Descriptive and Analytic Epidemiology: Person, Place, Time, Health Indicators, and Study Designs

Vocabulary practice flashcards covering fundamental terminology, demographic measures, vital health indicators, and analytic study designs from Chapters 5, 6, and 7.

Primary objectives of descriptive epidemiology

To describe the extent of public health problems according to person, place, and time; communicate findings with tables and graphs; identify high-risk groups; conduct surveillance; and generate causal hypotheses.

Person descriptors in descriptive epidemiology

Demographic and personal characteristics that address the question of who is affected, including age, sex, gender, race or ethnicity, marital status, family status, occupation, and education.

Family structure and genealogical descriptors in descriptive epidemiology

Descriptive characteristics regarding personal background, including family size, marital status, parental attitude, intelligence level, and family genetic diseases.

Place comparisons in descriptive epidemiology

Geographic analyses that address the question of where events occur, such as comparisons between geographic regions, populations before and after migration, and twins raised in different settings.

Temporal factors in epidemiology

Time-related elements that characterize disease occurrence, ranging from short-term incubation periods lasting hours to long-term chronic latency periods spanning decades.

Epidemiologic impact of the Black Death pandemic (1347-1351)

A historical plague outbreak that killed over $30\%$ of Western Europe (approximately $25{,}000{,}000$ individuals), with over $60{,}000{,}000$ deaths worldwide, including $100{,}000$ deaths in Venice alone.

Data (in public health)

Raw numerical observations or measurements derived from selected variables concerning a health phenomenon of interest.

Information (in public health)

Data that has been contextualized to permit analysis and determine public health meaning.

Message (in public health)

Information combined with interpretation and public health recommendations that explains why people should care and what action must be taken.

Measles progression from Data to Information to Message

Transitions from $1{,}000$ cases in a community (data), to an attack rate of $25$ per $1{,}000$ representing a $50\%$ annual increase alongside $500$ nationwide fatalities (information), to a directive instructing parents to vaccinate infants before age one because vaccination costs less than treating complications (message).

Primary role of tables and graphs in descriptive epidemiology

To summarize, present, and organize complex data sets into easily interpretable formats that highlight population patterns and high-risk groups.

Expansive population pyramid

A population structure featuring a broad base that rapidly narrows toward the top, reflecting high birth rates and high mortality rates.

Stationary population pyramid

A rectangular, uniform age-sex structure typical of nations with low fertility rates and low mortality rates.

Constrictive population pyramid

A population structure with an inward-tapering base, indicating a large older demographic, long life expectancy, low mortality, and low fertility.

Four stages of demographic population pyramid transition

Stage 1 is expansive with a wide base; Stage 2 remains expansive with growing cohorts; Stage 3 becomes stationary with balanced birth and death rates; Stage 4 turns constrictive with an aging base.

Dependency Ratio formula

Calculated as: $\frac{\text{Population } < 15 \text{ and } 65+}{\text{Population } 15\text{--}64} \times 100$.

Interpretation of a $51\%$ Dependency Ratio

There are $51$ dependent individuals (children and seniors) supported by every $100$ individuals of working age.

Projected United States dependency ratios for 2000, 2025, and 2050

A steadily climbing demographic burden measured at $51\%$ in $2000$, projected to rise to $61\%$ in $2025$, and reach $67\%$ in $2050$.

Five-level occupational classification hierarchy

A graded socio-economic scale categorizing work as: 1. Professional, 2. Intermediate, 3. Skilled, 4. Partly skilled, and 5. Unskilled.

Secular trends

Long-term directional movements or changes in health-related events spanning years or decades.

Short-term trends

Sudden, brief, and unexpected increases in health outcomes, typical of localized infectious disease outbreaks and epidemics.

Cyclic trends

Recurrent, periodic rises and falls in health events that happen at regular seasonal or multi-year intervals.

Secular trend of cigarette consumption and lung cancer

A multi-decade pattern demonstrating that population-wide rises and declines in per capita cigarette smoking precede matching shifts in lung cancer mortality by several decades.

Cyclic trend of varicella (chickenpox)

A predictable pattern of pediatric disease incidence that exhibits recurrent multi-year peak cycles and seasonal variations over an $8$-year monitoring span.

Public health surveillance

The systematic, continuous collection, analysis, interpretation, and timely dissemination of health data for public health planning and disease control.

Core purposes of gathering public health surveillance data

To identify sudden shifts in disease incidence, monitor secular trajectories, detect alterations in risk factors, and evaluate the efficacy of interventions.

