4.2 Descriptive study designs

Descriptive Study Designs

Descriptive study designs help epidemiologists understand how diseases are distributed across populations. They answer the "who, what, where, and when" of disease occurrence, and they're often the first step in investigating a health problem. While they can't prove causation, they generate the hypotheses that analytical studies then test.

Three core descriptive designs come up repeatedly in epidemiology: cross-sectional studies, ecological studies, and case series. Each collects and uses data differently, and each has trade-offs you need to know.

Types of Descriptive Study Designs

Cross-sectional studies measure exposure and outcome at the same point in time, giving you a snapshot of a population. They're the go-to design for estimating prevalence, which is how common a disease or condition is in a population at a given moment. A national survey measuring hypertension rates across age groups is a classic example.

• Useful for health planning and resource allocation
• Cannot establish causality because exposure and outcome are measured simultaneously, so you can't tell which came first (this is called temporal ambiguity)

Ecological studies use aggregate data rather than individual-level data. Instead of tracking what each person was exposed to, you compare groups or populations. For instance, you might compare average air pollution levels and asthma hospitalization rates across 50 cities.

• Efficient for studying large populations and generating hypotheses
• Prone to the ecological fallacy: just because a relationship exists at the group level doesn't mean it holds for individuals within those groups. A city with higher pollution and higher asthma rates doesn't prove that the specific people breathing more pollution are the ones getting asthma.

Case series describe a group of patients who all have the same condition, without any comparison group. Early reports on COVID-19 in Wuhan during January 2020 were case series describing the clinical features of the first 41 hospitalized patients.

• Valuable for identifying and characterizing new or rare conditions
• Cannot calculate risk or incidence rates because there's no comparison group of people without the condition

Strengths vs. Weaknesses of Study Designs

Cross-sectional studies

• Strengths: Relatively quick and inexpensive. Can study multiple exposures and outcomes at once. Provide prevalence data that directly informs community health planning.
• Weaknesses: Cannot determine causality. Susceptible to prevalence-incidence bias (also called Neyman bias), where you only capture people who currently have the disease, missing those who recovered quickly or died. You also miss how conditions change over time.

Ecological studies

• Strengths: Can efficiently cover entire populations using existing data (like census or environmental monitoring data). Useful for evaluating population-level interventions such as vaccination programs or seatbelt laws.
• Weaknesses: The ecological fallacy is the big limitation. You also can't control for individual-level confounders like socioeconomic status or personal behaviors, since you only have group-level data.

Case series

• Strengths: Easy and inexpensive to conduct. Often the first signal that something new is happening, such as an emerging infectious disease or an unusual drug reaction. Good for generating hypotheses about rare conditions.
• Weaknesses: No comparison group means you can't determine whether outcomes are more common in exposed vs. unexposed people. Prone to selection bias, especially when cases come from a single hospital that may not represent the broader patient population.

Applications of Descriptive Studies

Cross-sectional studies are commonly used to:

1. Estimate disease prevalence (e.g., surveying diabetes rates in a county)
2. Assess health-related behaviors (e.g., measuring smoking rates among teenagers)
3. Evaluate health service needs (e.g., mapping access to primary care in rural areas)

Ecological studies are commonly used to:

1. Investigate environmental exposures (e.g., comparing air pollution levels and respiratory disease rates across regions)
2. Evaluate public health interventions (e.g., whether states with seatbelt laws have lower traffic fatalities)
3. Study rare exposures or outcomes that would require enormous individual-level studies (e.g., radiation exposure and cancer rates near nuclear facilities)

Case series are commonly used to:

1. Describe clinical features of new diseases (e.g., early COVID-19 case reports in 2020)
2. Report on rare conditions or unusual clinical presentations (e.g., a cluster of patients with an unrecognized genetic disorder)
3. Generate hypotheses about treatment outcomes (e.g., documenting responses to a novel cancer therapy before a formal trial)

Role in Hypothesis Generation

Descriptive studies don't test hypotheses; they generate them. Here's how each function feeds into that process:

• Identifying patterns and trends. Descriptive data can reveal exposure-outcome associations worth investigating. The early observation that lung cancer rates were higher among smokers came from descriptive work before any analytical study confirmed causation.
• Characterizing populations. Baseline data on disease distribution helps identify high-risk groups. Age-specific cancer incidence data, for example, can point researchers toward particular age groups for further study.
• Raising new research questions. Sometimes unexpected findings spark entirely new lines of inquiry. Descriptive observations about lower diabetes rates among coffee drinkers led to analytical studies exploring that relationship.
• Informing analytical study design. Descriptive studies help researchers determine appropriate sample sizes, select key variables, and refine their research questions before launching more expensive cohort or case-control studies.
• Prioritizing public health resources. By revealing which problems are most common or most severe, descriptive studies help direct research funding and public health attention where it's needed most.