Epidemiology and Study Design Practice Flashcards

Comprehensive vocabulary flashcards covering basic epidemiology modules including study types, measures of frequency and association, bias, confounding, measurement error, RCTs, outbreak investigation, systematic reviews, and screening.

Descriptive Study

A study that measures a single variable and does not look at the association between two factors.

Analytic Study

A study that compares two variables and assesses their correlation to infer causation.

Randomised control trial (RCT)

The strongest primary study design involving random allocation of participants to exposure groups to balance confounders and directly test causality.

Cohort Study

An observational study with non-random allocation into groups based on exposure, where subjects are followed over time to calculate incidence and relative risk.

Case-control Study

A study where groups are selected based on the presence (case) or absence (control) of an outcome, looking back at exposure.

Study Base

The foundation of a case-control study defined by the place participants come from, the timeframe, and the person characteristics.

Cross-sectional Study

A study where exposure and outcome data are selected at the same time, measuring prevalence but unable to establish temporality or causality.

Ecological Study

A study that compares two variables at the population level rather than the individual level.

Ecological Fallacy

A limitation where associations found at the group level may not hold true at the individual level.

Crossover Trial

A study design where outcomes among an exposed group are compared to themselves when unexposed, meaning subjects serve as their own control.

Systematic Review (with meta-analysis)

The highest level of evidence that synthesises all high-quality studies on a question using reproducible methods and pools data to reduce random error.

Prevalence

The proportion of a population with an existing disease at a specific point or period in time; expressed as $\frac{\text{\# of people with attribute at time point}}{\text{total pop. at time point}}$.

Incidence

A true rate measuring how quickly people newly develop a disease during a specified period.

Person-time

The length of time people are at risk of a disease, measured in units like person-years, person-months, or person-days.

Incidence Density

Also known as incidence rate, calculated as $\frac{\text{\# new cases}}{\text{total person-time}}$, often expressed per $100,000$ person-years.

Incidence Proportion

The proportion of people at risk who develop disease during a specified period; also known as cumulative incidence or attack rate.

Symptom

Subjective indications of disease reported by the person.

Sign

Objective indications of disease apparent to the physician.

Rate Ratio / Relative Risk (RR)

An association measure indicating how much higher the rate of disease is in the exposed vs non-exposed; $RR > 1$ indicates a risk factor.

Risk Difference (RD) / Absolute Risk Reduction (ARR)

The difference in incidence between the exposed and unexposed groups ($\text{incidence in exposed} - \text{incidence in unexposed}$).

Odds Ratio (OR)

A measure of association used in case-control studies or cross-sectional analytic studies, especially when the outcome is rare.

Population Attributable Risk (PAR)

How much disease in the whole community can be attributed to a specific exposure.

Attributable Fraction (AF)

The proportion of outcomes in the exposed group that are specifically due to the exposure, calculated as $\frac{\text{Risk Difference}}{\text{Incidence in exposed}}$.

Number Needed to Treat (NTT)

The number of patients required to be given experimental therapy to prevent one adverse event, calculated as $\frac{1}{\text{ARR}}$.

Selection Bias

A systematic error resulting in a wrong result due to differences between people who are part of a study versus those who are not.

Intention to Treat Analysis

A methodology where all participants are analysed in the groups to which they were originally randomised, regardless of compliance, to preserve randomisation.

Confounding

A distortion of effects occurring when the relationship between exposure and outcome is confused by a third factor associated with both.

Stratification

The process of dividing study participants into separate groups (strata) based on a third variable to calculate separate association measures.

Effect Modification

A situation where the effect of one factor on an outcome depends on the presence or absence of a second factor.

Direct Standardisation

Applying variable-specific rates from a study population to a single, agreed-upon standard population to produce a standardised rate.

Indirect Standardisation

Applying variable-specific rates from a reference population to the study population to produce a Standardised Incidence Ratio (SIR).

Measurement Bias

Systematic error in estimated frequency or effect caused by measurement error or misclassification of factors.

Validity

The degree to which a measurement or test accurately measures what it is supposed to measure.

Reliability / Repeatability

The ability of a test to produce the same result when used repeatedly to measure the same factor.

Sensitivity

The proportion of people who have the outcome that receive a positive test result ($\frac{\text{Positive test}}{\text{Positive outcome}}$).

Specificity

The proportion of people who do not have the outcome that receive a negative test result ($\frac{\text{Negative test}}{\text{Negative outcome}}$).

Positive Predictive Value (PPV)

The proportion of people with a positive test who truly have the outcome ($\frac{\text{Positive Outcome}}{\text{Positive test}}$), which depends on disease prevalence.

Non-differential Error

When the same scale of measurement error exists in each study group, nearly always biasing results towards the null.

Differential Error

When the measurement error differs between study groups, which can bias results towards or away from the null.

Allocation Concealment

The process of hiding the group to which a subject is allocated from the person recruiting them to prevent selection bias.

Publication Bias

A bias occurring when studies are published or not based on the significance or direction of their results.

Endemic

The constant presence or baseline level of a disease or infectious agent in a specific geographic area or group.

Epidemic

Occurrence of a disease in a specific area and time period above the normal expected baseline.

Outbreak

An epidemic limited to a localized increase in disease incidence, such as in a school or hospital.

Pandemic

An epidemic occurring worldwide or over a very large geographic region across international boundaries.

Case Definition

Specific criteria (time, place, person, clinical/lab findings) used to identify who counts as a 'case' during an investigation.

Forest Plot

A graphical representation used in meta-analyses where boxes represent individual studies and a diamond represents the pooled estimate.

Heterogeneity ($I^{2}$)

A measure of inconsistency between studies in a systematic review; values of $75-100\%$ represent considerable heterogeneity.

Lead Time Bias

Bias occurring when survival time appears longer following screening because detection was earlier, even though the actual time of death remains unchanged.

Length Time Bias

The tendency for screening to preferentially detect slowly progressive diseases because they have a longer window to be picked up.

Wilson and Jungner Principles

The established set of requirements for a screening program, including that the condition must be an important health problem and treatment must be more effective before symptoms occur.

Volunteer bias

Bias arising from the fact that volunteers are generally more healthier