9ª AULA - BIOESTATÍSTICA E EPIDEMIOLOGIA - 52/25

Main ideas & concepts covered

• Course wrap-up (9th and final lesson)

  • Reminded students about deadlines for activities ending in the next two weeks.
  • Course staff thanked Professor Ricardo (instructor) and interpreter Adriana; Professor Débora was also thanked.

• Learning goals for the day

  • Calculate and interpret health indicators (mainly using contingency/2x2 tables).
  • Review epidemiological study designs and identify key characteristics.

• Why epidemiological studies are used

  • Epidemiology aims to address: What causes health problems, illnesses, and deaths in a population?
  • To answer this, epidemiology uses epidemiological studies.

---

Epidemiological study types

1) Observational studies

• Researcher does not interfere with participants’ health.
• Data may be collected retrospectively or prospectively.
• Research focuses on observing final health phenomena and analyzing exposure and outcomes without intervention.

Retrospective definition

• Studies information using data from the present and the past.
• Common methodologies mentioned:
  • Cross-sectional studies
  • Case-control studies
• Common indicators mentioned for these designs:
  • Prevalence indicators

Prospective definition

• Studies the future (follow participants forward in time).
• Common methodology mentioned:
  • Cohort study
• Common concept emphasized:
  • Incidence (incidence rate/measurements)

2) Experimental studies

• Researcher interferes with participants’ health (tests treatments/medications).
• Example used:
  • A participant with lung cancer joined experiments, underwent a medication/testing battery, and was reported as cured.

---

Contingency tables (2x2) and health indicators

Contingency table / “2x2” setup

• Data are arranged into a 2×2 (contingency) table to calculate indicators.
• Indicators depend on properly identifying:
  • Outcome/event (disease occurrence)
  • Exposure factor (risk factor such as smoking)
• The lecturer emphasizes the table can be arranged in two orientations, and if you transpose it, calculations can change.
• Key warning: be careful which value (A, B, C, D) corresponds to each part of the formula.

Indicators calculated and interpreted

• Odds Ratio (OR)

  • Compares the odds of an event between two groups (exposed vs unexposed).
  • Interpretation:
    • OR = 1 → no difference between groups
    • OR > 1 → exposure associated with higher odds of outcome
    • OR < 1 → exposure associated with lower odds of outcome (protective/negative association)
  • Note: discussion included the idea of negative correlation, with direction interpreted as protective vs harmful.

• Prevalence Ratio (PR / prevalence ratio)

  • Compares prevalence in exposed vs unexposed groups.
  • Interpretation:
    • PR > 1 → positive association (higher prevalence among exposed)
    • PR = 1 → no association
    • PR < 1 → negative association / protective effect (lower prevalence among exposed)

• Relative Risk (RR)

  • Compares probabilities of developing disease in exposed vs unexposed groups.
  • Interpretation:
    • RR = 1 → no difference in risk
    • RR > 1 → risk increases with exposure
    • RR < 1 → risk decreases with exposure

• Usage notes

  • OR/PR/RR are described as simpler than hazard ratio / more complex regression tools.
  • Regression tools exist but are less used clinically due to complexity.

---

Worked example (lung cancer and smoking)

Study framing

• Case-control study example:
  • Cases: confirmed lung cancer
  • Controls: individuals without lung cancer
• Smoking treated as the exposure.
• A contingency table is used (values summarized by the lecturer during calculation).

Counts mentioned

From the lecturer’s interpretation:

• Among those who developed lung cancer:
  • 815 smoked
  • 208 did not smoke
• Among those who did not develop lung cancer:
  • 115 smoked
  • 327 did not smoke

Other totals referenced:

• Total population: 1465
• Total lung cancer: 1023
• Total without lung cancer: 442

OR calculation and interpretation

• Computed/rounded result: OR ≈ 11
• Interpretation given:
  • Smokers have ~11 times greater odds/probability (as stated) of developing lung cancer than non-smokers.
• Additional transformation described:
  • Subtract 1 from OR: 11 − 1 = 10
  • Multiply by 100 to express as a percentage (as presented): “1000% greater chance” for exposed vs unexposed.

