Published in the April 2011 issue of Today’s Hospitalist
WHY IS IT SO HARD FOR PHYSICIANS to get DNR discussions right? A new study that asked that question found that despite all the attention being paid to end-of-life care and patient preferences, conversations about DNRs often fall into the category of too little, too late.
Studies have shown, for example, that DNR discussions often occur when disease is so advanced “within two to three days of death “that patients are no longer able to make decisions for themselves. Other research has found that DNR discussions often lack adequate information about patient goals and tend to downplay what can happen after “successful” resuscitations.
“This is a big problem,” says researcher Jacqueline Yuen, MD, lead author of an article posted online in February by the Journal of General Internal Medicine on what’s wrong with DNR orders. “We found that doctors often don’t have conversations with patients about their wishes before they get sick “and that when they do have those discussions, they do them so poorly.”
The team of researchers from Cornell University and the University of Michigan examined the existing body of research on DNR orders. The team found that many studies show that physicians often take a checklist approach to obtaining patients’ code-status preference, which doesn’t meet the informed-consent standard. Researchers also described other studies, which found that physicians accurately predict those preferences only about half the time, and that patients’ families overestimate a patient’s desire to receive CPR and other life-sustaining treatments.
Problems with policies and training
Dr. Yuen, who is a third-year internal medicine resident at New York-Presbyterian Hospital/Weill Cornell Medical Center in New York, says that her team learned that DNR discussions derail for a variety of reasons. Some of the causes are obvious, while others may come as a surprise.
One obvious challenge is the fact that physicians receive little if any training in how to conduct DNR conversations and they have scant access to good role models. Another DNR dead end, especially in teaching hospitals: having residents or even medical students “the least trained individuals on staff “conduct these discussions.
Then there’s the fact that hospital policies can be unclear or confusing, and they often differ from one hospital to the next. Even worse, policies may be inconsistent with state regulations or national guidelines.
A good example is hospitals that mandate a DNR conversation with every patient admitted, regardless of diagnosis, something that’s not required by the Joint Commission. Requiring hospitalists to discuss code-status preferences with young, healthy patients admitted for a relatively minor problem, Dr. Yuen says, leaves everyone with a bad taste about DNR discussions.
“That’s a problem,” she explains, “that can contribute to inadequate discussions.”
There’s also a factor known as the “technological imperative,” in which physicians pursue aggressive interventions ” think ICDs and dialysis “because the expertise is readily available in-house. “Physicians may not be mindful that they’re biased toward offering an intervention that’s life-prolonging,” explains Dr. Yuen, who plans to do a fellowship in geriatrics.
As a result, physicians may fail to have a real conversation with patients to determine whether such interventions are consistent with what the patient wants in the long run. And physicians might recommend a treatment without making sure that patients understand that choosing not to undergo treatment is an option.
Another problem: Dr. Yuen’s team found evidence in the literature that physicians don’t always have a clear picture of what exactly DNRs entail. “There are widespread misconceptions, even among doctors, about what the DNR order means and its scope,” Dr. Yuen says.
Studies have found, for example, that some physicians caring for patients with a DNR on file say they would withhold antibiotics or blood products because of that DNR. “I’ve seen physicians withhold intubating a patient or putting the patient on a respirator for a potentially reversible cause such as pneumonia because they falsely think DNR means the patient can never be intubated,” Dr. Yuen says.
Mentioning the downsides
Physicians’ understanding of code preferences can also affect their use. Some physicians, for example, may equate a DNR discussion with a goals-of-care discussion. That mistake can lead them to assume that a code-status preference “implies what the patient’s philosophy is toward care.”
Dr. Yuen explains that physicians should instead focus on identifying patients’ care goals “and what outcomes and benefits they expect from the treatment you’re prescribing. “From there, hospitalists can provide recommendations about whether certain treatments, including CPR, would be consistent with the patient’s wishes,” Dr. Yuen suggests, “in the same manner that they might discuss recommendations for placing a central line or doing a thoracentesis.” Conversations about code-status preferences and goals of care should be linked but separate, she adds, and doctors shouldn’t simplify issues in either.
Code-status conversations also need to focus on not only the upsides, but the downsides as well. “Sometimes, doctors will say, ‘If your heart stops and you can’t breathe, do you want CPR to try to bring you back?’ ” Dr. Yuen explains. But to make a truly informed decision, physicians need to understand and discuss patients’ chance of survival after resuscitation.
