How you die -- and what it costs -- depends largely on where you get care. That's the revelation of a major national database widely regarded as the best hospital-by-hospital look at the cost of dying.
It shows that Bay Area residents are about twice as likely to die in a high-cost, high-tech intensive care unit as people in Minot, N.D., or Portland, Ore. But they are far less likely to get ICU care than residents of Manhattan.
Within the Bay Area, the Dartmouth Atlas of Health Care depicts hospitals' widely different approaches to care for dying people: Hayward's St. Rose Hospital ranks twice as high as Stanford in "intensity of care," a measure that considers the amount of time that dying patients spend in the hospital and the aggressiveness of physician services.
Major differences appear even within cities: In San Jose, Regional Medical Center and Good Samaritan Hospital are 15 miles apart -- but Medicare spent an extra $23,000, or a third more, per patient in their last two years of life at Regional than Good Samaritan, because of different care strategies.
America's varied landscape of end-of-life experiences reflects different philosophies in how aggressively hospitals combat death, concludes Dartmouth's 2011 analysis. While some hospitals marshal doctors and devices to postpone death even when death is clearly certain, others favor care options that let the end of life proceed in comfortable settings, or even at home.
"We think that science drives clinical decision-making," said Dartmouth investigator Dr. Elliott Fisher. "But those decisions are sharply conditioned by how many beds are in our community and how our physicians have become accustomed to treating seriously ill patients.
"In the topsy-turvy world of health care, doctors and hospitals have a very powerful influence on how you are treated," he said.
Geographic and institutional variations in care are the subject of this installment of the ongoing series on the Cost of Dying. To understand those patterns, there is no better source of data than the Dartmouth analysis, which reveals patterns of care that patients and families can use to weigh where physicians are likely to be more, or less, aggressive.
California law states that physicians can not be required to provide medical care that they believe will be ineffective. But textbooks offer no standard formula for end-of-life treatment, and it can be hard to predict who will benefit or be harmed.
If patients have not specified in writing their end-of-life wishes in advance, their fates rest on doctors' accumulated subjective decisions about treatments and tests to order or to resist.
"These differences pan out because the way we practice is different," said Dr. Steven Pantilat, director of UCSF's Palliative Care Program. "It's not the patients."
For more than two decades, the Dartmouth Atlas Project has documented glaring variations in how medical resources are distributed and used in the United States. The project analyzes Medicare's billing records to obtain information about national, regional and local markets, as well as hospitals and their affiliated physicians. It aims to boost understanding of our health care system and forms the foundation for many efforts to improve care across America.
The project has its critics, who say its death analyses ignore the people who are still alive thanks to aggressive doctors.
Higher-spending hospitals seem to save more lives, according to one study of heart failure patients at six California teaching hospitals.
That finding is missed in the Dartmouth work because it focuses only on dead patients, according to author Michael Ong of UCLA. "For physicians, patients and patients' families, survival is a critical concern," he wrote.
But its supporters say the data shine a harsh light on "do everything" impulses in hospitals, which can prolong suffering or dying.
One end-of-life analysis is based on billing records of Medicare patients in their final two years of life between 2003 and 2007, the most recent years for which information was available. It assesses each hospital on an "intensity index," based on hospitalization of patients before their deaths.
Another analysis focuses more narrowly on care for people who died of cancer. It asks: In their final month of life, did they stay in the ICU, or have hospice care? Were they given chemotherapy or life-sustainers, such as ventilators and feeding tubes?
Among the findings revealed by the Atlas' data:
To ensure fair comparisons among regions, rates were adjusted to remove differences because of age, sex and racial compositions.
Intensity of care
Health care for elderly Americans at the end of life has been changing.
Dying patients are spending fewer days in the hospital and more time in hospice, investigators said.
But for patients who are hospitalized, the data show an increasing intensity of care -- with more doctors' visits and greater reliance on the high-cost ICU.
This holds true even for Bay Area hospitals like Stanford and UCSF, with strong palliative care programs that offer on-site relief from the pain and stress of a serious illness.
When 76-year-old Sarah Kyung Lee, of Foster City, fell and hit her head while on a walk in March, she was taken to Stanford, where she received IV lines, a urinary catheter and daily blood draws. Because she is a frail woman, with lifelong heart problems, the hospital sought to monitor her closely.
But her son, geriatrician Dr. Sei Lee of UCSF, insisted the hospital release her. So she was sent home to her board-and-care facility.
"It was exhausting, continually trying to prevent care that was too aggressive," he said.
She recovered in her own quiet bedroom, with a Bible, family photos and homemade Korean food.