Public health programs evaluated through surveillance data

Interventions aimed at increasing vaccination coverage, reducing cigarette smoking, increasing fruit and vegetable consumption, increasing physical activity, decreasing obesity, and expanding clinical screening.

Artifactual causes of changes in surveillance case numbers

Non-risk fluctuations caused by revised case definitions, inconsistent interpretation of definitions, modified reporting policies, improved diagnostic tests, increased physician awareness, shifts in population size or composition, changes in active case detection, random events, or heightened public awareness.

Criteria determining the value of a surveillance system

Whether the resulting data informs timely administrative decisions, prompts disease interventions, and aligns resource allocation with actual population need.

Generating etiologic hypotheses from descriptive time patterns

Observing distinct seasonal peaks (such as illnesses clustering exclusively in summer months) to direct researchers toward seasonal environmental factors or insect vectors.

Health indicator

A measurable population variable reflecting wellness, disease status, functional disability, access to care, or health resource distribution.

Seven primary categories of health indicators

Health and well-being, health resources, collective justice, social capital, collective capacity, resiliency, and functionality.

Global life expectancy milestones in 2016

Reached $72.0\,\text{years}$ overall ($69.8\,\text{years}$ for men and $74.2\,\text{years}$ for women), with the WHO African Region experiencing an increase of $10.3\,\text{years}$ from $2000$ to $2016$.

Primary advantage of birth and death metrics over morbidity indices

Vital statistics are legally mandated, standardized, and continuously registered, making them substantially more accessible and complete than morbidity registries.

Contraceptive prevalence

The proportion of women of reproductive age ($15\text{--}49\,\text{years}$) who are using, or whose partner is using, a contraceptive method at a designated point in time.

Crude Birth Rate formula

Calculated as: $\frac{\text{Number of live births during a specified period}}{\text{Total population from which the births occurred}} \times 1{,}000$.

Fertility Rate formula

Calculated as: $\frac{\text{Number of live births during a specified period}}{\text{Total population of women ages } 15\text{--}49} \times 1{,}000$.

Total Fertility Rate replacement threshold

A benchmark of $2{,}100$ births per $1{,}000$ women (or $2.1$ births per woman) required to maintain population stability without immigration.

Three primary categories of mortality causes in society

1. Degeneration of vital organs; 2. Disease states (infections and chronic conditions); 3. Environmental or societal forces (accidents, homicides, and natural disasters).

Jurisdictional authority of the U.S. vital statistics system

Legal authority resides individually within the $50$ states, Washington, D.C., and five U.S. territories, which collect records and report data centrally to the NCHS.

Vital events legally recorded by registry offices

Live births, fetal deaths, legal deaths, marriages, and divorces.

Core variables recorded on standard death certificates

Date of birth (useful for cohort studies), date of death, stated age, place of death, place of residence, occupation, sex, cause of death, marital status, and injury characteristics (type, time, and place of injury).

Underlying cause of death

The specific disease condition or injury that initiated the sequential morbid events resulting directly in a person's death, listed on the first line of the death certificate.

Crude Mortality Rate formula

Calculated as: $\frac{\text{Total deaths occurring during a given time period}}{\text{Total population from which the deaths occurred}} \times 100{,}000$.

Cause-Specific Mortality Rate formula

Calculated as: $\frac{\text{Deaths from a specific cause during a given time period}}{\text{Total population from which the deaths occurred}} \times 100{,}000$.

Infant Mortality Rate formula

Calculated as: $\frac{\text{Deaths among infants ages } 0\text{--}1\,\text{year during a specified period}}{\text{Number of live births in the same period}} \times 1{,}000$.

Neonatal Mortality Rate formula

Calculated as: $\frac{\text{Deaths among infants younger than } 28\,\text{days during a specified period}}{\text{Number of live births in the same period}} \times 1{,}000$.

Postneonatal Mortality Rate formula

Calculated as: $\frac{\text{Infant deaths occurring between } 28\,\text{days and } 1\,\text{year of age}}{\text{Number of live births in the same year}} \times 1{,}000$.

Perinatal Mortality Rate formula

Calculated as: $\frac{\text{Stillbirths plus infant deaths } \le 6\,\text{days of age}}{\text{Total births (live births plus stillbirths)}} \times 1{,}000$.

Maternal Mortality Rate formula

Calculated as: $\frac{\text{Deaths due to pregnancy and childbirth during a specified period}}{\text{Number of live births in the same period}} \times 100{,}000$.

Fetal Death Rate formula

Calculated as: $\frac{\text{Fetal deaths after at least } 20\,\text{weeks of gestation}}{\text{Number of live births plus fetal deaths}} \times 1{,}000$.

