W3L1: Study design

Study Design vs. Research Methods

Study Design

• A plan to ensure that the evidence obtained enables us to answer the initial question unambiguously.

• This involves specifying the population, sampling techniques, data collection methods, and analytical strategies used throughout the research process.

Research Methods

• Processes, procedures, and tools used to collect and analyze data.

• Includes sampling mode, data collection methods (e.g., questionnaire, observation), design of questions, and statistical analysis plans, which always confirm with study design.

Comparison of Study Designs

• Research Question/Problem: Determines the most appropriate

• Study Design: Implements specific types of study plans to determine expected results.

• Research Method: Affects the selection of suitable methods to collect and analyze evidence.

Study Designs

Types of Study Designs

Descriptive Studies

• Hypothesis generating without a comparator.

Analytical Studies

• Hypothesis testing with comparator group(s).

Descriptive Studies

• Describe characteristics of disease (outcome) or exposure (risk factor) concerning:

  • Populations: Demographics, socio-economic status, lifestyles.

  • Geographic Distribution: Place-based variations.

  • Frequency Over Time: Seasonal variations through time-trend studies.

• Research questions in format (PO question).

• Generate hypotheses for later testing in analytic studies.

Types of Descriptive Studies

Case Reports and Case Series

• Detailed descriptions of one or few patients with:

  • Unusual diseases or unexpected outcomes.

  • Speculated exposure leading to the disease (e.g., new syndrome in gay men with PCP pneumonia).

• Time Trend Studies: Measure disease occurrence over time to assess fluctuations correlated with community changes.

Cross-Sectional Studies (Surveys)

• Examine disease prevalence at one point in time by analyzing variables from a population sample.

• Gather data on exposure and disease simultaneously

Cross-Sectional Ecological Studies

• Population-level analysis rather than individual-level data.

• Example Question: Relationship between income and coronary heart disease (CHD) rates across different countries.

  • Findings suggested richer countries had higher CHD rates, but individual-level data revealed the opposite.

• Ecological fallacy: an assumption that an association observed at the group level applies at the individual level. → major weakness

Descriptive Studies: Advantages and Disadvantages

Advantages

• Quick and inexpensive.

• Utilize existing databases.

• Efficient resource allocation for education and health promotion.

• Answers non-clinical research questions

Disadvantages

• Cannot establish causality; conclusions can be misleading.

• Prone to ecological fallacy.

Analytical Studies

Observational Studies

• Primarily hypothesis testing using comparator groups.

• Goal: Determine if interventions or exposures affect outcomes.

Advantages and Disadvantages of Observational Studies

Advantages

• No intervention or manipulation is required—just observation.

• Useful for studying naturally occurring exposures or conditions where randomization is unethical.

• Efficient sampling opportunities (e.g., studying rare diseases).

Disadvantages

• More prone to bias and confounding than experimental designs.

Axes of Observational Analytical Study Designs

• Directionality

  • Exposure first → Cohort

  • Outcome first → Case-control

  • Both together → Cross-sectional

• Sample selection

  • Exposure → Cohort

  • Outcome → Case-control

  • Random/all → Cross-sectional

• Type of outcome

  • Incident → Cohort

  • Prevalent → Case-control/Cross-sectional

Cross-Sectional Studies

• Asses the prevalence of a condition at a specific point in time by collecting data from a representative sample.

• Shows correlations between exposure and disease

  • no directionality - do not track changes over time or determine cause-and-effect relationships.

Steps in Analytical Cross-sectional Study

1. select a sample from population

2. measure predictor and outcome variable at a single time point

3. estimate prevalence of outcome

4. asses correlation between exposure and outcome

   

Example Question:

• Research Question: In patients with Polycystic Ovarian Syndrome, is neuro-endocrine dysfunction associated with irregular cycles? What is the study design?

  • Cross-sectional analytical study

Case-control Studies

• Examine the people with/without disease and find differences in predictor variables that may explain why the cases got the disease and the control didn’t.

• Usually retrospective.

