A bit field variable defined as TWIPS represents a floating point defined in TWIPS. Load the value as a signed or unsigned integer and then divide it by 20. The floating point result is a precise dimension in pixel.
Please, see the [un]signed type for more information about fields.