Hi Anonymous,
it seems your problem is not the math itself, but the math under pressure. This is a very common problem for many candidates.
The majority of the people are actually fine with math when there is no pressure.
However they “suddenly” make a lot of mistakes in a case: missing zeros, simple math errors in divisions, wrong percentages, etc.
Why is that the case?
I have coached people in thousands of case interviews and I found there are two main reasons why candidates make mistakes under pressure:

1) Practicing math without a time constraint.
You may have a PhD in Math, but unless you practice math with a tough time constraint during the prep for consulting interviews, you will likely do mistakes.
Why? Because in a consulting interview you will face something you haven’t practiced before: pressure.
You know you are judged for each mistake you do and that you have to find a solution quickly.
This makes easy to do mistakes.
Unless you have trained for that in advance.
What is then a good solution for that?
Practice math with a time constraint that is almost unfeasible for you to satisfy initially.
Eg: if the average time you spend for percentage exercises is – say – 15 seconds, give yourself slightly less – say 12 seconds – for the following exercise.
Not only you will learn how to speed up your math, but forcing yourself to solve something in an “almost impossible” time will create pressure and will help you to get accustomed to it.

2) Not presenting a structure before doing the math.
Many people start their math exercise jumping to the numbers.
This is not the best approach to follow.
You should treat math in the same way you deal with the structure at the beginning of the case.
You don’t start a case going in details in one area, but presenting the overall approach you have, and THEN moving to the details.
In the math part you should do the same thing: present first your theoretical approach (the formula you are going to use), and THEN move to the numbers.
You may wonder how this may help to avoid math mistakes  in the end, you still have to go through the math anyway.
The reason is simple: your math mistakes in a case are usually not related to lack of math knowledge, but to the pressure you feel.
Presenting first your theoretical approach and aligning with the interviewer will not only help to understand if your approach is correct, but also to release pressure.
Indeed, once presented your approach, you know before the actual math if your logic is correct or not.
This usually makes also your math easier due to the fact you know you are following the right path.

If you exercise on the previous two points I am pretty sure you can improve quite a lot your math skills before your interview.
Hope this helps,
Francesco