Proportionate Mortality Ratio (PMR) formula

Calculated as: $\frac{\text{Deaths from a specific cause during a specified period}}{\text{Total deaths from all causes in the same period}} \times 100$.

Key difference between PMR and Cause-Specific Mortality Rate

The PMR denominator contains only deaths from all causes (a proportional ratio), whereas the mortality rate denominator includes the entire living population at risk.

Worked example: PMR versus Cause-Specific Mortality Rate across two populations

Two cities of $1{,}000{,}000$ people each have cancer death rates of $4$ per $100{,}000$ ($40$ cancer deaths). Metro City has an all-cause mortality rate of $400$ per $100{,}000$ giving a PMR of $\frac{40}{400} \times 100 = 10\%$. Suburban City has an all-cause mortality rate of $900$ per $100{,}000$ giving a PMR of $\frac{40}{900} \times 100 = 4.4\%$. Different total mortality baselines produce unequal PMRs despite identical cause-specific mortality rates.

Death-to-Case Ratio

The total deaths attributed to a disease divided by the number of new cases diagnosed in that same period, used to gauge disease virulence or severity.

Case-Fatality Rate

The proportion of diagnosed individuals with a specific condition who die from that condition within a designated follow-up period.

Years of Potential Life Lost (YPLL)

An epidemiologic index measuring premature mortality by summing the years lost when deaths happen before a predetermined benchmark age, such as $65\,\text{years}$.

Worked example: calculating total and average YPLL

For deaths occurring at ages $20$, $25$, $30$, $35$, and $40\,\text{years}$ with a benchmark of $65\,\text{years}$: Cumulative $\text{YPLL} = (65 - 20) + (65 - 25) + (65 - 30) + (65 - 35) + (65 - 40) = 175\,\text{years}$; Average $\text{YPLL} = \frac{175}{5} = 35\,\text{years}$ per deceased person.

Methodological purpose of the YPLL rate and direct standardization

Compensates for the distortion of comparing crude premature mortality counts across areas with unequal population headcounts (via the YPLL rate) and contrasting age distributions (via direct standardization).

Analytic epidemiology

The study design discipline focused on testing specific a priori hypotheses to identify causal mechanisms and explain why and how health events occur.

Observational analytic versus experimental analytic studies

Observational analytic designs observe exposure and disease relationships without intervention, whereas experimental designs involve active investigator assignment of exposures or treatments.

Case-control study design

An observational study where subjects are sampled based on disease status (cases vs. controls) and previous exposures are determined retrospectively.

Initial step in conducting a case-control study

Establishing explicit diagnostic criteria and a clear, objective case definition for the disease condition being evaluated.

Incident versus prevalent case selection in case-control studies

Incident cases consist of newly diagnosed cases presenting during a defined calendar period, while prevalent cases comprise all existing individuals living with the disease at a single evaluation point.

Case representation and restriction principles

Cases should be representative of all individuals with the condition in the target population; restriction narrows participant eligibility to enhance internal validity, though it limits external generalizability.

Standard sources for case recruitment

Hospitals, medical clinics, HMO registries, public health surveillance logs, and occupational registries.

Fundamental requirement for control selection in case-control studies

Controls must represent the same underlying base population that yielded the cases and possess equal opportunity for past exposure without possessing the disease outcome.

Selection methods for controls in case-control studies

Controls may be identified through random sampling, systematic selection (such as every $n^{\text{th}}$ listed person), restriction criteria, or multiple separate control groups when case numbers are limited.

Advantages and disadvantages of hospital-based controls

Advantages: readily available, low recruitment cost, and high compliance; Disadvantage: their concurrent medical conditions may not accurately represent the exposure baseline of the healthy population.

Advantages and disadvantages of general population controls

Advantages: excellent representation of source population exposure; Disadvantages: high acquisition costs, time requirements, potential nonresponse bias, and poorer recall.

Trade-offs of using special group controls (family and friends)

Advantages: high willingness to participate and control for confounding variables; Disadvantage: overmatching occurs if their exposures match the cases, diluting the estimated association.

Exposure ascertainment sources and protocols in case-control designs

Exposure histories are gathered from medical charts, interviews, self-administered questionnaires, or surrogates (spouses, siblings, employers); procedures and exposure time windows must be strictly identical across cases and controls.

Procedures to reduce exposure ascertainment bias

Blinding investigators and interviewers to case-control status, utilizing standardized survey questionnaires, and keeping identical recall time intervals across all participants.

