The I2C bus that the IMU Module is connected to cannot tolerate much in the way of noise on its bus, especially with long wire runs, which is probably causing your pressure fluctuations.
The I2C bus was never intended to be used with long cable/wire runs. It was intended to be used for short runs between PWBs. I believe that the I2C spec calls for one twist every inch or so in the cable to reduce noise susceptibility. I was using the I2C bus for thruster control and had a thruster that was 18 inches from the controller. I had to reduce the I2C bus clock frequency from 100kHz to 75kHz to maintain consistent communication with the thruster.
As an example, I have a test bench setup of just the wire harness, motors, and a IMU Module. When I kept the I2C bus wires at the original wire harness length, I saw the fluctuations in the pressure reading that you are experiencing. So I cut the test bench I2C bus wire harness wires to the shorter length that I cut them to on my ROV chassis. After shortening the I2C bus wires, the readings became stable.