I think you are referring to the value reported in the uhi_results.shp output. This is the average air temperature within the polygon (from T_air.tif), minus the “reference air temperature” value.
I think this question of getting results in terms of actual temperature change may be addressed in this thread. It sounds like the CC index is not appropriate for this, but a scenario-based approach using the T_air outputs might be.
This is another great thread on what the CC index can tell you: