Unit 1 - Definitions and Equations Flashcards
Attack Rate
Incidence Proportion
Proportion of the population that develops illness during an outbreak
= # of new cases / # of people at risk of illness (or in pop)
Case Definitions
Set of criteria that a case must meet in order to be definitively diagnosed with a disease
Need to be determined before we can define the “who” of Descriptive EPI
Based off of this, we split cases into confirmed vs probable
Confirmed vs Probable Cases
C: person who meets all of the specified criteria and can be diagnosed
P: person who meets the majority but not all of the criteria in order to be definitively diagnosed. Even if health care provider is super sure that they qualify for they diagnosis, they cannot be confirmed until meet all criteria
Case Fatality Rate
Proportion of persons with a disease who die from it
Cannot die from a disease if you don’t have it
= # of cause specific deaths / # cases of disease
Cause- Specific Morbidity Rate
The morbidity rate from a specified cause for a population
= # of persons with cause- specific disease / # of people in population
Cause- Specific Mortality Rate
Mortality rate from a specified cause for a population
= # of cause- specific deaths / # of people in population
Cause- Specific Survival Rate
of cause- specific cases alive / # of cases of disease
Cluster
Outbreak of disease
Crude Morbidity Rate
= # of persons with any and all disease / # of persons in population
Crude Mortality Rate
= # of deaths / # of persons in population
Cumulative Incidence
Incidence calculated from the incidences of non-dynamic populations at different points in time.
Combining all of the incidences from a given time period from a non-dynamic population
Disease Registry
Running list gathered by the CDC based on input from health care providers about what disease they are seeing in private practice/ the public.
Part of passive surveillance requires health care providers to report diseases to the CDC so they can update / track registries/ diseases
Endemic
When a population has a higher than normal presence of a disease as their baseline compared to other populations
Epidemic
When a population experiences a higher than normal presence of disease compared to baseline presence
Occurs on a larger scale than outbreaks
Community and period is clearly defined
Fertility Rate
= # of live births / 1,000 women of childbearing age (15-44)
Fixed vs Dynamic Populations
F: non-dynamic
- little movement, stable
- used to calculate cumulative incidence
D: movement, less stability
- where we need standardization of some sort bec we can be less certain about size of population or lengths of time
- at risk group estimated by: population at start, middle, or end of year or average population over entire year
- used to calculate incidence density
National Notifiable Disease Surveillance System (NNDSS)
Gather information about diseases from the community
Requires that physicians and other health care providers report certain diseases to them and the CDC in order to keep tracking, update information
Allows them to notify others/ make resources available should an outbreak/ epidemic occur
Incidence
Also known as risk or attack rate
The occurrence of new cases of disease or injury in a population over a specified period of time (usually 1 year)
Can mean number of new cases in community or number of new cases per unit population
Used for people who develop a condition during a period of time
Measure of how fast (risk)
= # of new cases of disease / # people at risk
Incidence Density
Combination of the incidences over time from a dynamic population
Incidence Rate
Person - time rate
Measure of incidence that incorporates time directly into denominator
Describes how quickly a disease occurs in a population
= # of new cases / person-time each person was observed
Incubation Period
Time between exposure and onset of disease
Induction Period
Synonymous with incubation period
Time between exposure and onset of disease
Infant Mortality Rate
Most common measure for comparing health status among nations
Ratio not a proportion
= # of deaths among children < 1 year old / 1000 live births
Infectivity
Ability to invade a patient (host)
Incidence
= # infected / # at risk
Latency Period
Time between onset of disease and disease diagnosis
Live Birth Rate (Natality)
