• Evidence Based Medicine
• Grades of evidence
• 1a
• Met analysis of RCTs
• 1b
• At least 1 RCT
• 2a
• At least 1 controlled non-randomised study
• 2b
• At least 1 quasi-experimantal study (eg cohort study)
• 3
• Well designed non-experimental study (eg case-control series, case series)
• 4
• Expert opinion
• Grades of recommendations
• A
• Based on level 1 evidence
• B
• Based on level 2 evidence or extrapolated from level 1 evidence
• C
• Based on level 3 evidence or extrapolated from level 1 or 2 evidence
• D
• Based on level 4 evidence or extrapolated from level 1, 2 or 3 evidence
• GPP
• Good practice point based on the view of the Guideline Development Group
• Criticisms of EBM
• Doesn't guarantee absence of bias
• Impossible to fully practice EBM
• 'Cookbook' medicine promoted
• Can't provide evidence for most clinical scenarios
• Observational study uses
• Ethical problems (e.g. Abortion-breast cancer hypothesis)
• Lack of control of investigators (e.g. smoking)
• Impracticalities (e.g. rare disease)
• Trial design
• Design Features
• Ethics approval
• Formulate null hypothesis
• Set controls and outcomes
• Define subject selection
• Calculate sample size (power)
• Trial
• Randomization
• Blinding
• Data collection
• Minimise bias
• Analysis
• Statistical analysis
• Assess clinical significance
• Phases of clinical trial
• Phase 0
• Animal studies/in vitro studies
• Phase 1
• Healthy volunteers: pharmacodynamic/kinetic information, small doses
• 20-50 people
• Phase 2
• Healthy volunteers: pharmacodynamic/dynamic information, different doses and frequencies
• 50-300
• Phase 3
• RCT, comparisons with existing treatments, assessments of S/E
• 250-1000+
• Phase 4
• Post-marketing surveillance
• 2000-10,000+
• CEACCP
  Link:ceaccp.oxfordjournals.org/content/6/4/148.full.pdf