As if we needed further evidence that medicine is growing increasingly impersonal, the Canadian Medical Association Journal has published a study that claims a computer program can predict when seniors have six months to live. From the Global News story:
Amid a lack of proper support for Canadians receiving home-based support towards the end of their lives, a new risk calculator is helping predict how long seniors have left to live.
The Risk Evaluation for Support: Predictions for Elder-Life in the Community Tool — dubbed ‘RESPECT’ for short — can predict death within six months, and was developed using data from more than 491,000 community-dwelling adults aged at least 50 years who used home care between 2007 and 2013.
Always with the acronyms to hide utilitarian protocols and procedures. Euthanasia in Canada is called MAID (medical assistance in dying), and now, RESPECT. Good grief.
“The RESPECT calculator allows families and their loved ones to plan,” said Dr. Amy Hsu, investigator at the Bruyère Research Institute and lead author of the study.
“For example, it can help an adult [or] child plan when to take a leave of absence from work to be with a parent or decide when to take the last family vacation together.”
Or it could be used to restrict care and/or push euthanasia. As one Canadian bioethicist noted:
If the calculator would ever be introduced to Canada’s healthcare system, Bowman believes that it would be interfaced with the country’s medical assistance in dying (MAiD), and could possibly shape the attitude of palliative care and end of life decisions.
“It will also shape the attitude of health care workers and it also raises a deeper question of who will interface with the broader question of what types of life are worth living and who decides, which is profoundly important stuff,” he said.
People don’t die by the numbers. Much depends on the kind of care they receive, their mental states, and individual differences that can be immeasurable. Even the study’s authors note a very big problem.
As with many prediction models, RESPECT is less well-calibrated at the extremes of the distribution. In particular, we found that RESPECT overpredicted the mortality risk of patients in our top 3 risk bins.
The idea that crucial and intimate decisions about patient care could soon be driven by a computer-modeling system — rather than individual assessments — is very alarming. And it will often be wrong. I know of several patients given six months or less to live who got kicked out of hospice because their health improved unexpectedly. This includes the humorist Art Buchwald, who left hospice when he didn’t die from kidney failure and lived long enough to write his last book.
But then, with the quality-of-life ethic taking hold in medicine throughout the West, a “follow the science!” approach would make it much easier for clinicians, socialized-medicine bean counters, and family to abandon frail patients to comfort-care-only regimens — or worse — and still get a good night’s sleep.