Observational Studies: Ecological, Cross-Sectional, Case-Control, and Case-Crossover Designs

Comprehensive vocabulary flashcards covering observational study methodologies from Gordis Epidemiology, including case reports, ecological studies, cross-sectional designs, case-control studies, and case-crossover designs, alongside core epidemiological biases and statistical concepts.

Etiology

The cause, set of causes, or manner of causation of a disease or health condition; epidemiological research investigates whether an observed factor-to-outcome relationship represents a genuine causal link.

Observational Study

A study design conducted without active intervention or variable manipulation by the investigator, where exposures and behaviors occur "naturally" in the study subjects.

Hierarchy of Evidence

A pyramidal arrangement that ranks study designs by methodological rigor, spanning from Case Reports at the base, through Case Control Studies, Cohort Studies, and Randomized Controlled Trials, up to Practice Guidelines, Systematic Reviews, and Meta-Analyses at the top.

Case Report / Case Series

A descriptive study at the individual level detailing the clinical course of one or a small group of patients (such as early Zika-microcephaly reports); it lacks a comparison or control group and has very limited generalizability, but serves to generate hypotheses.

Ecological Study

An observational study of group characteristics where the unit of observation and comparison is the population or geographic unit (such as a country) rather than the individual, meaning individual exposures and outcomes are not linked.

Ecological Fallacy

A logical error that occurs when aggregate characteristics of a group are erroneously attributed to an individual member who does not in fact possess those characteristics.

Cross-Sectional Study

An observational study design (also called a prevalence study) that measures exposure and disease status simultaneously at a single point in time ("snapshot" or "slice") for each individual in a defined population.

Prevalence

The proportion of a defined population that possesses a specific disease or condition at a designated point in time, calculated as $\text{Prevalence} = \frac{\text{Number of existing cases}}{\text{Total population}}$.

Cross-Sectional $2 \times 2$ Analysis

The analytical evaluation of a cross-sectional study that compares either the prevalence of disease among exposed individuals ($\frac{a}{a+b}$) versus nonexposed individuals ($\frac{c}{c+d}$), or the prevalence of exposure among diseased individuals ($\frac{a}{a+c}$) versus nondiseased individuals ($\frac{b}{b+d}$).

Survival Bias

A form of selection bias occurring in cross-sectional studies where individuals who died or did not survive long enough are missing from the point-in-time assessment, skewing observed exposure-disease associations.

Correlation Coefficient ($r$)

A numerical statistic ranging from $-1$ to $+1$ that indicates the direction and strength of a linear association between two continuous variables, where values farther from $0$ indicate a tighter scatter along a line.

Positive (Direct) Correlation

A linear association where both variables increase or decrease together, indicated by a positive sign or no sign and a correlation coefficient $r > 0$.

Negative (Inverse) Correlation

A linear relationship where one variable decreases as the other increases, denoted by a negative sign ($-$) and a correlation coefficient $r < 0$ (such as $r = -0.801$ in the exercise and blood sugar dataset).

Case-Control Study

An observational study design that begins by selecting individuals who have the disease (cases) and individuals without the disease (controls), and then looks backward to determine and compare their past exposure histories.

Cohort Study

An observational study design that begins by selecting individuals according to their exposure status (exposed versus unexposed) and follows them over time to compare disease incidence rates.

Retrospective

A temporal study direction that looks backward into the past to gather historical data on exposures, events, and behaviors that occurred prior to the inquiry.

Prospective

A temporal study direction that looks forward into the future, tracking subjects over time to record newly occurring exposures and incident disease outcomes.

Selection Bias

Systematic error introduced by the procedures used to recruit or select study participants (such as hospital-based controls), causing comparison groups to differ systematically in factors other than the disease under study.

Telescoping

A specific limitation in recall where a participant erroneously brings events that occurred in the distant past forward in time, remembering them as having happened more recently.

Confirmation Bias

A form of recall distortion where participants selectively remember exposures or events that reinforce their personal explanations or theories while forgetting details that do not.

Matching

A methodological procedure in case-control studies where controls are selected to match cases on potential confounders (such as age, gender, education, and socioeconomic status) at the individual or group level.

Case-Crossover Design

An epidemiological study design in which each case acts as their own control, comparing exposure levels during an "at-risk period" immediately preceding an acute event (such as an MI) to exposure during an earlier "control period."

Repeated Measures Design

A study design where subjects are assessed multiple times and compared across periods, effectively controlling for inter-individual variability because intra-individual variability is substantially lower.

Helicobacter pylori

The bacterium discovered in $1982$ by Barry Marshall and Robin Warren in peptic ulcer disease (PUD) patients, demonstrating that ulcers stem from bacterial colonization adaptable to gastric acid rather than solely hyperacidity, a discovery awarded the Nobel Prize in Physiology or Medicine in $2005$.