Text this: Can routinely collected data be used to reliably follow-up participants in large randomised trials?