Predicting the time to death – when a doctor has a 50% chance of being wrong

Predicting the time to death – when a doctor has a 50% chance of being wrong

According to the assisted dying bill, two doctors will have to certify the patient has less than 6 months to live. But, has anyone actully questioned what this means?

Remember the Al Megrahi case and his release: at the time he was given just 3 months to live. At the time he had prostate cancer, and doctors said the cancer had spread so extensively that he could not live for much longer.

Yet, he lived exactly 2 years and 9 months from the point of his release. The oncologist who made the prognosis said ultimately he was embarrassed by his decision, and the error occured  because he had given the average life expectancy.

And there you have the problem: life expectancy is estimated as an average – using either the mean or the median. The latter can give wildly inaccurate estimates.

A prospective cohort study in 4 multi-professional palliative care teams showed the accuracy of doctors estimates was about 70% right only when the minimum estimate of life expectancy amongst patients was less than 14 days. The conclusion was straightforward,  offering a prognosis estimate for death is very often inaccurate, except when very close to death.

In 1018 patients with advanced cancer survival estimates given by  Multidsciplinary teams, doctors’ and nurses’ were 57.5%, 56.3% and 55.5% accurate when compared with actual survival.

Therefore when a doctor certify someone has 6 months to live – he has at least a  50% chance of being wrong.

No Comments

Post a Comment