  • look back in time to examine how certain exposure/risk factors may have influenced the occurrence of the outcome

Steps in a Case-control Study

1. Classifies individuals based on presence/absence of disease

   1. select a sample from a population of people with the diseases (cases)

   2. select a sample from a population at risk that are free of the diseases (controls)

   3. each case is matched with a similar control

2. Compare exposure history between cases and controls

Study Base

• a defined study population

• both cases and controls should be selected from the same study base

• important to minimize bias

Case-control Example Question

• Research Question: In patients with Reye’s syndrome, is there an association between the use of aspirin and the development of Reye’s syndrome?

  • 30 patients with Reye’s syndrome were accessible to the investigator for study

  • 60 patients were drawn from the much larger population of accessible patients who have had minor viral illnesses without Reye’s syndrome.

  • Subjects in both groups asked about history of aspirin use.

  • Subjects who had minor viral illness and took aspirin had 9 times the odds of developing Reye’s syndrome.

• What is the study design?

  • Case-control

Odds Ratio

• Measure of strength of association

• Odds in cases/Odds in controls

  • a/b / c/d

• can be greater than/less than 1

  • OR > 1 = Positive association

  • OR = 1 = No association

  • OR < 1 = Negative association

Advantage

• Quick & Inexpensive

• only method for studying rare disorders with long lag times between exposure&disease

• fewer subjects require

• useful for answering clinical questions about etiology and harm

Disadvantages

• susceptible to bias

• potential for confounding

• can only study one outcome

• cannot calculate incidence rates for exposed vs unexposed patients

Cohort Studies

Design

• Individuals classified by exposure status: can be prospective or retrospective.

Purpose

• Examine incidence of disease over time, collecting data on outcomes compared to initial predictors.

  • prospective → investigator defines the sample and collects data about the predictor before any outcomes have occurred.

  • retrospective → investigator defines the sample and collects data about the predictor variable after the outcomes have occurred.

Steps in a Prospective Cohort Study

1. Select a sample from the study cohort

2. eligible participants need to be free from the disease under study

3. measure exposure first (risk factor present/absent at the time of study onset)

4. follow-up the cohort overtime until outcome occurs

5. measure outcome variable (developed disease or did not develop disease)

Relative Risk

• Measure of strength of association

• RR>1 = positive association, higher risk

• RR=1 = no association

• RR<1 = negative association, exposure is protective

Steps in a Retrospective Cohort

1. Identify a cohort that had been assembled in the past (historical cohort)

2. Eligible participants need to be free from the disease under study at that past period

3. collect data on predictor variables (exposure measured first, but in the past)

4. Follow-up the cohort regarding outcome occurrence

5. Collect data on outcome variables (measured in the past as per medical records or in the present)

Advantages

• allows measurement of mortality rates in exposed vs unexposed

• subjects can be matched for possible confounders

• can study rare exposure

• can examine multiple outcomes

• easier and cheaper

• useful for answering clinical questions about etiology, harm, or prognosis

Disadvantages

• expensive and time consuming

• lost to follow-up → serious threat to validity

• potential for bias and contamination

• blinding of subjects and investigators difficult

• inefficient for uncommon outcomes

Experimental Studies

Randomized Controlled Trials (RCTs)

• Investigates the cause-effect relationship by randomly allocating participants to either intervention or control groups.

Steps in an RCT

1. Select a sample and

2. randomly divide it into experimental and control groups.

3. The experimental group receives an intervention

4. Measure outcome variables on both groups.

Advantages

• minimal bias

• gold standard design for examining treatment efficacy

Disadvantages

• Expensive

• Time consuming

• Loss to follow-up

Quasi-Experimental Studies

Description

• Lack random assignment; groups are pre-existing, limiting ability to infer cause-effect relationships.

Example

• Investigating pain levels in burn patients by comparing treatment across different centers.

Advantages

• practical when random assignment is not feasible due to ethical or practical constraints

Disadvantages

• Highly [one to bias and confounding due to lack of randomization → provides lower quality evidence

Conclusion

Types of Clinical Questions

• Etiology/Harm: Questions regarding negative impacts from interventions.

• Therapy: Questions surrounding treatment efficacy.

• Prognosis: Questions about disease progression.

• Diagnosis: Questions about identifying disorders in patients presenting with symptoms.

Selection of Study Designs

• Different study designs optimize answers to varying clinical questions:

  • Etiology (cohort, case-control),

  • Therapy (RCT),

  • Prognosis (cohort),

  • Harm (cohort, case-control),

  • Diagnosis (cross-sectional).