Abstract
The so-called mark variable which combines both quantity and quality of a patient's life is a tempting measure in evaluation of therapies. The survival function of such mark variable, referred to as mark survival function in this note, is a primary interest in many applications. Much effort has been made on the estimation for this survival function. Nevertheless, we show in this note that mark survival function is unidentifiable under conventional assumption. Illustrative examples are provided to confirm the non-identifiability of mark survival function.