Experimental and Observational Studies

• Study designs include experimental and observational studies

Experimental Studies

• An investigator manipulates an intervention and controls its distribution
• Used to establish causality

Observational Studies

• There is no intervention manipulation
• An investigator observes exposures and outcomes
• Types of observational studies include cross-sectional, cohort, case-control and ecological

Cross-Sectional Study

• Snapshot of exposure and outcome at a single point in time

Cohort Study

• Follows a group over time to see who develops the outcome

Case-Control Study

• Compares people with and without a disease to find past exposures

Ecological Study

• Uses group-level data instead of individual data

Key Study Design Differences

• Randomized controlled trials use randomization, quasi-experimental studies do not
• Cross-sectional assess prevalence at one time, while cohort studies follow participants over time
• Cohort studies start with exposure and follow forward, case-control studies start with outcome and look back

Randomized Controlled Trials (RCTs)

• Randomization eliminates selection bias
• Randomization controls for confounders, both known and unknown
• Randomization ensures valid statistical comparisons

Types of Randomization

• Simple randomization is like flipping a coin
• Blocked randomization ensures equal group sizes
• Stratified randomization balances key variables like age and disease severity

Control Groups in RCTs

• Placebo control uses an inactive substance
• Active control uses an existing treatment
• Delayed intervention controls get treatment later
• Usual care control is the standard treatment used in clinical practice

Allocation Concealment vs. Blinding

• Allocation concealment prevents selection bias before randomization
• Allocation concealment ensures researchers are unable to influence group assignment
• Example of allocation concealment includes sealed opaque envelopes, computer-generated randomization
• Blinding prevents bias after randomization

Types of Blinding

• Single-blind: Only participants are unaware
• Double-blind: Participants and researchers are unaware
• Triple-blind: Participants, researchers, and analysts are unaware
• Blinding is not always feasible, such as in surgery trials

Measures of Disease Frequency

• Prevalence measures existing cases at a point in time
• Types of prevalence include point and period prevalence
• Point prevalence is a snapshot of disease at a specific time
• Period prevalence measures cases over a set period
• Incidence measures new cases over time
• Types of incidence include Cumulative Incidence (Risk) and Incidence Rate (Incidence Density)
• Cumulative Incidence formula: New cases / Population at risk
• Incidence Rate formula: New cases / Total person-time at risk

Measures of Association

• Risk Ratio (Relative Risk) is used in cohort studies
• Risk Ratio formula: Risk in exposed / Risk in unexposed
• Risk Ratio interpretation: RR > 1 = Exposure increases risk; RR < 1 = Exposure decreases risk
• Odds Ratio is used in case-control studies
• Odds Ratio formula: (a x d) / (b × c) (from a 2x2 table)
• Odds Ratio interpretation: OR > 1 = Higher odds of disease in exposed group
• Prevalence Proportion Ratio and Prevalence Odds Ratio are used in cross-sectional studies
• Prevalence Proportion Ratio Formula: Prevalence in exposed / Prevalence in unexposed
• Prevalence Odds Ratio Formula: (a × d) / (b x c)
• Prevalence Odds Ratio is used when prevalence is high

Bias and Limitations

• Key biases in studies include selection, information, confounding, and survivor bias
• Selection bias is due to improper participant selection
• Information bias includes errors in measuring exposure or outcome
• Confounding occurs when another factor influences the relationship
• Survivor bias, (incidence-prevalence bias), means that disease cases that die quickly may not be captured in cross-sectional studies
• Ecological fallacy occurs in ecological studies
• Applying group-level associations to individuals incorrectly is an ecological fallacy
• Example of ecological fallacy: Cities with high ice cream sales have higher drowning rates (confounder: hot weather)

Choosing the Right Statistical Test

• Chi-Square Test compares percentages or frequencies in two or more groups (Categorical Data)
• T-Test compares means between two groups (Continuous Data)
• ANOVA compares means across three or more groups
• Regression Analysis is used to measure the effect of multiple variables on an outcome

Study Design Applications

• RCTs are used when testing a new intervention
• Cohort studies are used when studying risk factors over time
• Case-control when the disease is rare and you want to look back at exposures
• Cross-sectional studies are used when estimating disease prevalence
• Ecological studies are used when using group-level data for comparisons

This material discusses experimental and observational study designs. Experimental studies involve intervention manipulation to establish causality. Observational studies, including cross-sectional, cohort, case-control, and ecological studies, observe exposures and outcomes without intervention.