If all we taught were the most current, most accurate, most complete versions of scientific theories, then it would be very difficult to teach anything below graduate level. If the lower, simplified levels of understanding didn't exist, then it would be all but impossible to get into science at all.
We know that Newtonian mechanics is fundamentally inaccurate. Nevertheless, it’s a useful and necessary component of any physics curriculum because it’s a valid approximation in the vast majority of cases, and because it provides a foundation for more advanced studies. The scientists developing these theories didn’t pull them out of the air; they built on the work of previous generations. Moreover, it’s often true that the easiest way to understand a scientific concept is to follow the progression of logic that led to its discovery. We study the earlier model, point out the holes and inconsistencies in it, and then look at how the subsequent models get around those issues.
The Bohr model was an improvement on the understandings that came before it. It did, of course, have some flaws that could not be solved within a classical understanding, (such as the fact that an orbiting electron would radiate away its energy and spiral into the nucleus within a fraction of a second) but it also made possible some approximations that are still useful today. Quantum mechanics is more accurate, but it sacrifices a great deal of simplicity to do so. While it is theoretically possible to solve the wave function for larger atoms with many electrons, or even molecules, it’s unlikely to happen outside supercomputers. For most practical applications, a simpler model or an approximation is the best way to get a useful result.