Prognostic Indices Offered for Decisions with Older Patients

/kidneynews/4_3/6/graphic/6f1.jpg

Many clinical guidelines—including a recent one on dialysis—recommend taking a patient’s life expectancy into account in selecting treatments, but accurate prognostic tools are hard to find and use, especially amid the time constraints of a busy practice.

A new website could make life expectancy judgments for older patients easier and more accurate by offering automated calculators that provide patient-specific statistical likelihoods after a few clicks of a mouse. The calculators are backed up by a recent review article in the Journal of the American Medical Association (1).

“Ignoring prognosis and life expectancy can lead to poor care,” said study coauthor Sei Lee, MD. Patients are often treated with therapies that they will not live long enough to benefit from. Those with life-threatening conditions are often referred to hospice too late to appreciate its benefits. And age-based recommendations may withhold appropriate treatment from those who are unusually hale and hearty for their chronologic age.

“Life expectancy is often not accounted for in medical decision-making, so we tried to make it easier for doctors and other health-care providers by collecting all of the life-expectancy calculators that we could find in a systematic review and putting them in one place so that people could just go to one place and find what they needed,” Lee told ASN Kidney News. He is an assistant professor of medicine in the geriatrics division at the University of California, San Francisco.

After a literature search, the researchers screened some 20,000 prognostic indices. They ruled out disease-specific indices, focusing on all-cause mortality in patients over 60 years old. They found 16 indices that passed the test of being developed in one cohort and validated in another with a level of accuracy deemed “moderate” to “very good.” The indices apters delineated in each report to create automated calculators and published them at a website, Eprognosis.org. The researchers urge caution in their use because none of the prognostic indices has been completely tested for routine use, but they propose that the indices provide some objective information beyond a physician’s intuition and experience.

None is specific to nephrology, but measures of kidney health are important contributors to some. For example, the Inouye burden-of-risk illness score for nonterminal hospitalized persons 65 years and older has an accuracy rating of “good.” It gives a patient who on admission has chronic renal failure, an albumin level of 3.5 g/dL or lower, a creatinine level above 1.5 mg/dL, and no other risk factors (such as cancer, stroke, congestive heart failure, diabetes with end-organ damage, or dementia) a 32 percent 1-year mortality risk. The addition of a single additional risk factor raises this risk to 61 percent.

“I think that it is a very important review,” said Mark A. Swidler, MD, a nephrologist and associate professor of medicine, geriatrics, and palliative medicine at Mt. Sinai School of Medicine in New York City, who was not involved in the review.

“It draws attention to the importance of prognostic indices because we have an aging population that is living longer with a greater amount of comorbid conditions and geriatric syndromes, some of whom are facing dialysis decisions or are on dialysis. It is important to have methods to quantify the contributions of those conditions and syndromes to the patients’ survival. However, we’re not only talking about survival. Geriatric decision-making is also about quality of life, which is most reflected in optimizing mental function and functional status. Eprognosis is useful because it provides calculators, so all you have to do is put in the appropriate numbers and then you get an answer,” Swidler said. A clinician could bring up the calculator on a smart phone while talking to a patient but would be unlikely to perform the calculations required otherwise.

Swidler agreed with the review authors, who noted that more work remains to be done to make prognostic indices more helpful for routine use. Also, although prognostic information is important in a patient’s decision to choose or forego dialysis therapy, these indices have not been validated in dialysis or other nephrology populations.

Prognosis is especially relevant to high-impact treatments such as dialysis and transplantation. The 2010 edition of the Renal Physicians Association guideline on initiating and withdrawing dialysis emphasizes the need to estimate prognosis and survival time. The chair of the panel that drew up the guideline, Alvin H. Moss, MD, told ASN Kidney News, “The physician should learn the patient’s values, wishes, and goals for care and make a treatment recommendation, also taking into account the patient’s prognosis and overall condition. It is a shared decision-making process, about what course of treatment the patient would want given the patient’s condition. The prognostic information is very helpful in that process.”

Although some indices have been developed that are more applicable to nephrology patients than are those at Eprognosis.org, most are not as accessible as calculators. However, Moss helped create an easy-to-use calculator for patients already undergoing dialysis, “The Surprise Question—Dialysis Mortality Predictor.”

Rethinking dialysis in the elderly?

The consideration of prognosis could lead to some rethinking about dialysis, especially because the fastest-growing age group to be starting dialysis is made up of those 75 and older. The average life expectancy of a 75-year-old starting dialysis is 1.5–2 years, so the wisdom of the treatment was called into question by a study showing that the start of dialysis is associated with a substantial and sustained decline in functional status in nursing home residents with ESRD, published in 2009 in the New England Journal of Medicine by Manjula Kurella Tamura, MD, and associates (2).