= # live births / 1,000 population
Maternal Mortality Rate
= # deaths of women while pregnant or within 42 days of termination of pregnancy / 100,000 live births
Morbidity
Any departure, subjective or objective, from a state of physiological or psychological wellbeing
- disease, injury, disability
Measures characterize the number of persons in a population who become ill (incidence) or are ill at a given time (prevalence)
Mortality
Death of an individual from any cause
Neonatal Mortality Rate
Neonatal period = birth - 28 days old
= # deaths of children under 28 days old / 1,000 births
Outbreak
Higher than normal frequency of disease in a localized population compared to baseline values
Occurs on small scale than epidemic
Interchangeable with cluster
Pandemic
When an epidemic spreads to cross geographic borders
Multi-national, Multi- continent
Determined by intensity of spread, multi-continent, and lethality
Pathogenicity
Ability to cause clinical disease
= # with clinical disease/ # infected
Period Prevalence
Prevalence measures over specific time period (usually mid year)
Person- Time
Generally calculated from a long-term cohort study
Assumes probability of disease during study period is constant
Allows us to standardize population when we don’t know the population size or the time period
Postnatal Mortality Rate
Postnatal period = 28 days - 1 year
= # deaths of children aged 28 days - 1 year / 1,000 live births
Prevalence
Proportion of people in a population who have a particular disease or attribute at a specified point in time or over a specified period of time
Includes all cases - new and preexisting
See for people who have a condition during a period of time
Used for measuring chronic disease
Measure of how much - burden of disease
Prevalence Rate
Proportion of the population that has a health condition at a point in time
Point Prevalence
Proportion of people with a particular disease or attribute on a particular date
Prevalence measured at a particular point in time
Proportion
Comparison of 2 related things
Represents a part of a whole
Have to factor numerator into denominator
Proportional Mortality Rate
Proportion of deaths in a specified population over a period of time attributable to different causes
= # of cause- specific deaths / total # deaths in population
Rate
Comparison of 2 things over time or a specified time period
Like a proportion but it must include time
Part over whole over time period
Ratio
Comparing 2 non-related things
Numerator and denominator are separate
Risk
Another term for incidence or attack rate
Secondary Attack Rate
Looks at the difference between community transmission of illness vs transmission of illness in a household/ other closed populations
= # of cases among contacts of primary cases / total number of contacts
Sentinel / Index Case
Case that epidemiologist’s retrospectively label as the first incidence of disease
- look at max incubation period and min incubation period
Where we link start/ spread of disease to
Cannot always be defined as 1 person - sometimes it is a range or nonexistent
Depends on illness and nature of contact with others
Active Surveillance
Going out into the community and asking questions and looking for new disease/ cases
More involved - need a lot of time, people, and resources
Need to ask questions, conduct research, observe people you are looking at
Ex: John Snow and Broad Street Pump
Passive Surveillance
Health care providers have to alert the CDC/ NNDSS when they see a certain disease
Passive bec they wait for updates from health care providers and then enter information into database - no boots on ground
Let’s us track disease frequency and occurrence over time and within populations
Syndromic Surveillance
Another form of passive surveillance
Used when person presents a certain set of symptoms but a definitive diagnosis cannot be given at moment.
Use already defined similar diagnosis to guide decisions and protocol until something more definitive can be done
Almost- could be phenomenon
Virulence
Ability to cause death (case fatality rate)
= # deaths / # with infectious disease
John Snow
Father of Epidemiology
Recognized a lot of people in a London neighborhood were sick. He conducted descriptive EPI by going into community and asking questions and keeping track of his information on a map.
