Studi Deskriptif

01Overview and Core Concept

Descriptive studies form the foundation of epidemiological research. They answer the most basic question in public health: What is happening? They do not answer why it is happening. This distinction is the single most tested concept regarding descriptive studies on the exam.

The framework used to organize descriptive data is Person, Place, and Time. "Person" variables include age, sex, race, occupation, and socioeconomic status. "Place" variables include geographic region, urban versus rural setting, and country. "Time" variables include seasonal trends, secular (long-term) trends, and epidemic curves. Every descriptive study collects data along at least one of these axes.

Because descriptive studies lack a comparison group, they cannot calculate relative risk, odds ratios, or attributable risk. They can only estimate disease frequency. Any question that presents a study calculating prevalence from a single survey and then asks whether the study can prove causation is testing your understanding of this limitation. The answer is always no.

Descriptive studies are considered hypothesis-generating. They observe a pattern (for example, a cluster of liver cancer cases near an industrial site) and propose a possible explanation that must then be tested by an analytical study. The exam frequently frames a vignette where an investigator notices something and then asks what the "next best step" is. When the current study is descriptive, the next step is to design an analytical study (case-control or cohort) to test the hypothesis.

02Types of Descriptive Studies

Case Report is the simplest form of medical literature. It describes something unusual or new in a single patient. It has no denominator, no rate, and no statistical analysis. Its value is in alerting the medical community. The classic example tested on exams is identifying a novel adverse drug effect.

Case Series expands on the case report by collecting multiple patients with the same condition or exposure. It still has no control group. It can describe patterns (for example, "7 out of 12 patients with this condition were male and presented before age 40") but cannot compare those patients to anyone. A common exam trap is presenting a case series and asking whether it proves an association. It does not.

Cross-Sectional Study (also called a prevalence study or survey) is the most commonly tested descriptive study. It measures exposure and disease status at the same point in time. Because both are measured simultaneously, you cannot determine which came first, meaning temporality cannot be established. This is why cross-sectional studies measure prevalence, not incidence. Prevalence captures existing cases at one moment, while incidence requires following a population over time to detect new cases. The formula for prevalence is:

$$Prevalence=\frac{\text{Number of existing cases at a point in time}}{\text{Total population at the same point in time}}Prevalence\; =\; \backslash frac\{\backslash text\{Number\; of\; existing\; cases\; at\; a\; point\; in\; time\}\}\{\backslash text\{Total\; population\; at\; the\; same\; point\; in\; time\}\}$$

A key nuance: cross-sectional studies can calculate an odds ratio when analyzing the association between exposure and disease, but they still cannot establish a causal relationship because the temporal sequence is unknown.

Ecological Study (also called a correlational study) uses group-level data rather than individual-level data. Instead of asking individual patients about their diet and checking their disease status, an ecological study compares average dietary patterns of entire countries against national disease rates. The fatal flaw of ecological studies is the ecological fallacy: the assumption that associations observed at the group level apply to individuals. For example, a country with high average sugar consumption and high diabetes rates does not prove that the individuals who consumed more sugar are the ones who developed diabetes.

03Measures of Disease Frequency in Descriptive Studies

Prevalence and incidence are related by the equation:

$$Prevalence\approx Incidence\times DurationPrevalence\; \backslash approx\; Incidence\; \backslash times\; Duration$$

This relationship is tested frequently. If a disease has a long duration (e.g., diabetes), its prevalence will be much higher than its incidence because patients accumulate in the "existing cases" pool. If a disease is rapidly fatal (e.g., pancreatic cancer), its prevalence will be lower relative to its incidence because patients exit the pool quickly through death.

A common exam question will describe a new treatment that extends survival for a chronic disease and ask what happens to prevalence and incidence. The answer: incidence stays the same (the treatment does not prevent new cases), but prevalence increases (patients live longer with the disease, so existing cases accumulate).

04Strengths and Limitations

Survival bias in cross-sectional studies deserves emphasis. Because a cross-sectional study captures only people who are alive and present at the time of the survey, it systematically misses patients who died before the survey. This means diseases with high early mortality will appear less prevalent than they actually are. It also means that chronic, non-fatal conditions will be overrepresented. The exam tests this concept by describing a survey that finds a surprisingly low prevalence of an aggressive disease and asking what explains the finding.

05Comparison with Analytical Studies

The hierarchy of evidence frequently appears on the exam. From lowest to highest: case report, case series, cross-sectional, case-control, cohort, randomized controlled trial, systematic review/meta-analysis. Descriptive studies occupy the bottom tiers. A common question format is asking which study design provides the "strongest evidence" for a given clinical question, and the answer is almost never a descriptive study unless the question is about estimating prevalence.

One important exception: when the exam asks what study design should be used to estimate the prevalence of a disease in a population, the correct answer is cross-sectional study, not a cohort or RCT. The study design must match the research question.

06Traps and High-Yield Pearls

The most common mistake students make with descriptive studies is confusing association with causation. When a cross-sectional study shows that smokers have a higher prevalence of depression, students are tempted to conclude that smoking causes depression. The correct interpretation is that there is an association, but the temporal relationship is unknown: depression may have preceded smoking, smoking may have preceded depression, or both may be caused by a shared confounder.

The second major trap involves ecological studies and the ecological fallacy. If a question states that countries with higher ice cream sales have higher drowning rates, students may conclude that ice cream causes drowning. The correct reasoning is that both variables are driven by a confounder (hot weather) and that the group-level correlation does not apply to individuals.

A third trap concerns the cross-sectional study's relationship to incidence. Some vignettes will describe a single survey and ask for the "incidence" of disease. If only one measurement was taken at one point in time, the study can only measure prevalence. Incidence requires follow-up over time to detect new cases.

Finally, remember that when a question describes a physician who notices an unusual cluster of cases and asks what the "first step" is, the answer is usually a case series to characterize the pattern. The "next step" after the case series is an analytical study (typically case-control for rare diseases or cohort for common exposures) to test the hypothesis that emerged from the descriptive observation.

The core competency being tested across all descriptive study questions is your ability to match the correct study design to the research question, recognize what each design can and cannot conclude, and identify the logical next step in the research sequence.