Stanford Chief Medical Officer Dr. Norman Rizk responded that some families would be uncomfortable with the early discharge of a relative. "We acknowledge there is a wide variety of sentiments about the levels of care appropriate for elderly patients," he said. "Care cannot be generic; it must be about personal preferences."
What's happening is that "end of life is becoming more health care intensive," said Judy Citko, director of the Coalition for Compassionate Care in California. "It's a natural progression of two trends: People are living longer with more chronic conditions, and there are more things that the health care system can do."
Yet with all our new tools, it's harder to know what helps, and what doesn't, she said.
"For example, there are a lot more chemotherapy options," she said. "But there's not a clear demarcation when the transition occurs from chemo that has a good chance of being helpful to chemo that is a complete long shot."
Culture and resources
The Medicare data don't examine how hospitals help -- or don't help -- patients make these choices. And it doesn't reveal reasons for the variations in medical care at the end of life.
But research by Dartmouth and others suggests that treatment trends result from different medical cultures and resources.
In medicine, supply drives demand. So regions with a lot of doctors tend to have more hospitals, offer more medical services and charge more.
"The most expensive piece of medical equipment is the doctor's pen," which signs orders for hospitalizations, tests, treatments and referrals to pricey specialty doctors, writes Atul Gawande, author, public health researcher and surgeon at Brigham and Women's Hospital in Boston.
When a city or hospital has a lot of empty beds, local physicians unconsciously adapt to this higher capacity and admit more patients, the Dartmouth team said. Research at Dartmouth and elsewhere shows that when ICU beds are readily available, less severely ill patients fill them and stay longer.
Imagine that a patient's chronic condition worsens; if the nearby hospital has a lot of room, a doctor may reason it is safer and easier to treat them there.
But if the hospital doesn't have available beds, doctors may look harder for stay-at-home alternatives, said UCSF's Pantilat.
This explains the differences in care between San Francisco and Los Angeles. "In the San Francisco Bay Area, where hospitals are fuller, we see less intensive care than in L.A., where beds are readily available," said Maribeth Shannon of the California HealthCare Foundation in Sacramento.
Medical school curriculums also drive regional differences: Medical residents often practice where they studied, reinforcing the region's culture of care.
At UCSF, medical students hear lectures in pain management and palliative care and participate in discussions about dying.
They even get lessons, using paid actors, in how to break bad news. Studies show that doctors who don't feel comfortable discussing end-of-life options are more likely to propose more treatment, no matter how futile.
The hospitals that take the time for vital conversations are the ones that keep dying patients out of ICUs, off ventilators and free of desperate chemotherapy, Pantilat said.
They give patients impartial and accurate information about end-of-life choices. They describe options of palliative care or hospice. They mention the complications of treatment. They probe families' goals, such as pain control, dignity and quality of life.
Those are some of the steps being taken by Daly City's Seton Medical Center, which in 2007 ranked high among Bay Area hospitals for its frequency of treating end-of-life patients in its ICU.
"We began to realize that we were protracting death, not ensuring life, and that wasn't right," said Seton Chief Medical Officer Dr. Irwin Shelub. The hospital hired a palliative care coordinator and chief medical officer. It insists on frequent family conferences. It has even put into place a "Medically Ineffective Care" policy, which forestalls unnecessary or ineffective care.
Under the policy, when Seton's medical team agrees that a patient would be better served by "comfort care," it meets with family members to say so. If the family disagrees, the hospital has the authority to discharge the patient and helps families find another facility.
"We've dramatically changed our direction," Shelub said. "We don't want to engage in futile care."
Even more ambitious city and statewide efforts are happening outside the Bay Area.
A model of "low intensity" end-of-life care, Gundersen Lutheran Health System of La Crosse, Wis., encouraged every city resident to complete an advance directive, written instructions specifying patients' end-of-life care preferences.
The result is that doctors there can tailor their treatments to each patients' desires. It also has hired social workers to talk with patients and families about end-of-life options and provide other support.
Oregon, which also ranks low in intensity of care, has a tradition of debating end-of-life issues. So residents know they have choices, said the California HealthCare Foundation's Shannon. The state recently pioneered an electronic registry for more than 120,000 POLST (Physician Orders for Life-Sustaining Treatment) forms for the aged or very ill, in which patients set highly specific limits on the kinds of care they desire.
"The best places have a system in place -- so that patients can make their preferences clear, and physicians feel trained and supported," said Dartmouth's Fisher.
"That's how people get the right kind of care, when it matters most."
Contact Lisa M. Krieger at 650-492-4098.