Foundations of Epidemiology: Hypothesis Testing and P-Values

Disease Prevention Strategies

• Primary Prevention: Focus on preventing disease occurrence.

  • Examples: Immunizations, tobacco control programs.

• Secondary Prevention: Early detection in clinical settings or through screening programs.

• Tertiary Prevention: Improve treatment, minimize morbidity, and prevent new sequelae.

  • Example: Development of novel therapies.

Validity and Reliability of Tests

• Validity: Measures whether the test accurately indicates disease presence or absence.

• Reliability: Assesses if test results are consistent upon repeated measures.

  • Intra-subject variability: Fluctuation in a variable over time.

  • Intra-observer variability: A single observer gets different results over time.

  • Inter-observer variability: Different observers obtain varying results on the same individual.

Evaluating Screening

• 2 x 2 Table for Evaluating Screening:

  • Must know true disease state based on a gold standard.

  • Test results compared to findings from new tests.

Validity Measures

1. Sensitivity: % of people with disease who test positive.

   • Formula:

     • Sensitivity = [True Positives / (True Positives + False Negatives)] x 100

2. Specificity: % of people without disease who test negative.

   • Formula:

     • Specificity = [True Negatives / (False Positives + True Negatives)] x 100

3. Positive Predictive Value (PPV): Likelihood that a positive test accurately indicates disease.

   • Formula:

     • PPV = [True Positives / (True Positives + False Positives)] x 100

4. Negative Predictive Value (NPV): Likelihood that a negative test indicates disease absence.

   • Formula:

     • NPV = [True Negatives / (True Negatives + False Negatives)] x 100

Example Scenario for Validity

• 50,000 study subjects with a 0.5% prevalence of colorectal cancer.

• 1,800 subjects tested positive using a new molecular screening test; 200 confirmed via gold standard.

Applying Given Data to Calculate Sensitivity, Specificity, and PPV

• Sensitivity:

  • (200 disease positives / 250 total disease state) x 100 = 80%

• Specificity:

  • (48150 true negatives / 49750 total no disease) x 100 = 96.78%

• Positive Predictive Value:

  • (200 true positives / 1800 total positives) x 100 = 11.11%

Prevalence Impact on PPV

• PPV influenced mainly by specificity and disease prevalence.

Statistical Foundations in Hypothesis Testing

• Null Hypothesis (H0): Assumes no association between exposure and disease.

  • Relative Risk or Odds Ratio = 1.0.

• Alternative Hypothesis (Ha): Suggests there is an association.

• P-value indicates the probability of observing a test statistic as extreme as the one calculated under the null hypothesis.

• **Type I Error (α)**: Incorrectly rejecting H0 when it's true. Commonly set at α = 0.05.

• Type II Error (β): Failing to reject H0 when it's false.

• Power = 1 - β: Probability of detecting an effect if one exists.

Confidence Intervals

• A range indicating where the true population parameter lies.

• Example: For RR = 1.5, CI = 1.2 - 1.8 indicates plausible values.

• As sample size increases, confidence interval width decreases, enhancing precision.

Random vs. Systematic Error

• Random Error: Chance-related errors in measurements.

• Systematic Error: Errors arising from identifiable biases or confounding effects.

Internal and External Validity

• Internal Validity: Reflects the accuracy in determining cause-effect relationships.

• External Validity (Generalizability): Ability to apply findings to broader populations.

Enhancing Study Precision

• Increase sample size to reduce random error.

• Conduct repeated measures to improve reliability.

Use of Humor in Epidemiology

• Illustrative examples in teaching public health concepts, including references to Type I errors and hypothesis rejection.