Allocation concealment in randomised controlled trials: improving the implementation, reporting and developing methods to detect and prevent selection bias