1) People change a lot. My med school gave out the list of what our choices had been as incoming students to us on match day and I think less than a 1/3 were actually unchanged.
2) The things that drive a lot of change are varied but in my experience in dealing with a lot of students and residents heading into 3rd year were the following:
- Step 1 scores - more so in terms of weeding people out, a high score doesn't mean someone automatically starts thinking derm when they never considered it before, but that derm hopeful who scores a 225 suddenly has to make sense of their world.
- Course work - there's a lot in med school that many people have never been exposed to before, so sometimes that's very eye opening when a person finds out they love neuroanatomy or cardiac physiology. Similarly, they have done a ton of labwork/bench research in undergrad and find that they are most comfortable in the path lab and decide they'd rather stay in that realm because it's comfortable.
- Patient encounters - either standardized patients or with pre-clinical preceptors. Some people really struggle with that transition and determine early on that the less they see patients the better.
3) I agree with whoever said there are few "aha" moments for most people. It becomes a gradual sort of thing that reveals itself a lot of times.
4) Clinical rotations are important for sure, because while everyone has experience being a patient in the outpatient settings, inpatient hospital based medicine is a very different beast. And even fewer have had much experience in the OR.
5) I think the worst thing that a students can do (yet happens all the time) is to base their specialty choice on a great experience based on the people they were around in that field. So many students have a great attending or love their team while on a rotation and think that means it's their best choice to spend the rest of their career without realizing they really don't like the actual medicine they'll have to practice. The thing is though, the medicine doesn't change but the people do. You can find degrading peds residencies and neurosurgery programs that reaffirm your soul, but if you don't like the actual pathology walking through the door, you're going to be unhappy.
So what's a person to do?
For your specific research question, do something that sounds interesting. If its in your desired field, awesome, if not, who cares, better to have it than nothing.
After that, I think the algorithm for most people should be as follows:
1) Do I like the OR or not? If yes, then youre down to surgery specialties and anesthesia. If no, you can rule those fields out.
2) Kids or no kids and by that I mean, will not seeing a particular patient population leave a hole in your professional soul? Most people have one they prefer immensely and the other is at best a "nice to have, not a deal breaker if I never took care of those patients again".
2B) is do I want to only take care of women, if yes, then go directly to OB/GYN as all other questions are largely irrelevant
3) Do I really want to interact with patients that much? If no, then you need to start considering the Path/Rads options
4) Is there a specific organ system I love more than the others? Or would I rather be a generalist? For the surgical fields, this pushes you into your various subspecialties, keeping in mind you could still decide to do pediatric fellowships out of any of them, re: question #2. If you prefer to be a generalist that pushes you towards Gen Surg and to a lesser degree Plastics, and then you can further subspecialize later. For the non-surgical person, this question gets you into the subspecialties, but also includes things like PM&R, EM, Family Practice. Within Peds, the organ specific vs not pushes you towards Peds Critical Care, Neonatology, Adolescent Med and PEM pathways in addition to General Peds. There's less "generalist" fields in Medicine (although geriatrics definitely is) but still is an important question to answer.
5) If you're a non-surgical person and you're still undecided, then you can start to consider things like OP vs IP time. Very few fields are entirely one or the other so you'll mainly be comparing relative amounts of time.
6) Lastly, if you're still stuck between several options, go with what pays the most...