Statistical recommendations for count, binary, and ordinal data in rare disease cross-over trials