He traced the bad water which was leading people to develop/ spread cholera to the Broad Street Pump
Epidemiology
A public health basic science that studies the distribution and determinants of health related states and events in specific populations to control disease and illness and promote health
Objectives of EPI
Identify patterns / trends
Determine extent
Study natural course
Identify causes of or risk factors for
Evaluate effectiveness of measures
Assist in developing public health policy
Looking to affect changes on population level
Epidemiological Assumptions
- Disease occurrence is not random
- Systematic investigation of different populations can identify associations and causal/ preventive factors and impact changes on health of population
- Making comparisons is the cornerstone of systematic disease assessments/ investigations
Distribution of Disease
Frequencies of disease occurrence
- counts of disease
- counts of disease in relation to size of population
Patterns of disease occurrences
- Person (Who)
- Place (Where)
- Time (Where)
= Descriptive EPIDescriptive EPI
Person, Place, Time
Who, Where, When
Can be used to know if a location is experiencing disease occurrence more frequently than usual for that locale or other locations
Determinants of Disease
Factors of susceptibility/ exposure/ risk
Etiology/ causes of disease
Modes of transmission
Social / Environmental / biological elements that determine the occurrence/ presence of disease
Looks at associations vs causes
= Analytic EPI
Analytic EPI
Why, How
6 Core Functions of EPI
Surveillance Field Investigation Analytic Studies Evaluation Linkages Public Policy
Public Health Surveillance
Portray ongoing patterns of disease occurrence so investigations, control, and prevention measures can be developed and applied
Skills: data interpretation
- designing and using data collection instruments
- data management
- scientific writing and presentation
Field Investigation
Determine source/ vehicle of disease; to learn more about the natural history, clinical spectrum, descriptive EPI (WWW), and risk factors of a disease
Analytic Studies
Advance the information generated by descriptive EPI techniques
Hallmark of analytic studies is use of comparison group
Skills: design, conduct, analysis, interpretation, and communication of research study data and findings
Evaluation
Systematically and objectively determine relevance, effectiveness, efficiency, and impact of activities
Linkages
Collaborate/ communicate with other public health and health care professionals and the public themselves
Policy Development
Provide input, testimony, and recommendations regarding disease control and prevention strategies, reportable disease regulations, and health care policy
Natural History of Disease Timeline
Stage of Susceptibility
- environmental, biological, or other
Stage of Subclinical disease
- body knows you have something but you are pre-diagnosis/ asymptomatic
- subclinical = asymptomatic
- things are advancing physiologically / pathological changes
- induction period = time between exposure and onset of disease
Stage of clinical disease
- patient and doctor know that disease is presence; may know diagnosis
- latency period = time between onset of symptoms and diagnosis
Stage of recovery, disability, or death
Emergency of International Concern
Epidemic that alerts the world to the need for high vigilance
Stage between epidemic and pandemic
Epidemic Curve
Graphical, time based depiction generated during an outbreak/ epidemic reflecting number of cases by date
Represents epidemic frequency change over time
Incorporates all 3 elements of descriptive EPI
What does epidemic curve depict?
Magnitude and Timing of Disease Occurrence
- how quick progression is - speed and intensity of disease
- sentinel case/ peak/ outliers
- start/ stop/ duration
Patterns of disease occurrence
- shape
- Common/ point source
- PropagatedCommon/ Point Source EPI curve
Not person to person spread
Can be continuous (not repeated) or intermittent (repeated)
Disease is derived from a common, single point source for outbreak
May or may not have index case
Propagated EPI curve
Person to person spread
Outbreak is spread as infected subjects infect others (secondarily) who then infect others
See start to disease and then saw toothing up and down
Probable Exposure Period
When have min / max, you can look back to determine range of onset
- need to know what illness you are working with
When look at max, min, and avg incubation time, you can determine the probable exposure period (range)
Factors in Comparing Measures of Disease Frequency between Groups
- Number of people affected/ impacted (frequency/ count)
- Size of the source population (from which disease cases or outcomes arose) or those at risk
