While I don’t think it is the greatest of ideas and considering that it is a flow it may not be possible.
However, if you change your range from say 0-100 " H2O to -100 -0 " H2O you should be able to achieve the same mA out put to your control system therefore getting a correct reading. Essentially no different than using a wet leg on you LP side of a level.
Also I do not see why you would damage your diaphragm seeing how both sensors would be sitting in the same line pressure essentially. Then again I am not certain this would work just spit balling an idea.
The easiest fix would just be to fix the mistake and swap your impulse lines rather than doing a bunch of paper work and working on my theory.