Health, Science, Environment
Can knowing when you’ll die help you live?
A note to our listeners and readers: We’ve changed the name of the patient in this story for privacy reasons.
It’s a busy Monday morning at the San Francisco Veterans Association Medical Center. Palliative care physicians Alex Smith and Katherine Aragon, along with their intern Sara Murray, gather in the lobby. They’re on their way to see a patient, Bill Jones – a gentleman in his 60s who has a form of slow-growing cancer.
On the way to Jones’ room, the physicians pause by the elevators to discuss Jones’ case, and estimate his prognosis. To do so, they not only look at Jones’ charts – they pull out their iPhones.
Doctors usually rely on their experience when offering a patient their prognosis, especially patients who are suffering from multiple illnesses. But these doctors are using an additional tool called ePrognosis, a website developed by Smith and a team of researchers (they’re currently testing an app of the site). The tool uses existing indices to help physicians with the complicated task of estimating life expectancy.
This morning, ePrognosis is prompting Smith with a series of questions like: Does he have shortness of breath? Does he have heart failure? Does he leave more than 25% of his food uneaten?
All in all, it takes Smith and Aragon about five minutes to complete the questionnaire on their phones. At the end, ePrognosis spits out an answer: Jones has between six months and a year to live – an answer that confirms Smith’s clinical judgment. Physicians don't always determine an accurate prognosis, however.
“Doctors take their training and their professionalism very seriously … they really don’t like being wrong. And whenever you make predictions about the future, you’re almost guaranteed to be wrong sometimes,” explains Dr. Sei Lee, a geriatrician and assistant professor at UCSF and SFVA. Lee is also one of the authors of ePrognosis. He says that when it comes to predicting death, physicians aren’t just sometimes wrong – they’re frequently wrong.
“Doctors, when they use their clinical intuition to estimate life expectancy for a patient, have a tendency to be overly optimistic,” Lee says. On average, physicians estimate that a patient will live twice as long as they usually do. And that miscalculation can lead to bad medical decisions. Physicians might approve risky tests for frail patients who likely won’t live to see the benefits of those tests. And it can lead to worse quality of life at the end of life, Lee says.
“One of the things that we’re struggling with in this country is that patients are not entering hospice until very late,” says Lee, “and on average, that time can be as short as a week or two before death, which is really not enough time to maximally relieve symptoms and really support the family.”
ePrognosis can’t predict a patient’s death with 100% accuracy – nothing can – but combined with a physician’s intuition, the tool can help inform end-of-life conversations with patients – if they want to have them.
“It’s certainly not an easy topic to think about,” says Lee. “But in my work as a doctor – I feel like so much of the heartache and hardship that comes at the time of death are because things are were not well-prepared beforehand, where things that needed to be said weren’t said. Those things could have been avoided if there had been a more frank discussion about life expectancy and prognosis up front.”
Back at the SFVA nursing home, Dr. Aragon is attending to patient Jones while Smith and Murray look on. Jones looks grizzled and grayed, and his speech is a little garbled from the cancers in this throat and head. But overall, he’s clear-minded and upbeat. He says that his symptoms have been changing a lot – that there have been good days and bad days.
When Aragon asks him whether he’s ever wondered how much time he has left, Jones says that it could be weeks, months, or even tonight, but he doesn’t seem to be afraid of dying. Jones has put his money in order and has arranged to leave an inheritance to his mother. He’s also been speaking with ministers and priests. “I don’t think I’m going to hell if you believe in that sort of stuff,” Jones says. “The pain is the thing I fear the most, I think.”
Overall, the conversation with Jones goes well, but oftentimes, end-of-life conversations are difficult for both the patient and the physician. “It’s hard because the goal is you want to tell someone that they’re going to get better,” Aragon admits.
Furthermore, as Lee explains, doctors have not been well-trained in having these types of discussions in the past. Still, despite historic trends, Lee says that some medical schools are starting to prepare young doctors to conduct end-of-life conversations. And he’s hoping that tools like ePrognosis can help. The site’s been active for two months, and received about half a million page views in its first week. Visits to ePrognosis have tapered off from that initial spike but remain steady – a sign that a core group of physicians could be utilizing the tool in their practice.
Meanwhile at the SFVA, Bill Jones is planning for the last few months of his life. He isn’t accepting any more chemo – he says he doesn’t know what kind of a “vegetable” he’d be if he continued treatment. Instead, Jones is spending the rest of his days in hospice care at the SFVA Community Living Center. And that might make his last few months a little more livable.