- length of time the ‘population’ is followed
- frequency is more more likely to occur at time inc
Exposure
Exposure to disease, treatment, medicine
Outcome
Resulting in disease or another health related event
Counterfactual Theory
Describes the illogical best way to compare groups - impossible
In same group, all else being equal, looks at the outcome if the exposure didn’t occur
- pretend like the exposure didn’t occur
- ex: smoker’s risk of Coronary Heart Disease
- Would pretend that they got CHD but not from smoking so they would try to look at CHD separately from smoking
Exchangeability
Comparability with respect to all other determinants of outcome
Assuming groups are as similar as possible without that one factor/ exposure
In order to compare groups, need to assume this
The way that counterfactual theory becomes possible
Absolute Differences
Subtracting Frequencies (count)
= A - B
Ex: 65 surgeries in males and 27 surgeries in females
- males had 38 more surgeries or females had 38 fewer surgeries (65- 27)
Smaller than relative differences
Relative Differences
Division of Frequencies (ratio)
- = A / B
- ex: 65 surgeries in males and 27 surgeries in females
- males had 2.4 times the number of surgeries compared to females ( > 140% increase)
Division of Proportions
- ex: 65 / 70 surgeries in males (93%) and 27/ 59 surgeries in females (102% increase)
- females had just 49.5% of proportion of surgeries compared to males
- ratio of group proportions
Risk
Probability of outcome in an individual group based on exposure or non-exposure
Proportion
Helps us understand the impact of exposure on disease
Used to see the association between exposure and outcome, the changes exposure can have, the impact/ relationship of outcome and exposure between 2 groups
Also known as incidence risk
- proportion
Probability of Outcome in exposed group
= A / (A + B)
Outcome = numerator because it is the new case that is occurring in the population
Probability of outcome in non-exposed group
= C / (C + D)
If people are not exposed, why are they getting the disease/ outcome?
Absolute Risk Reduction
= B - A
Risk difference of the outcome attributable to exposure difference between groups
Also known as attributable risk
Can attribute risk to difference in exposure
Relative Risk Reduction
= ARR/ R unexposed
Need to ask what the baseline risk is - what would happen if I didn’t take the medicine?
Number Needed to Treat (NNT)/ Number Needed to Harm (NNH)
= 1 / ARR
- always round up to next whole number
Number of patients that need to be treated in order to receive the stated benefit or harm
- How many patients need to be treat vs harm in order for the effects to occur?
To prevent harm, want NNT to be small and NNG to be large
Risk Ratio
Also known as relative risk or incidence
Ratio of risks from 2 different (unrelated) groups
= risk of outcome in exposed / risk of outcome in unexposed
If ratio = 1
- outcome is equally likely for both groups
- numerator = denominator
If ratio > 1
- outcome is more likely to occur in comparison group (numerator)
- numerator > denominator
If ratio < 1
- outcome is less likely to occur in comparison group (numerator)
- numerator < denominator
Risk Ratio Interpretations
= 1.0
- no difference/ increase/ decrease in risks, odds, hazard
- n = d
> 1.0
- increased ratio
- 1.01 - +1.99 = use decimal value as percent
- this will never be considered ‘decreased’
- n > d
> 2.0
- increased/ greater risk/ odds/ hazard
- use phrase ‘x times greater’
- n > d
< 1.0
- decreased ratio
- 0.0001 to 0.99 = subtract from 1.0 and convert answer to percentage
- this will never be considered ‘increased’
- n < d
What are the 3 things to look for when interpreting ratios?
- Group comparison orientation
- exposed vs non-exposed
- Direction of words
- increased vs decreased
- above 1 vs below 1
- Magnitude
- 80% = 1.8 times vs 20% = 0.80 times
Forest Plots
Visual representation of ratios
Puts multiple ratios together
Components
- center line and 1.0 = equalness - 1:1 ratio
- to the right or left of center line = either greater or lesser than 1
Odds
Frequency of an outcome occurring vs not occurring
- Likelihood of event occurring / likelihood of event not occurring
Ratio
- comparing 2 different things
Frequency of exposure- in cases
A / C
Frequency of non-exposure (controls)
B / D
Odds Ratio
Ratio of the odds from 2 different groups
= Odds of exposure in disease / Odds of exposure in non-diseased
= AD / BC
= (A / C) / (B / D)
Confounding
Lack of Exchangeability (comparability)
A 3rd variable that distorts a measure of association between exposure and outcome
Alternative explanation of the association
What are the 3 requirements of confounders?