An assistant professor of medicine at Stanford University, Kurella Tamura has a prognosis-oriented article coming out in Kidney International that provides a framework for individualizing ESRD management decisions in older patients by incorporating life expectancy and patient preferences to assess the risks and benefits of competing treatment strategies (3). “We tried to look at decisions like vascular access placement or referrals for kidney transplant, because life expectancy has a substantial effect on the potential benefits of those interventions,” she told ASN Kidney News.

Most guidelines recommend an arteriovenous (AV) fistula rather than an AV graft or a catheter as the first access type in patients beginning hemodialysis, but the recommendation may not apply equally to all. AV fistulas have fewer complications like access-related bloodstream infections than do AV grafts or catheters, but they take longer to mature, so patients with limited life expectancies may not realize the benefits. Kurella Tamura and her team estimated that for the average 75-year-old patient, one would need to treat 25 patients with an AV fistula rather than an AV graft to prevent one episode of access-related infection. “That to us seems like quite a large number of patients. In contrast, you would only have to treat two patients with an AV graft vs. a catheter in order to prevent one bloodstream infection. That suggests that a fistula may not be the access of first choice for some patients,” she said.

The article says that perfectly accurate predictions of life expectancy are not needed: “Reasonable estimates of whether a patient is above or below the median life expectancy for his or her age will allow clinicians to make better assessments of the risks and benefits of various management strategies.”

The article also contains life expectancy estimates for dialysis patients of different ages broken into quartiles. For example, in the 75–79 age group, 25 percent of the patients can be expected to live 3.7 years, 50 percent to live 1.7 years, and 25 percent to live 6 months or less. Swidler said that the Eprognosis indices could be helpful in placing patients into these quartiles and talking meaningfully to them about how they want to optimize their quality of life and spend their remaining time.

In an editorial in JAMA that accompanied the prognostic indices review, Thomas M. Gill, MD, of Yale cautions, “Despite the proliferation of prognostic indices for mortality, there is currently no evidence that their routine use improves patient outcomes. To determine whether use of a previously validated prognostic index is better than usual care, an impact study must be conducted.”

The review article agrees that “further research is needed before general prognostic indices for elderly individuals can be recommended for routine use.” But Lee said that he would “absolutely encourage” clinicians to use the indices “with a grain of salt” to improve on the use of clinical experience alone.

Physicians too optimistic?

Studies have shown that physicians tend to be too optimistic in estimating life expectancy. “When you compare clinician intuition vs. an index vs. a combination of both, the combination always wins, and so I would argue that this piece of information is a valuable adjunct to clinical intuition and has been shown to lead to more accurate predictions,” Lee said.

A potentially controversial aspect of Eprognosis.org is that its presence on the Internet makes it accessible to the general public. Patients can access it simply by clicking the button saying that they are health professionals. The researchers left it accessible because anything that would have made it harder for the public to use would have made it less accessible to physicians. Lee acknowledged that even sophisticated patients may not understand the limitations of the indices.

Public accessibility can be seen as a part of the movement toward shared decision-making, observers said. “We’re moving toward an age where consumers are better informed,” said Moss, a nephrologist and medical ethicist at West Virginia University. “But drawing conclusions from Eprognosis.org is not something that patients should do independent of having a discussion with their doctor.”

“I think families and patients have to be involved and be given the choice of getting the information,” Swidler said. “Dialysis in certain subgroups of the elderly ESRD population is very challenging. You are signing up for a treatment program that is a big commitment. And up until now, I don’t think there has been enough available information for the public to really know what the reality is and make good decisions.”

Lee said that he has been using prognostic indices for years in his geriatrics and palliative care practice for discussions with patients: “It really opens the door. Some patients quickly let me know that they don’t want to talk about it, and I recommend specific care incorporating life expectancy into my recommendations, but I don’t ever explicitly talk about it. For other patients, they have been thinking about it, and it feels like flood gates are opening.”

References

1.Yourman LC, Lee SJ, Schonberg MA, et al. Prognostic indices for older adults: a systematic review. JAMA 2012; 307:182–192.

2.Kurella Tamura M, Covinsky KE, Chertnow GM, et al. Functional status of elderly adults before and after initiation of dialysis. N Engl J Med 2009; 361:1539–1547.

3.Kurella Tamura M, Tan JC, O’Hare AM. Optimizing renal replacement therapy in older adults: a framework for making individualized decisions. Kidney Int 2011, in press.

March 2012 (Vol. 4, Number 3)