Click image to open full size
Study Design
RCT, Cohort, Case-Control, Cross-Sectional
Related Tools
Related Labs
Related Worksheets
Medical research uses different study designs to answer different kinds of questions about treatment, risk, and disease patterns. Knowing how randomized controlled trials, cohort studies, case-control studies, and cross-sectional studies work helps students judge the strength of evidence. These designs differ in how participants are selected, whether exposure is assigned or observed, and whether time moves forward, backward, or stays at one point. Choosing the right design affects cost, speed, bias, and how confidently we can infer causation.
A randomized controlled trial assigns an intervention and compares outcomes between groups, making it the strongest design for testing treatment effects. Cohort studies begin with exposure status and follow people to see who develops an outcome, while case-control studies begin with outcome status and look back for prior exposures. Cross-sectional studies measure exposure and outcome at the same time, making them useful for estimating prevalence but weak for establishing temporal order. Students should connect each design to its main measure, common bias, and best clinical use.
Key Facts
- RCT: participants are randomly assigned to intervention or control, which reduces confounding and supports causal inference.
- Cohort study: starts with exposed vs unexposed groups and usually estimates risk ratio, RR = incidence in exposed / incidence in unexposed.
- Case-control study: starts with cases and controls and usually estimates odds ratio, OR = odds of exposure in cases / odds of exposure in controls.
- Cross-sectional study: measures exposure and outcome at one time point and is commonly used to estimate prevalence, prevalence = existing cases / total population.
- Incidence describes new cases over time, while prevalence describes all existing cases at a specific time.
- Temporality matters: RCT and prospective cohort designs can show exposure before outcome, but cross-sectional studies usually cannot.
Vocabulary
- Randomization
- Randomization is the process of assigning participants to groups by chance to reduce systematic differences between them.
- Confounding
- Confounding occurs when a third factor is associated with both the exposure and the outcome and distorts the true relationship.
- Incidence
- Incidence is the number of new cases of a disease that develop in a population during a specified time period.
- Prevalence
- Prevalence is the proportion of a population that has a disease or condition at a given time.
- Odds ratio
- An odds ratio compares the odds of prior exposure in cases with the odds of exposure in controls.
Common Mistakes to Avoid
- Assuming every observational study can prove causation, which is wrong because lack of random assignment leaves more room for confounding and bias.
- Confusing incidence with prevalence, which is wrong because incidence counts new cases over time while prevalence counts all existing cases at one point or period.
- Using risk ratio as the standard measure in a case-control study, which is wrong because investigators usually do not directly measure incidence and instead calculate an odds ratio.
- Thinking cross-sectional studies show which came first, which is wrong because exposure and outcome are measured at the same time so temporality is usually unclear.
Practice Questions
- 1 An RCT tests a new antihypertensive drug in 200 patients. One hundred receive the drug and 100 receive placebo. After 6 months, 12 patients in the drug group and 24 in the placebo group have uncontrolled blood pressure. Calculate the risk in each group and the risk ratio.
- 2 In a cohort study, 300 smokers and 500 nonsmokers are followed for 10 years. Lung disease develops in 45 smokers and 20 nonsmokers. Calculate the incidence in each group and the risk ratio.
- 3 A researcher surveys 1000 college students once and records vaping status and current asthma symptoms on the same day. Identify the study design and explain one major limitation of this design for deciding whether vaping caused the symptoms.