Letters

New method of expressing survival in cancer is popular

EDITORWe are delighted that our method of calculating the "normal remaining life" to express survival in cancer aroused so much interest, and we would like to respond to some of the issues raised in letters.¹ Tan suggests that the life expectancy figures that we used represent the median survival for a population cohort.² This is untrue. We used cumulative yearly probability of death to calculate our figures of life expectancy, which is the age at which the probability of death reaches 100%. The median survival, on the other hand, is the age at which the probability of death reaches 50%. When we say that a woman aged 40 has a life expectancy of 75 we mean that the cumulative probability of her dying by age 75 is 100%, not 50% (median survival) as Tan assumes.

There were several suggestions that living one's full normal remaining life does not equate with cure, because cure is thought to mean complete biological elimination of disease. We thought that a patient would really be interested in a personal cure, meaning that she will not die of the disease before she has lived her full normal life span. To keep things simple, we used only mortality data in our paper; inclusion of data on relapse would give an estimate of the chance of living the full normal life span without the disease relapsing.

Finally, what has given us the most satisfaction is that our method appealed to a patient with breast cancer, Harrison. Patients like hera doctor herselfare the real reason for our paper, especially because she laments the fact that her prognosis was not calculated on the basis of our method when she developed cancer. Her feelings are being echoed in an ongoing study of patients' preferences about communication of the prognosis that we are conducting.