People need to seriously just go back to DSM.
If this person meets criteria for Bipolar I, then she has bipolar I... etc. go prescribe meds from there.
If this person meets criteria for ADHD, write ADHD meds.
If someone meets criteria for BPD, refer for DBT/TFP/Schema.
If someone meets criteria for a substance use disorder, do MI/CBT/network therapy + evidence based meds (antabuse, naltrexone, maybe high dose ADHD meds, opioid meds, etc. if you don't know what you are doing, consult a specialist)
Seriously. If you haven't memorized the criteria just pull out that god damn handbook and go down the check list. It's designed to be fool proof. There's a reason all clinical trials use it.
You get done with that. Then *maybe* you go off label for symptomatic control using mood stabilizers for people who are "bipolar NOS". Maybe you add a cocktail med (lamictal, yuck) or two briefly to see what happens. MAYBE you start playing the old witch and make a borderline caldron with multiple antidepressants. Maybe you start to get fancy and think about neurotransmitters and check the latest on RDoC. MAYBE.
But really, you err on the side of pharmacological minimalism. Everyone here knows that if someone meets criteria for borderline personality disorder, invariably if they just have a commitment to sit in 3 times a week formal DBT for a year it's better than any med cocktail. And this is by the way is 100% evidence based.
What I find problem in the community with "bipolar" is that people are too lazy to use the DSM and justify their laziness by reporting that they are somehow "better".
No you are not better than evidence based medicine. Just stop. I would seriously prefer a robot over you. By the way, just to throw one more wrench: in my experience, NPs are way worse in this regard than MDs.