Independently associated with exposure
Independently associated with outcome
Not directly in the causal- pathway linking exposure and outcome
Shown in DAG - direct acyclic graph
Crude measure of association
Looking at exposure and outcome relationship while ignoring other factors
- not looking at confounds
Unadjusted association
Calculation of odds ratio/ risk ratio
What is an example of a classic confounder?
Age
Adjusted Association
Outcome measure of association between exposure and outcome for each individual level of the 3rd variable
Weighted average of all levels
Authors must tell what variables are used in the adjusting
Steps for Confounding testing
Calculate crude association
Calculate adjusted association
Compare crude vs adjusted measures
= absolute difference of crude and adjusted measures / crude measure
- if crude and adjusted estimates are different by 15%: confounding variable is present
Impacts of Confounders
Magnitude of association
- strength of association - association more extreme or less extreme than crude association - becomes not as impactful because exaggerates the real measure of association
Direction of Association
- produces association in opposite direction
Why do we want to control for confounders?
To get a more precise/ accurate estimate of the measure of association between exposure and outcome
How to control confounding?
During Study Design stage:
- Randomization, restriction, matching
Analysis of Data stage:
- stratification, multivariate statistical analysis
Randomization
Allocates an equal number of subjects with the known and unassessed confounders into each intervention group
Strengths:
- will be successful with large enough sample size
- stratified randomization more precisely assures equalness
Weaknesses:
- sample size may not be large enough to control for all unknown/ unassessed confounders
- process doesn’t guarantee successful, equal allocation between all intervention groups for all known and unknown confounders
- only practical for intervention studies
Restriction
Study participation is restricted to only subjects who do not fall within pre-specified categories of confounder
Strengths:
- straight forward, convenient, and inexpensive
- doesn’t negatively impact internal validity
Weaknesses:
- sufficiently narrow restriction criteria may negatively impact ability to enroll subjects - reduced sample size
- if restriction criteria is not narrow enough, it will allow introduction of residual confounding effects
- eliminates researchers ability to evaluate varying levels of the factor being excluded
- can negatively impact external validity
Matching
Study subjects selected in matched pairs related to the confounding variable, to equally distribute confounder among each study group
Strengths:
- intuitive, some feel it gives greater analytic efficiency
Weaknesses:
- difficult to accomplish, can be time consuming, and potentially expensive
- doesn’t control for any confounders other than those matched on
Stratification
Descriptive/ statistical analysis of data evaluating association between exposure and outcome within various strata within the confounding variable
Strengths:
- intuitive, straight forward, and enhances understanding of data
Weaknesses:
- impractical for simultaneous control of multiple confounders, esp those with multiple strata within each variable being controlled
Multivariate analysis
Statistical analysis of data by mathematically factoring out the effects of the confounding variable
Strengths:
- can simultaneously control for multiple confounding variables
- OR’s can be obtained and interpreted
Weakness:
- process requires individuals to clearly understand and interpret the data
- can be time consuming
Effect Modification
Also known as interaction
A 3rd variable that modifies the magnitude of effect of a true association by varying it within different strata of a 3rd variable
- modifies the effect across the strata
If present, we must report the measures of association for each strata individually
- do not want to control for/ adjust these
Is present when the odds ratio changes substantially according to different strata of the effect modifying variable
Steps in testing for Effect Modification
Calculate crude measure of association between exposure and outcome
- odds ratio, risk ratio
Calculate strata specific measures of association between exposure and outcome for each strata of the 3rd variable
- odds ratio, risk ratio
Compare each of the strata specific measures of associations between each other
- measure of association between the lowest and highest strata of the effect-modifying variable will be 15% different if effect modification is present
Cause
A precursor event, condition, or characteristic required for the occurrence of the disease or outcome
Help us understand things better
Not just one cause - usually multiple- component factors
Associations - definition and types
Relationships between an exposure/ treatment and outcome/ disease
- Artifactual associations
- Non-causal associations
- Causal associations
Artifactual Association
Arise from bias and or confounding
Something generates false measure of association
Makes us think there is a relationship but there isn’t
Non-causal associations
Occurs in 2 different ways
1. Disease may cause the exposure rather than exposure causing disease
- Disease and exposure are both associated with a third factor (confounding)
- can create mild distortion
- there is still a relationship between disease and exposure
Causal associations
Exposure directly linked to outcome
Sufficient Cause
Type of causal relationship
Precursor event that is sufficient enough on its own to induce disease and does so every time.
