12.3 Biostatistics, Evidence-Based Medicine & Ethics

12.3 Biostatistics, Evidence-Based Medicine & Ethics

Diagnostic Test Performance & 2x2 Contingency Tables

Evaluating diagnostic tests requires mastering 2x2 contingency tables and understanding how prevalence influences predictive values.

The 2x2 Diagnostic Matrix

Core Diagnostic Equations

1. Sensitivity (True Positive Rate): The proportion of individuals with the disease who test positive. $\text{Sensitivity} = \frac{TP}{TP + FN}$
   • Clinical Utility: High sensitivity tests have very few false negatives (SnNout: a Sensitive test when Negative rules OUT the disease). Used for initial screening.
2. Specificity (True Negative Rate): The proportion of individuals without the disease who test negative. $\text{Specificity} = \frac{TN}{TN + FP}$
   • Clinical Utility: High specificity tests have very few false positives (SpPinn: a Specific test when Positive rules IN the disease). Used for diagnostic confirmation.
3. Positive Predictive Value (PPV): The probability that a patient with a positive test actually has the disease. $\text{PPV} = \frac{TP}{TP + FP}$
4. Negative Predictive Value (NPV): The probability that a patient with a negative test is truly free of the disease. $\text{NPV} = \frac{TN}{TN + FN}$

Critical Exam Concept: Impact of Disease Prevalence

• Sensitivity and Specificity are intrinsic characteristics of a diagnostic test and do not change with disease prevalence.
• PPV increases as disease prevalence increases in the screened population.
• NPV decreases as disease prevalence increases in the screened population.

1. Likelihood Ratios (LR): Combine sensitivity and specificity into a single metric independent of prevalence.
   • Positive Likelihood Ratio ($LR+$): $\frac{\text{Sensitivity}}{1 - \text{Specificity}} = \frac{\text{True Positive Rate}}{\text{False Positive Rate}}$. An $LR+ > 10$ strongly confirms the presence of disease.
   • Negative Likelihood Ratio ($LR-$): $\frac{1 - \text{Sensitivity}}{\text{Specificity}} = \frac{\text{False Negative Rate}}{\text{True Negative Rate}}$. An $LR- < 0.1$ strongly rules out disease.

---

Clinical Study Designs, Measures of Association & Trial Analytics

Epidemiology categorizes study designs by intervention status, temporal sequence, and sampling method.

Observational & Interventional Study Designs

1. Cross-Sectional Study: Measures exposure and disease status simultaneously at a single point in time.
   • Primary Measure: Prevalence. Cannot determine temporal sequence or causality.
2. Case-Control Study: Retrospective study comparing individuals with the disease ("cases") to individuals without the disease ("controls") to assess prior exposure history.
   • Primary Measure: Odds Ratio (OR) ($OR = \frac{a \times d}{b \times c}$).
   • Biases: Highly susceptible to Recall Bias (cases remember past exposures differently than controls) and Selection Bias.
3. Cohort Study: Prospective or retrospective study tracking exposed and unexposed groups over time to observe disease incidence.
   • Primary Measure: Relative Risk (RR) ($RR = \frac{\text{Incidence in Exposed}}{\text{Incidence in Unexposed}} = \frac{a / [a+b]}{c / [c+d]}$) and Attributable Risk (AR).
   • Advantage: Establishes temporal sequence; directly measures disease incidence.
4. Randomized Controlled Trial (RCT): Gold standard for evaluating therapeutic efficacy. Participants are randomly allocated to treatment vs control groups.

Therapeutic Trial Metrics & Calculations

Consider a clinical trial comparing a novel oral anticoagulant against standard treatment for stroke prevention in atrial fibrillation:

• Control Event Rate (CER): $\frac{\text{Events in Control Group}}{\text{Total Control Patients}}$
• Experimental Event Rate (EER): $\frac{\text{Events in Treatment Group}}{ ext{Total Treatment Patients}}$
• Absolute Risk Reduction (ARR): $ARR = CER - EER$
• Relative Risk Reduction (RRR): $RRR = \frac{CER - EER}{CER} = \frac{ARR}{CER} = 1 - RR$
• Number Needed to Treat (NNT): The number of patients who must receive the intervention to prevent one additional adverse outcome. $NNT = \frac{1}{ARR} = \frac{1}{CER - EER}$ (Always round NNT UP to the next whole integer to avoid overestimating clinical benefit).
• Number Needed to Harm (NNH): $NNH = \frac{1}{\text{Absolute Risk Increase (ARI)}} = \frac{1}{EER_{harm} - CER_{harm}}$

Analytical Strategies: ITT vs Per-Protocol

• Intention-to-Treat (ITT) Analysis: Analyzes every patient in the group to which they were originally randomized, regardless of non-compliance, protocol deviations, or withdrawal.
  • Purpose: Preserves baseline randomization balance, reflects real-world clinical effectiveness, and prevents Attrition Bias.
• Per-Protocol Analysis: Analyzes only patients who fully completed the trial protocol as assigned. Evaluates maximal potential efficacy but risks selection bias.

---

Biomedical Ethics, Informed Consent & Medical Law

Clinical ethics questions on the MRCPI assess decision-making frameworks under UK/Irish law and international ethical guidelines.

The Four Core Principles of Biomedical Ethics (Beauchamp & Childress)

Decision-Making Capacity & Informed Consent

Decision-Making Capacity is time-specific and decision-specific. A patient is presumed competent unless proven otherwise. To possess capacity for a specific medical decision, a patient must be able to:

1. Understand the information relevant to the decision (nature, purpose, risks, benefits, alternatives).
2. Retain that information long enough to make the choice.
3. Weigh or use the information as part of the decision-making process.
4. Communicate their decision (by speech, sign language, or clear gestures).

If an adult patient lacks capacity:

• Inspect for valid Advance Healthcare Directives (Living Wills) or designated Durable Power of Attorney / Healthcare Proxy.
• If no advance directive exists, treatment decisions must be made in the patient's Best Interests, involving family/carers, multi-disciplinary teams, and independent advocates.

Confidentiality & Legal Exceptions to Disclosure

Confidentiality is a fundamental duty, but disclosure without patient consent is legally and ethically justified under specific conditions:

1. Express Consent: Patient explicitly consents to sharing information.
2. Statutory Requirement: Mandated by law (e.g., reporting specified communicable infectious diseases like tuberculosis or measles to public health authorities).
3. Public Interest / Harm Prevention: To prevent serious, imminent harm to third parties or the public (e.g., reporting a patient with uncontrolled epilepsy who continues driving against medical advice, or warning identifiable targets under duty to protect frameworks).