Page 18 - PCC13

• •
• •
Analysis
Pre-specified analysis plan
Analysis by intention to treat
Avoids bias due to non random exclusion Less likely to overestimate a positive result
Multiple analysis and subgroup testing
Hypothesis generating and pre specified
Only pre-specified interim analysis
Reporting trials
• Compliance with statistical analysis plan
• Consort statement
• Highlighting strengths and weaknesses
• What do the results mean?
• Rational interpretation of data
18
Page 12 of 84