274 / 1983
Prefer individual data
• Improved quality of data
•
check random procedure
•
update follow-up
• Improved analysis
•
uniform endpoint definition
•
intent to treat
•
assess heterogeneity between trials
• Contact with investigators