Oncologists provide imprecise but unbiased estimates of survival time in elderly cancer patients

Oncologists often have imprecise estimates of expected survival time (EST), but these are generally unbiased and helpful for establishing best- and worst-case scenarios, a recent study has found.

The study included 102 elderly cancer patients (median age, 74 years) who were starting a new line of palliative chemotherapy. Upon enrolment, oncologists estimated the EST and rated the patients’ frailty status. An estimate was considered precise if it fell within 0.67–1.33 times the observed survival time (OST).

Fifty-eight deaths were reported after a median follow-up of 19 months, resulting in a death rate of 57 percent. The median OST was 15 months. Although the median EST estimate was 15.5 months, most were imprecise, with only 30 percent of patients receiving an EST within 0.67–1.33 times the actual OST.

Nine percent of the patients lived for at most a quarter of their EST, while 56 percent lived half to double their ESTs. The resulting discriminative value of the EST was moderate, with a c-statistic of 0.64.

Despite this, estimates were unbiased, showing no systematic tendency towards over- or under-estimation. In practical terms, 46 percent of the participants lived below their EST while 54 percent enjoyed longer OSTs.

Multivariable analysis showed that the EST was minimally but significantly correlated with OST, such that each additional month increase in the estimate reduced the risk of death by 4 percent (hazard ratio, 0.96, 95 percent CI, 0.93–0.99; p=0.03).