Survival-based comparison of hypothyroidism risk for nine bipolar therapies: approaches to address biases and competing risks in large scale observational studies