Glossary

Term

Definition

A

Accuracy

The extent to which a given measurement is close to the true or accepted value.

Acquiescence bias

The tendency to agree with a statement when in doubt.

Active surveillance

Disease surveillance that involves proactive case-finding.

Additive interaction

Interactions that occur on an additive scale.

Aggregation bias (also known as “ecological fallacy”)

When incorrect conclusions about individual-level phenomena are made based on aggregated data instead of individual-level data.

Attack rate (AR)

The number of cases divided by the number of persons who have been exposed. Akin to “incidence”. Commonly used in outbreak investigations.

Attributable risk (AR) (also known as “risk difference (RD)”)

The incidence among the exposed that could have been prevented had the exposure been removed, assuming a causal association between the exposure and the outcome.

Attributable risk percentage (AR%)

The proportion of incidence (converted to percentage) among the exposed that could have been prevented had the exposure been removed, assuming a causal association between the exposure and the outcome.

B

Base population

The population that gave rise to the cases.

Baseline

The beginning period in a cohort study and experimental studies in which characteristics are measured for the first time or when intervention (or lack thereof) is allocated to the participants.

Bias

Systematic deviation of an observation from the truth.

Burden of disease

The impact of a health problem on a population.

C

Cases (in case-control studies)

Individuals with the outcome of interest.

Case-control study

A type of observational study in which individuals with the outcome of interest are compared to individuals without the outcome of interest with regard to the exposure of interest.

Case fatality rate (CFR)

The number of patients who die from a given health problem divided by the number of patients with the health problem.

Case series

The description of a group of individuals with the outcome of interest.

Cause

An agent that modifies a state of health, without which the modification would not have occurred.

Causal inference

The process of generating a conclusion about whether a given exposure is causally associated with a given outcome.

Causality

The relationship between cause and effect.

Censoring/censored

A condition in which the value of a measurement or observation is only partially known. Commonly mentioned in cohort studies and randomized trials.

Clinical trials

Experimental studies in biomedical or behavioral science where the investigators allocate a given intervention to the participants in a clinical setting.

Cohort

A group of individuals who share a common characteristic.

Cohort studies

A longitudinal study on a given cohort of individuals.

Collinearity

A strong correlation between two independent variables that makes it difficult to distinguish the influence of individual variables

Concurrent validity

The extent to which a criterion (outcome) obtained at the same time as the test (or the question) corresponds to the test or the question.

Content validity

The extent to which the test items (or questions) are a sample of a universe in which the investigator is interested.

Confounder

An extraneous variable whose effect is mixed with an exposure that influences the distortion in the extent to which the exposure is associated with an outcome (see also “Confounding”).

Confounding

The mixing of effect between an extraneous variable (a confounder) and an exposure that distorts the extent to which the exposure is associated with an outcome.

Construct validity

The extent to which test items (or questions) can be interpreted as a measure of an attribute or quality that is not operationally defined.

Control group

Individuals with the outcome of interest (in case-control studies) or without the exposure/intervention of interest (in cohort studies and experimental studies).

Controls (in case-control studies)

Individuals with the outcome of interest.

Comparison group

A group to which a comparison is made with regard to an outcome or an exposure of interest.

Competing risks

Events that prevent or augment the primary outcome of interest.

Component cause

An event or condition that contributes to the development of the disease but is not sufficient to cause the disease on its own.

Correlation

The extent to which two variables are associated.

Counterfactual model

A thinking model used to describe the occurrence of an observed outcome (or exposure) of interest had other circumstances been different.

Cross-sectional study

An observational study that analyzes data collected at a single point in time.

Cumulative incidence

The number of new cases divided by the number of the population at risk within a specific time interval.

D

Data visualization

Presentation of data using graphics to enable information to be easily communicated and understood.

Determinant

Any cause (biological, physical, chemical, or psycho-social-economic factors) that can increase or decrease a disease’s likelihood of occurrence.

Differential loss to follow-up

When loss to follow-up occurs at different extents between comparison groups (see also “Loss to follow-up”).

Differential misclassification

When misclassification occurs at different extents between comparison groups (see also “Misclassification”).

Disease eradication

Reduction of the disease’s prevalence to zero cases.

Disease-specific mortality

Mortality attributed to a particular disease as the primary cause of death.

E

Ecological fallacy (also known as “aggregation bias”)

When incorrect conclusions about individual-level phenomena are made based on aggregated data instead of individual-level data.

