no misclassification errors;
2) second, another (experimental) test purportedly measuring the same thing but may be cheaper, shorter,
less invasive….
◼ This allows for each respondent to be classified into one of four groups
Sensitivity vs specificity
What are complements of sensitivity and specificity
◼ Complementary means probabilities add up to 1.
◼ The complement of sensitivity is the false-negative rate (ie., 1-sens=fn rate).
◼ The complement of specificity is the false-positive rate (ie., 1-sp=fp rate)
two main sources of bias
misclassification bias
selection bias
Random vs Measurement vs Selection error
sensitivity equation, false negative rate
tp/tp+fn = sensitivity
1- sensitivity= FN rate
specificity formula, false positive rate
Specificity = tn/tn+fp; 1 – Sp = false positive rate
What is reliability? 2 focuses of reliability
Does the instrument measure something in a reproducible fashion
- Test/method or observer/rater
- within or between test/methods or observer/rater
3 different types of reliability
How to calculate reliability
calculate percent agreement
What is the kappa statistic
Measure of reliability
The kappa statistic is the excess agreement over that expected by chance (EP), divided by the potential excess
Kappa = (OP-EP)/(1-EP)
OP= (a+d)/n, where
n = a+b+c+d
EP = [(a+b/n) X (a+c/n)] + [(c+d/n) X (b+d/n)]
1-0, 1 = complete agreement, 0= agreement is equal to that expected by chance, negative indicates agreement less than expected by chance
Selection bias
Type of systematic error related to participation
Can lead to missing data Participants only providing partial information, dropping out of study etc
Probability sampling
probability of selecting a person from the population into the sample is known
Selection probability
The probability that a member of the target population is selected into a sample
Positive and negative predictive value (PPV & NPV)
PPV: Probability that a person has the disease given that he or she tests positive
NPV: the probability that a person is disease free given that he or she tests negative
2 tests of validity
- Other measures of the same thing are available (already talked about this)
- No other (good) measure exists
Convenience samples
selection probabilities are unknown
Simple random sampling
each member of a population has the same chance of being selected into a sample; that is, the selection probability for each member of the population is the same
which probability sampling does statistics Canada employ
use complex methods sampling methods to ensure estimates are representative
Most important procedures to handle confounding are:
Stratification
Standardization
Direct vs indirect standardization
Direct: selecting a standard population and usign the age structure from the standard population and multiplying it by stratum-specific disease frequency rates form the two populations being compared (DSMR)
Indirect: Used when there is missing data and/or cell size may be too small. Selects a standard population and uses the stratum-specific disease frequency rats from the standard population and multiplying them by the age structure of the population with missing data
Is it necessary to standardize for age when comparing measures of disease frequency
no, two criteria must be met
Age must be associated with place
age has to be associated with the disease of interest
How to know if there was confounding
Calculate Rc and compare to DSM
If DSMR = Rc then no condounding was performed
Rc= Crude pop A/ Crude pop B
When DSMR =/ Rc standardized rates should be reported
Indirect standardization
the measure of interest is called the Standardized Mortality (or morbidity) Ratio (SMR)
Indirect steps
Step 1: calculated expected # of deaths if the age-specific death rates were the same as in the standard population
Step 2: Calculate SMR
SMR = Observed deaths/Expected deaths
Comparisons of crude, specific, and adjusted rates