EBRO 2017

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