[Hospital statistics as tool in epidemiologic studies: heart failure in Trieste].