Ecological studies

Epidemiological studies that describe the association between an exposure and an outcome using aggregated data at the population level.

Effect modification

The extent to which the association between an exposure and an outcome varies according to the level of an extraneous variable (see also “Interaction”).

Effect modifier

An extraneous variable upon which the strength of association between an exposure and an outcome varies.

Effectiveness

The effect of an intervention on an outcome in real-world circumstances.

Efficacy

The effect of an intervention on an outcome in ideal settings.

Epidemic

An unusually high level of disease occurrence in a given population at a particular time.

Epidemiology

The study of the distribution of health outcomes (and their determinants) in a given population and the application of such knowledge to disease prevention and control.

Experimental study

Scientific studies in which the investigator allocates a given exposure to the participants.

Exposure (also, non-exposure)

Any variable or phenomenon that may affect health.

Exposure misclassification

Misclassification of individuals with an exposure as those without the exposure or vice versa.

F

Financial epidemiology

A study of the extent to which personal finance and associated psychological attributes are associated with health outcomes in a defined population.

Follow-up

Data collection among the participants of longitudinal observational or experimental studies to measure the outcome(s) of interest.

Fomite

An object that may carry an infection.

G

Genetic epidemiology

The field that studies the extent to which genetic attributes are associated with health outcomes in a defined population.

Genome-wide association study (GWAS)

An observational study on the association between a genome-wide set of genetic variants and the outcome(s) of interest.

Germ theory

A scientific theory that microorganisms can cause diseases.

Global health

A study of the health of the populations in a global context.

Gold standard

A widely accepted standard test against which new tests (including screening tests) are assessed with regard to validity.

I

Incidence

The number of new cases divided by the susceptible population (see also “risk”).

Incidence density (also known as “incidence rate”)

The number of new cases divided by cumulative person-time at risk.

Incidence rate (also known as “incidence density”)

The number of new cases divided by cumulative person-time at risk.

Incident cases

Cases recruited after being newly diagnosed.

Information bias

Bias that occurs based on the method and circumstances in which information was collected.

Intention-to-treat analysis

A method of analysis of experimental data that classifies exposure (intervention vs. control status) based on allocation, regardless of whether a participant actually received the intervention.

Interaction

The extent to which the association between an exposure and an outcome varies according to the level of an extraneous variable (see also “Effect modification”).

International health

Cross-border healthcare with an emphasis on the north-south approach.

Intervention group

A group of participants in an experimental study that is allocated the exposure by the investigator.

K

Koch’s Postulates

A set of observations and requirements designed to establish a causal relationship between a microbe and a disease.

L

Loss to follow-up

Participants enrolled in a cohort or experimental study who cannot be contacted for a follow-up visit and thus become lost.

M

Matching

Process of finding a comparison group that shares a similar attribute.

Misclassification

When a participant from one comparison group is mistakenly classified as belonging to the other group.

Monitoring

Routine observation of general health status in a population.

Mortality

Death per a given number of a population.

Multiplicative interaction

Interactions that occur on a multiplicative scale.

N

Necessary cause

An event or condition that contributes to the development of the disease, without which the disease cannot occur.

Negative predictive value (NPV)

Proportion of the population with a negative screening test result that actually does not have the disease.

Nested case-control study

A case-control study conducted among cohort participants.

Non-differential misclassification

When misclassification of an exposure status or an outcome status occurs evenly between comparison groups (see also “Misclassification”).

O

Observational study

Studies in which the investigator does not allocate an exposure to the participants.

Observer bias

Information bias that occurs due to the influence of an individual conducting the measurement or observation.

Odds

The number of persons with an exposure divided by the number of persons without the exposure, or the number of persons with an outcome divided by the number of persons without the outcome.

Odds ratio

The ratio of the odds of the outcome between the exposed and non-exposed groups, or the ratio of the odds of the exposure between the cases and the controls.

Outcome

The disease, behavior, or any other occurrence whose distribution the investigator wishes to describe in a given population.

P

Pandemic

A widespread occurrence of a disease at a global level.

Passive surveillance

Disease surveillance that does not involve active case-finding.

Pathway variable

A variable that occurs as an intermediate phenomenon in the potential causal pathway between an exposure and an outcome.

Percentage agreement

The extent to which two measurements agree with one another.

Period prevalence

The proportion of a population with a characteristic of interest at any point during a given period of time.

