I know that given a differential equation, one that is separable, it is not fully correct to handle the $\frac{dy}{dx}$ as a ratio. Meaning that it is not simply a small difference in $y$ over a small difference in $x$, as the difference is infinitesimal. It is simply a notation created by Leibnitz that is equivalent to $f'(x)$.
My question is if this is the case then why is this method for ODE still used and taught? Is the answer always correct when dealing with ODE using this method or is does it lead to wrong results in some cases?
When a seemingly illegal mathematical manipulation has been assigned a logically consistent foundation and when the users are fully aware of what they are doing, then the manipulation is adopted for convenience. There is a handy example: the symbol"$dx$" in a definite integral $\int_{a}^{b}f(x)dx$. One can argue that writing $\int_{a}^{b}f$ suffices to preserve all the things relevant here; so why the "redundant" one is still in use? A reason is that it is mnemonic and hence convenient! It reminds people of the construction of (Riemann) integration. Once the foundation of limit is solid, mathematicians won't be confused by the mysterious "dx" anymore; instead, they use symbols wisely not pedantically. The seemingly abuse of symbol sometimes does in fact help, as long as one knows his stuff.
So don't be played by symbols; play the symbols.