Odds Ratio formula ($2 \times 2$ contingency table)

Calculated as: $\frac{a / c}{b / d} = \frac{a \times d}{b \times c}$, where $a$ and $c$ are exposed/unexposed cases, and $b$ and $d$ are exposed/unexposed controls.

Interpretation of an Odds Ratio ($OR$)

An $OR = 1.0$ indicates null association; an $OR > 1.0$ indicates an elevated exposure risk; an $OR < 1.0$ indicates a protective association.

Berkson's bias

A selection bias in hospital case-control designs caused by differing hospital admission rates between exposed and unexposed cases and controls.

Prevalence-Incidence bias (Neyman's bias)

A selection bias where prevalent cases overrepresent prolonged or mild disease courses, missing rapidly fatal cases and distorting exposure estimates.

Recall bias

An observation bias occurring when case participants recall past exposures more thoroughly or differently than unaffected control subjects.

Interviewer bias

An observation bias where investigators query, probe, or record exposure information more intensely among cases than among controls.

Differential versus nondifferential misclassification

Differential misclassification occurs when exposure or disease errors differ between comparison groups, whereas nondifferential misclassification occurs uniformly across groups, biasing associations toward the null.

Confounder

An extrinsic variable that is an independent cause of the outcome and is also associated with the exposure, without existing on the intermediary causal pathway.

Matched-pair case-control Odds Ratio formula

Calculated as: $\frac{b}{c}$, using discordant participant pairs where $b$ represents exposed case / unexposed control and $c$ represents unexposed case / exposed control.

Strengths of case-control studies

Well suited for investigating rare disease outcomes, requiring smaller sample sizes, having short study durations, and incurring lower overall financial costs.

Weaknesses of case-control studies

Inability to demonstrate direct temporal sequence, limitation to a single disease outcome, susceptibility to recall bias, and inability to calculate true incidence.

Cohort study design

An observational analytic design where a disease-free group categorized by exposure status is followed forward over time to measure and compare disease incidence.

Selection eligibility and exclusion rules for cohort studies

Study members must be at risk of developing the outcome; exclusions include prevalent cases who already have the condition and people not biologically at risk (such as individuals with surgical organ removal).

Application of restriction in cohort study designs

Selecting cohorts with narrow exposure levels, uniform work environments, or specific behaviors to increase study focus and feasibility while accepting limited generalizability.

Prospective cohort study

A longitudinal study where exposure status is established in the present prior to disease occurrence, and subjects are tracked into the future for incident outcomes.

Retrospective cohort study

A historical study design where exposure status and subsequent disease events are reconstructed using existing historical records covering past time periods.

Trade-offs between prospective and retrospective cohort designs

Prospective cohorts provide greater investigator control over participant selection, exposure classification, and outcome follow-up, but are more expensive, take longer, and assess one exposure; retrospective cohorts are faster, cheaper, and evaluate multiple exposures from existing records, but offer less control over measurement precision.

Cohort effect (generation effect)

Shifts in disease patterns observed as a specific birth or generational cohort moves through successive historical time periods and chronological ages.

Primary measures of incidence in cohort investigations

Cumulative incidence (attack rate) measuring the proportion of at-risk subjects developing disease, and incidence density measuring disease frequency per person-time unit.

Risk Ratio (Relative Risk) formula

Calculated as the ratio of cumulative incidences: $\frac{a / (a + b)}{c / (c + d)}$.

Rate Ratio formula

Calculated as the ratio of person-time incidence rates: $\frac{a / PT_e}{c / PT_0}$.

Relative Risk percentage change formulas

If $RR > 1$, percent increase is: $(RR - 1) \times 100$; if $RR < 1$, percent decrease is: $(1 - RR) \times 100$.

Attributable Risk (Risk Difference) formula

Calculated as: $I_e - I_0$, representing the absolute incidence of disease in the exposed cohort that is directly caused by the exposure.

Attributable Risk Percent (AR%) formula

Calculated as: $\frac{RR - 1}{RR} \times 100$, representing the proportion of illness in exposed individuals that would be prevented if the exposure were eliminated.

Population Attributable Risk (PAR) formula

Calculated as: $I_{\text{total}} - I_0$, quantifying the excess disease incidence across the total general population attributable to an exposure.

Population Attributable Risk Percent (PAR%) formula

Calculated as: $\frac{I_{\text{total}} - I_0}{I_{\text{total}}} \times 100$, representing the overall percentage drop in population disease incidence if the exposure were eliminated.

Healthy worker effect

A selection bias in occupational cohorts occurring when working participants exhibit lower baseline morbidity than the general public, which includes institutionalized and ill individuals.