Prevalence ratio calculation and interpretation

• Computed prevalence ratio: PR ≈ 2.25 (after correcting a mistaken intermediate note)
• Interpretation:
  • Lung cancer prevalence is 2.25 times higher among smokers than non-smokers.
  • Since 2.25 > 1, this indicates a positive association between smoking and lung cancer.

---

Epidemiological study designs (review)

Cross-sectional studies

• Measure prevalence (current state: “at a given time”).
• Used to:
  • Assess prevalence of outcomes (illness, cure, death, survival chances)
  • Identify predictive factors and generate hypotheses
  • Support public health actions and evaluate services/effects
• Data collection characteristics:
  • Collected at one point in time without follow-up of individuals’ entire health trajectory.
• Causal limitation:
  • Exposure and outcome measured at the same time → can’t establish cause-effect, only suggest associations.
• Retrospective element mentioned:
  • Uses present + past information (discussed as “retrospective” character in the talk).

Case-control studies

• Best suited for:
  • Rare diseases or diseases with long latency
  • Outbreaks when extent/complications are unknown
• Key features:
  • Compares cases (with disease) vs controls (without disease), ideally similar in other characteristics.
  • Mentioned as fast and low cost compared to alternatives.
• Temporal framing described:
  • Retrospective: goes from present to past.
• Example referenced:
  • Early COVID-19, when outcomes were still emerging/uncertain (pneumonia-like mention).
• Homework/exercise mention:
  • Multiple-choice question; correct option identified as A for rare disease suitability.

Cohort studies

• Observational and prospective.
• Characteristics:
  • Select a population and follow them over time.
  • No experimental intervention; only observe predictors/exposures and outcomes.
• Uses:
  • Focus on incidence (new outcomes).
  • Aim to identify:
    • Etiology (causes/origins of disease)
    • Risk factors
    • Prognosis (what may happen)
• Claims made:
  • Cohort studies are described as the only observational studies capable of developing etiological hypotheses and testing them statistically.

Confounding and cohort-group similarity

• A quiz response emphasized:
  • Exposed/unexposed groups must be similar or adjusted for confounding factors to avoid biased conclusions about association.

---

Instructional / methodological bullet points (how to compute & interpret indicators)

• Step 1: Build a correct 2×2 contingency table

  • Identify which column/side represents:
    • Outcome/event (disease: yes/no)
    • Exposure/risk factor (exposed: yes/no)
  • Assign the correct cell labels (A, B, C, D) consistent with the formulas being used.
  • Be careful: transposing the table (switching rows/columns) changes where values land and therefore can change computed results.

• Step 2: Identify which indicator matches the study context

  • OR often discussed for case-control style comparisons.
  • PR compares prevalence in exposed vs unexposed groups.
  • RR compares probabilities of disease between exposed and unexposed groups.

• Step 3: Compute Odds Ratio (OR)

  • Use the OR definition as odds of event in exposed group vs odds in unexposed group.
  • Interpret:
    • OR = 1 → no association
    • OR > 1 → exposure increases odds/risk
    • OR < 1 → exposure decreases odds/risk (protective)

• Step 4: Compute Prevalence Ratio (PR)

  • Compute: prevalence in exposed ÷ prevalence in unexposed
  • Interpret:
    • PR > 1 → positive association
    • PR = 1 → no association
    • PR < 1 → negative association / protective

• Step 5: Interpret results in plain-language terms

  • Translate numeric ratios into direction:
    • > 1 → exposure associated with higher prevalence/risk
    • < 1 → exposure associated with lower prevalence/risk

• Step 6: For assignments

  • Exercises 6 and 7 left as homework.
  • Answers to be provided later (via tutoring/moderators).

---

Speakers / sources featured (as named in the subtitles)

• Professor Ricardo (course instructor)
• Professor Débora (appears in greetings/thanks)
• Adriana (interpreter)
• Moderators of the Nursing and Biofar R courses (mentioned as assisting in chat)