Loss in heat instead of heat flux at boundary?

Is there a way to simulate a defined loss in heat at a boundary? So instead of having W/(m^2K) you'd have J/(m^2K).

Right now heat flux is the only time dependant variable that prevents my model from being stationary.