A set of minimal conditions/ events that inevitably produce disease
- can still have multiple, required components
Rare - except for genetic abnormalities
Component Cause
= Risk Factors
Type of causal relationship
A factor/ element that if present/ active increases the probability of a particular disease
- Multiple, required components that collectively act to induce disease
Some patients must be primed or susceptible to disease before component causes induce disease
- usually involves age
- more risk factors + more time = more likely to get disease
Necessary Cause
Type of Causal Relationship
A cause precedes a disease and has the following relationship with it:
- cause must be present for disease to occur but cause may be present without disease occurring
Cause is not sufficient to induce disease by itself but is needed to make the diagnosis
Synergism
Biological interaction of 2+ component- causes such that the combined measure of effect is greater than the sum of individual effects
Factors work together; both
Parallelism
Biological - interaction of 2+ component- causes such that the measure of effect is greater if either is present
Needs one or the other but not both
Factors work in parallel; either
Occurs if both risk factors are present but measure of association doesn’t change
Multiple Causation
Multiple risk factors working together to collectively become sufficient- causes
Component- causes have correlations
- collectively and with enough time present in patient
Hill’s Guidelines
Asks “what do we need in order to make the jump from association to cause?”
- Strength
- Consistency
- specificity
- Temporality
- Biologic Gradient
- Plausibility
- coherence, experiment, analogy
The higher number of criteria met, when evaluating an association, the more likely it may be causal
What process is Hill’s Criteria a part of?
Causal Inference Process
- an interpretive, application process
Strength
First category of Hill’s criteria
The size of the measure of association
The greater the association, the more convincing it is that the association might be causal
- the larger the number, the more important it is
Ignores the concept of changing exposure
Dependent on disease and exposure and their relationship
Consistency
Second category of Hill’s criteria
Also known as reproducibility
The repeated observations of an association in different populations under different circumstances in different studies
May still obscure the truth
- observational studies over long time can lead us astray
- Menopausal Hormone Therapy
- disproved via the Women’s Health Initiative Study
Temporality
Third category of Hill’s criteria
Reflects that the cause precedes the effect/ outcome in time
Time- order:
- Proximate causes = short term interval - distant causes = long term interval
Want to start looking at closest
- there could be delayed hypersensitivity - ex: drug side effect
Example of Temporality - Fact, inference, actuality
A number of studies have shown higher lung cancer rates among former smokers during the first year after cessation compared to those who continue to smoke
Inference: continuing to smoke must decrease the risk of lung cancer
Actuality: substantial portion of those who voluntarily stop smoking at any given time do so because of early symptoms and signs of the already-existing but as yet undiagnosed illness
Biologic Gradient
Fourth category of Hill’s criteria
Presence of a gradient of risk associated with degree of exposure
- dose response
Looking at intensity
- whatever relationship is with less exposure, as exposure increases, the risk is changed
- wrt negative component cause: more of it should be bad
Some demonstrate a threshold effect
- no effect until a certain level of exposure is reached
- ex: lead exposure and mental retardation
Plausibility
Fifth category of Hill’s criteria
Presence of a biological feasibility to the association, which can be understood and explained (biologically, physiologically, medically)
Want to answer the questions: “does it make sense? Can we explain it?”
