Some statistical methods for multiple endpoints in clinical trials