Per-protocol analysis

A method of analysis of experimental data that only includes participants who completely receive the allocated exposure.

Person-time at risk

The cumulative amount of time at risk of disease among all susceptible persons.

Population attributable risk (PAR)

Incidence that could have been prevented in the entire population of interest had the exposure been removed, assuming a causal association between the exposure and the outcome.

Population attributable risk percentage (PAR%)

The proportion of incidence (converted to percentage) in the entire population of interest that could have been prevented had the exposure been removed, assuming a causal association between the exposure and the outcome.

Positive predictive value (PPV)

The proportion of the population with a positive screening test result who actually does have the disease.

Predictive validity

The extent to which a criterion (an outcome) obtained after a test is given (or a question is asked) corresponds to the test or the question.

Prevalence

The proportion of individuals with an outcome in a given population.

Prevalent case

Cases recruited regardless of when the diagnosis happened.

Prospective cohort studies

Cohort studies conducted by investigators who recruit a new group of participants and make baseline measurements and plan for prospective follow-up visits.

Point prevalence

The proportion of a population with a characteristic of interest at a given point in time.

Public health

Actions taken to protect and improve health at a population level.

Q

Quarantine

Isolation of members of a population who have been exposed to an agent of disease.

R

Randomization

The allocation of a status in an unpredictable manner.

Randomized clinical trial

An experimental study in an ideal clinical setting in which the investigator randomly allocates an exposure of interest (“intervention”) to a group of participants and allocates the non-exposure (“control”) status to another group of participants.

Randomized trial

An experimental study in which the investigator randomly allocates an exposure of interest (“intervention”) to a group of participants and allocates the non-exposure (“control”) status to another group of participants.

Recall bias

Information bias in case-control studies that occurs from cases recalling past exposure differently compared to the controls.

Relapse

When an outcome re-occurs after a period of improvement or remission.

Reliability

The precision of a measurement.

Restriction

A method to control for confounding in which a characteristic that can be a potential confounder is restricted or excluded from a study.

Retrospective cohort studies

Cohort studies conducted by investigators who measure the outcome (i.e., conduct follow-up measurements) among members of a defined cohort.

Risk

Number of new cases divided by susceptible population (see also “Incidence”).

Risk difference (RD)

The incidence among the exposed that could have been prevented had the exposure been removed, assuming a causal association between the exposure and the outcome (see also “Attributable risk (AR)”).

Risk ratio (RR)

The ratio of the risk of the outcome among the exposed, divided by the risk of the outcome among the non-exposed.

S

Screening

Early detection of diseases.

Selection bias

Bias that occurs from the selection of participants and the occurrence of loss to follow-up in a study.

Self-serving bias

Information bias that occurs from the need of participants to maintain self-esteem while providing answers.

Sensitivity

The proportion of the population with a positive screening test result divided by all screened individuals with the outcome of interest.

Sensitivity analysis

An analysis of the extent to which the findings of a study are affected by uncertainty in its inputs.

Social desirability bias

Information bias that occurs from the need of participants to avoid telling the truth about legally or socially sensitive topics.

Specificity

The proportion of the population with a negative screening test result divided by all screened individuals without the outcome of interest.

Statistical adjustment

A method to control for confounding post hoc (i.e., after data collection) by using stratified and multivariable analyses.

Stratification

Division of a study population into sub-groups.

Stratified analyses

Assessment of potential confounding or effect modification by analyzing the extent to which an exposure is associated with an outcome after stratifying the participants by the extraneous variable that could be either a confounder or an effect modifier.

Study arm

Comparison groups in an experimental study.

Study design

The process of designing a study’s methods for participant selection and data collection.

Sufficient cause

The full set of events or conditions that will invariably contribute to the development of a disease of interest.

Surveillance

Specific observation of a particular health outcome in a population.

Susceptible-infectious-recovered (SIR) model

An epidemiological model used in the prediction and forecasting of infectious disease outbreaks and epidemics.

T

Temporality

The notion that a given exposure must happen before a given outcome in order to make an inference regarding the causal relationship between the exposure and the outcome.

Triangulation

The practice of obtaining more information about a given measurement from additional sources to confirm its accuracy and validity.

V

Valid

When a variable measures what it intends to measure.

Validity

The extent to which a variable measures what it intends to measure.

W

Window of exposure

A time period in which an exposure of interest influences an outcome of interest.