*rolls up revision sleeves*
*digs out preclinical development lectures*
I'm only in my second year so take this with a pinch of salt, but from what I recall (and I hope I remember better in the exam!), animal testing tends to have a role in looking at the pharmacokinetics of a drug, and especially how it's metabolised in vivo. It's all very well if it works in a test tube in a lab, but you need to know whether it'll work on a living thing. Sometimes animals are used for this, but since the animals used in labs have to be specially bred for the purpose, they're more expensive than hiring the researchers themselves (if I heard my lecturer correctly), as LiquidCherry said, where it's possible not to use animals, they don't. If I remember rightly I think safety and efficacy tests are performed on both animals/cell cultures and then later in healthy human volunteers, before moving onto tests with patients who actually have the condition. Animals are also sometimes used to look at the function of different genes, to help identify possible targets for drug development. Like if you knockout the leptin gene from a mouse it doesn't know when to stop eating.
Also tests on the effects of drugs on pregnancy and lactation are done solely on animals, not humans, if they're done, you can't test on pregnant women. So in that sense it's used to check whether a drug would be safe in pregnancy, and that'll appear in the literature allowing healthcare professionals to make a judgement on whether or not they believe the benefit of taking a particular medication outweighs the risk. Obviously there are some we know cause definite harm to the foetus, thalidomide being the most famous example, and a few medicines that are so old that they've been used in pregnancy before and haven't given any ill effects.
I hasten to add that this is all rather sketchy knowledge and it's mostly me trying to regurgitate lecture notes as a vain attempt at revision.