Research has shown that the life-saving potential of DNR is often overestimated. Only 10% to 18% of patients resuscitated in-house survive to discharge, Dr. Yuen says. “That’s for all comers and all ages, and it hasn’t improved much over the last 20 years. Patients should understand that most of the time CPR doesn’t work, and that age is not a strong determinant of outcome.”
A discussion of other potential downsides of CPR is also warranted. “Some doctors don’t provide information about potential complications,” she says. “They don’t explain that most patients don’t just suddenly wake up and are back to normal. They don’t mention that patients often are transferred to the ICU, intubated and put on a respirator, and given IV medications to support their blood pressure, and that even then their chance of survival is very low. Many patients don’t foresee that as an outcome.”
To put that concern into perspective, Dr. Yuen cites research showing that only 4% of residents whose DNR conversations were recorded addressed CPR survival in clear terms. Nearly all failed to bring up comfort care measures and natural death as alternatives.
Who to have the talk with
Dr. Yuen acknowledges that DNR discussions might be particularly difficult for hospitalists, who have to broach a complex topic with patients they’ve literally just met. Still, she thinks hospitalists have a duty to raise the issue with very ill patients if there’s no preference documented in the hospital records or elsewhere.
Dr. Yuen recommends a code conversation with patients or their surrogates for patients with diagnoses such as end-stage COPD or class IV heart failure, terminal cancer or advanced dementia. It’s also warranted for patients who’ve suffered irreversible loss of consciousness or have very poor functional status due to severe, irreversible illness or injury, and for those whose likelihood of surviving resuscitation is low and risk of cardiac or respiratory arrest high.
Dr. Yuen’s research team also recommended that hospitalists or other attendings conduct the conversation (don’t pass the job off to an intern or resident), and that physicians with inadequate skills and residents get training. If firsthand, onsite training isn’t available, educational resources such as standardized patient-simulation courses, online modules or videos can help. (See “Getting better at DNR discussions” on page 23.)
Moving toward incentives
Finally, Dr. Yuen says that hospitalists should take advantage of their unique vantage point on what can be wrong with DNR orders.
“If hospitalists see that a lot of patients with poor prognoses are getting very aggressive care that has no benefit,” she explains, “they can bring together physicians to address the situation with educational activities.” Physicians should, for example, think of incorporating DNR and EOL care discussions “or the lack thereof “in M&M conference proceedings.
“It’s important to know what was addressed prior to the patient’s decompensation,” Dr. Yuen says. “Hospitalists could also talk to hospital leadership, and the nurses and social workers who are very attuned to these issues, to come up with creative strategies to change the culture of the institution.” That culture needs to be geared toward recognizing the importance of clarifying patient goals and providing end-of-life care that is consistent with them.
In the meantime, she points out, studies show that patients and family members who’ve been able to discuss end-of-life preferences are more satisfied health care customers than those who don’t. As more hospitals (and, eventually, hospitalists) have their patient satisfaction scores publicly reported, Dr. Yuen says she looks forward to DNR discussions becoming a quality measure “complete with financial incentives.
Using financial incentives such as pay for performance to reward quality DNR discussions is certainly one way to go, she points out “as long as measures don’t just promote more DNR conversations. “Measures need to improve the quality of end-of-life care,” she says, “and be geared toward having timely discussions early enough in a hospitalization so patients can be part of the decision-making process. Measures need to ensure that patients or surrogates have the information they need to make informed decisions.”
Certainly, if Medicare rewards hospitals for improving the quality of end-of-life care, that will create a powerful incentive for hospitals to train physicians in these discussions, she adds. In the meantime, Dr. Yuen says, “Hospitals can incentivize physicians in nonmonetary ways by stressing improved patient satisfaction scores or recognizing physicians who serve as good role models for end-of-life communication.”
Bonnie Darves is a freelance health care writer based in Seattle.
Getting better at DNR discussions
WANT HELP LEARNING how to conduct more effective discussions on DNRs? Here is a list of resources that can help:
- The University of Rochester Medical Center has put together modules geared to ACGME competencies, including holding a family meeting about CPR and DNR.
- The University of Chicago faculty have spearheaded the CHAMP (Caring for the Hospitalized Aging Medical Patient) Web site, with this discussion of advanced directives.
- The Medical College of Wisconsin’s EPERC site (End of Life/Palliative Education Resource Center) offers a variety of case studies with teaching points.
- The Center to Advance Palliative Care has produced educational modules geared to critical care fellows to improve communication skills.