Case Control Study Explained: Concept, Design & Odds Ratio Numericals | Dr. Priyanka Sachdev

Main ideas / lessons conveyed

• Purpose of epidemiological studies

  • Epidemiological studies are conducted to prove a hypothesis.
  • A hypothesis typically links a cause/exposure/risk factor to an effect/disease (e.g., smoking → lung cancer).

• Types of epidemiological studies

  • Observational studies
    • Descriptive studies: formulate the hypothesis by describing time, place, person; do not measure association.
    • Analytic studies: test the hypothesis by measuring association (focus of the session).
  • Experimental studies
    • Example: Randomized Controlled Trial (RCT), where the investigator intervenes.

• Analytic studies: Case-control vs Cohort

  • Case-control study
    • Starts with disease/outcome status (cases vs controls) and looks back to past exposure.
    • Retrospective / backward-looking
    • Described as effect → cause.
  • Cohort study
    • Starts with exposure status (exposed vs unexposed) and follows people forward to see who develops disease.
    • Prospective / forward-looking
    • Cause → effect (outcome in the future).

---

Case-control study: concept, design, and workflow

Key concept

• Called “case-control” because it compares:
  • Cases = people with the disease present
  • Controls = people with the disease absent
• The study then compares prior exposure (e.g., smoking history) between cases and controls.

Step-by-step methodology (as taught)

1. Select the population

   • Choose a defined population (e.g., Delhi/Mumbai/community).

2. Select cases

   • Case definition: a person in whom the disease under study is present.
   • Must use diagnostic/eligibility criteria (e.g., lung cancer confirmed by biopsy/CT criteria as appropriate).
   • Example: take 100 lung cancer patients from a cancer hospital.

3. Select controls

   • Control definition: a person in whom the disease under study is absent.
   • Other diseases may exist, but the target disease (e.g., lung cancer) must be absent.
   • Matching idea: make cases and controls similar in factors like age group and gender (and generally other confounders) except for disease status.

4. Measure exposure (ask for history)

   • Collect exposure history for each participant (e.g., smoking in the last 10–15 years):
     • Exposure presence/absence (smokes vs does not smoke)
     • Optionally intensity/frequency (e.g., cigarettes per day)

5. Create the 2×2 classification

   • Divide into four groups:
     • A: exposed + disease present
     • B: exposed + disease absent
     • C: unexposed + disease present
     • D: unexposed + disease absent

6. Decide retrospective design

   • If you start at the present (disease already diagnosed) and go back using history, it is:
     • Retrospective (backward-looking)
     • Effect → cause

---

Cohort study: design contrast (as taught)

Detailed design points

• Start point = exposure

  • Take exposed group (e.g., smokers) and unexposed group (non-smokers).
  • Do not intervene; observe naturally (as described in the lecture).

• Follow-up

  • Follow for a long time (e.g., 10 years).
  • Periodically check whether participants develop the disease.

• Therefore

  • Prospective (forward-looking)
  • Cause → effect

• Primary contrast in timing

  • Case-control: start with disease → look back for exposure
  • Cohort: start with exposure → observe disease in the future

---

Advantages & disadvantages (comparison provided)

Case-control study

• Advantages
  • Fast (can be completed quickly; lecture implies “in one day” for concept level)
  • Inexpensive/low cost
  • No ethical issue (you’re not intervening; you’re asking history)
• Disadvantage
  • Recall bias
    • Exposure is based on participant memory/history, which may be inaccurate.

Cohort study

• Advantages
  • No recall bias
    • Disease occurrence is observed during follow-up rather than relying solely on memory.
• Disadvantages
  • Time-consuming (requires long follow-up)
  • Expensive (calls, visits, emails, etc.)
  • Possible ethical concerns
    • In smokers vs non-smokers, investigators may know risks but are not intervening in the described cohort setup.

---

Analysis in case-control: exposure rate and Odds Ratio (OR)

1) Build the 2×2 table (required rule)

• Disease categories across the top
  • Disease present | Disease absent
• Exposure categories along the side
  • Exposure present | Exposure absent
• Cell mapping:
  • A (exposed + disease present)
  • B (exposed + disease absent)
  • C (unexposed + disease present)
  • D (unexposed + disease absent)

2) Exposure rate (association present or absent)

• Exposure rate among cases = A / (A + C)

• Exposure rate among controls = B / (B + D)

• Decision rule

  • If exposure rate among cases > exposure rate among controls → association present
  • If exposure rate among cases < exposure rate among controls → association absent/negative

3) Odds Ratio (strength/direction of association)

• Odds Ratio formula (cross-product ratio):

  • OR = (A × D) / (B × C)

• Interpretation

  • OR = 1 → no association
  • OR > 1 → positive association (higher odds of disease with exposure)
  • OR < 1 → negative association / protective effect (lower odds of disease with exposure)

---

Numerical example (Odds Ratio) solved in the lecture

• Context: smoking effect on renal cell carcinoma

• Given

  • 50 patients (disease present):
    • 30 smokers → A = 30
    • Total disease present A + C = 50 → C = 20
  • 50 controls (disease absent):
    • 10 smokers → B = 10
    • Total controls B + D = 50 → D = 40

• Odds Ratio

  • OR = (A×D)/(B×C) = (30×40)/(10×20) = 1200/200 = 6

• Result

  • OR = 6 (strong positive association in this example)

---

Exam-focused MCQ conclusions included

• OR = 1 implies no association.
• Case-control is suitable for rare disease.
• Cohort is suitable for rare exposure.
• Recall bias is a key drawback of case-control (so “minimal bias” is not correct).
• Odds ratio (as cross-product) is tied to the case-control 2×2 setup.

---

Speakers / sources featured

• Dr. Priyanka Sachdev (speaker/instructor for the session)