iStockphoto/Getty Images/iStockphoto
THE QUESTION
My aunt was diagnosed with advanced ovarian cancer. Her doctor said she had about six months to live. But she died just over one month later. Are doctors usually this bad at predicting how long patients will survive?
THE ANSWER
Unfortunately, doctors get it wrong more often than they get it right, and most of the time, they overestimate how long patients will live.
Many physicians will shy away from making specific predictions. But patients and their families often want some idea about how long they might have left. And that desire to have a window on the future is certainly understandable.
However, the difficulty of making an accurate prediction was revealed in a study led by Dr. Debbie Selby, a palliative-care physician at Sunnybrook Health Sciences Centre in Toronto.
The researchers reviewed data collected on 1,622 patients diagnosed with incurable diseases. Palliative-care doctors assessed each of the patients. During the assessments, the doctors recorded details about the patients and gave an estimate of survival time. In particular, the doctors said if they thought a patient would live less than a day, one to seven days, one to four weeks, one to three months, three to six months, six to 12 months or more than a year. The estimates were not given to the patients, so the doctors weren't hampered by how their predictions might be received.
An analysis of those estimates showed that the doctors' predictions often missed the mark.
"We are not good at being very accurate," Selby says. "The vast majority of patients have less time than we predict."
The doctors were right only half the time in cases involving patients who had less than a month to live and also for those who survived longer than a year, according to the study published in the Journal of Pain and Symptom Management.
Predictions of between one month and one year were even less accurate. The doctors were correct only one-third of the time. "And we tended to systemically overestimate the time left."
Selby notes that the majority of terminally ill patients fit into the middle group – with a survival time of a month to a year – the period that seems the most difficult to predict accurately.
So, why are the predictions unreliable? For starters, many of these patients are going through a "transition" in their medical condition, and it is not easy to determine if they are in a steady or accelerating rate of decline. An unexpected turn of events – such as a bout of pneumonia – may suddenly speed up the dying process.
There are numerous measures – such as medical tests, physical exams and the patient's history – that can also be used to produce a statistical likelihood of surviving a specific length of time. Yet even these calculations "are not any more accurate than the physicians' predictions of survival," she says.
Selby is convinced that not much can be done to significantly improve the estimates. As a result, she believes it's important for health-care providers to let patients know that the estimates are simply best guesses. "We can't tell the future," she says, "but we can use our best guess to help people think about what may be important to them in whatever time they have left."
Without those discussions, there is a risk that some patients may continue to focus on their ongoing medical treatment, neglecting everything else. Even when patients respond positively to a treatment – such as getting a reduction in the size of a cancerous tumour – it may not actually help them live substantially longer.
"I don't have any qualms if somebody is getting chemotherapy until three days before dying, so long as that was right for that person," Selby says. "But did that person have the opportunity to explore other options and decide what was right for them? I am not sure everyone does."
For instance, some patients may want to take a trip to visit a special place, or see a loved one for a final time. Still others may want to be present at a certain event – such as the marriage of a family member. Sometimes, weddings can be moved forward, she notes. But other milestones – such as the birth of a grandchild – are simply not going to happen in time. In those cases, some patients may want to write cards that can be read by family and loved ones at key points in the future.
"I don't think that we, in the medical community, do enough to ask patients what matters most to them," Selby says. "We should help patients use their remaining time in the way that is right for them – recognizing that it is probably shorter than we think."
Paul Taylor is a patient navigation adviser at Sunnybrook Health Sciences Centre. He is a former health editor of The Globe and Mail. You can find him on Twitter @epaultaylor and online at Sunnybrook's Your Health Matters.