Schema for Multicentre Trials
Use a study ID consisting of 3 components in the format XXXXXX.XXXX:
- 2 digit study number
- 4 digit site code
- 4 digit randomisation number
Study IDs are stored as character values to ensure that zero place holders are not dropped. The randomisation number is assigned sequentially at each site after randomisation. A decimal point is placed between the site code and randomisation number. This allows sites to monitor their recruitment and ensures the ID can be easily converted to a numerical value if manipulations are required. The study number ensures that samples can be identified across studies.
The following site codes have been used in previous Australasian multicentre trials and have been retained for continuity:
Auckland City Hospital
Hawke’s Bay Hospital
North Shore Hospital
Tauranga Hospital (Bay of Plenty)
Randomisation in multiple pregnancy
To deal with unit of analysis issues, there are two common solutions:
- Add a suffix where 0 is mother; infants are coded 1, 2 and 3 according to birth order in multiple pregnancy or 5 for singleton. A surviving twin would be coded as 1.
- Assign a separate "randomisation number" to the fetuses, e.g., start mothers at 1000 and babies at 5000.