skip to Main Content

by: Alex Smith @alexsmithMD

What if the above form was the default for patients with serious illness?  Most current advance directive forms and the POLST have no default – although one could argue that our default without a form is full code. But what if we could set a default on these forms, so that when a patient received a diagnosis of a serious life limiting-illness, the default option was Do Not Resuscitation (DNR)?

Scott Halpern and colleagues tried this approach in a study published in Health Affairs of 132 seriouly ill outpatients with incurable diseases.  Patients were randomly assigned to complete one of three advance directives:

  • Comfort default: default of DNR.
  • Life-Extension default: default “full code.”
  • Standard advance directive: patients chose preferences for rescusitation.

Patients in the two “default” pathways could change their advance directive by crossing out the default, initialing the cross out, and selecting another option.  You can see examples of all three advance directive forms in the appendix.

What did they find?

  • Nearly 80% of patients in the comfort-care default accepted the default.
  • 61% in the standard advance directive chose comfort care
  • 43% in the life-extension default crossed out the default and changed to comfort care
  • After explaining the manipulation of the study forms into 3 possible advance directives, only 2.1 of patients changed their selection.  People were pretty content with what they selected the first time around.
  • Satisfaction with the advance directive form was high for all 3 versions of the form.

What’s it all mean?  Defaults matter.  Defaults are part of the behavioral economics revolution, a subject we’ve discussed previously on GeriPal.  This study suggests that peoples preferences for resuscitation are not deeply held, they are highly influenced by the (somewhat) arbitrary choice of the default options.  This may be true  because people have very little previous experience to ground their preferences, and often have no idea what sort of care they would prefer in a future state near death.

Should we change the default choice for patients with seriuos life-limiting illness to DNR?  Why not?  Seriously.  As long as you give the patients a clear path out of the default, why not? 

To be sure, the forms matter far less than the discssion.  But let’s not kid ourselves here, folks.  Even if we try our best to educate all clinicians about how to have high quality discussions, the forms matter.  Forms will inevitably be used to guide the discussions and influence patient choices.

So I’m waiting to hear of the first health system to try it.  Once we have one or two, others will be brave enough to follow.  Why aren’t we changing?  Because, as this study shows, it’s hard to go against the default!

NOTE: This is the second in a series of posts this week for “Code Discussion Week.” Come back everyday this week for a new post focused on CPR AND DNR Discussions.

Here is a running list:

This Post Has 13 Comments

  1. At least one of the problems with this study is the false forced choice presnted to study subjects between life prolonging measures versus comfort measures, as if the two are mututally exclusive. If you threaten people with unmitigated suffering, we should not be surprised that they opt for comfort measures. But this is not (in reality) an either-or choice, though it makes for a clean research design. The decision for or against CPR should be made not based on fear of suffering (which can and should be mitigated regardless of the patient's goals for care), but on whether CPR may help a patient achieve an important goal. This is, obviously, critically dependent on contextual factors that are not addressed in this study design. CPR in a post op cardiac patient with a rhythm disturbance is quite likley to restore life of a reasonable quality. Most of us would recommend a trial of CPR under such circumstances. Conversely, if the patient has advanced organ system failure or advanced cancer or advanced dementia, CPR is not likely to achieve such meaningful goals.
    I think it is irresponsible to present patients with a false set of defaults (misery+CPR versus comfort+no CPR) and then suggest that they are making an informed choice. This is the kind of inasccurate and potentially manipulative thinking that gets demagogues going. As a field we should be really careful not to get caught up in the quick (but wrong) fix.

  2. Thanks Diane for these strong words of caution. You're absolutely right that we need to be very careful here. The conversation about goals of care is far more important than the form.

    And I worry too about the false dichotomy between life-prolonging and comfort oriented treatment. To be fair, this is the way such choices are currently presented on most advance directive and POLST forms, not a specific shortcoming of this study.

    My understanding is that in many parts of the rest of the world, patients with incurable diseases of the chest, such as those in the study, would not be offered CPR. It would not even be a choice.

    We have put CPR on a pedestal in this country – partly for legal reasons, such as the Nancy Kruzan supreme court ruling allowing states to set their own evidentiary standards for refusing life-sustaining treatment. And partly, I suspect, for cultural/religious reasons having to do with the sancity of life, and deeply held convictions that we should be allowed freedom and autonomy to make our own choices about our healthcare.

    Any proposed intervention around CPR/DNR/Code that is seen as manipulative will be hard to sell to the American public. However, one of the keen insights from behavioral economics is that we are inescapably making such choices all the time – even the mere ordering of choices on a form, or reporting of what most people would do, or our current default of full code – all of these choices "manipulates" patients. We just don't recognize these decisions as being manipulative, when in fact they are just as manipulative as the alternatives presented in this study.

    The idea is to be more thoughtful about the presentation of these choices in a way that balances concerns about risks/harms/outcomes for most patients, and still allows people the freedom to choose.

  3. Folks

    I agree (as an author strongly). The forms did not have that wording. I get your worry about the wording in the article but it was just to distinguish the different types of advance directives. We said

    One-third of patients were assigned to receive a “comfort default advance directive” that defaulted to the goal of relief of pain and suffering and nonreceipt of life-sustaining interventions. Patients were instructed to make other choices if they preferred by crossing out the default options and initialing lines next to their selections.

    Another one-third of patients, in the “life-extension default advance directive” group, received a directive that defaulted to the goal of life extension and receipt of potentially life-sustaining interventions. Again, patients were shown how to make alternative selections.

    Patients in a third group were assigned to receive a “standard advance directive” that required patients to actively choose their goals of care or preferences for specific interventions. As in usual practice, if patients did not make active choices, surrogates and clinicians would make decisions if patients lost capacity.

    The ACP was the "check list" one mostly cause that is what is out there in the literature

    I hope this explains some stuff

  4. I do worry a bit about a default option for all of the reasons that Diane lists. Forms are vitally important, but it is well known that limited health literacy prevents people from understanding these forms at baseline without pre-marking something on the form that they may not understand. In addition, this vulnerable population is much more susceptible to framing – either positive or a negative framing. I worry that without a thoughtful discussion, many people will sign these forms without making an informed choice.

    Should any of these forms be allowed in a medical record without someone at least verifying that the check box marked actually aligns with the patient’s preferences? In my geriatrics clinic, I have had more than one experience where a patient handed me a signed and notarized directive. However, when I summarized what was on the form and how that would translate into the care they would receive in the hospital, they were shocked to realize they had not filled out the form correctly or that they had interpreted their choice differently than a medical professional would.

    I also wonder how a default option would go over in minority communities who often already feel distrustful of the healthcare system (and often for good reason).

    Although forms are vitally important, I feel that we lose the actual person and their story among the list of check boxes. Reading Diane’s post, I am reminded that no 2 people are the same and no 2 situations are the same. An AD for a young healthy person going in for surgery is different than an AD for a frail individual with widely metastatic cancer. Whether we intend for them to or not, the same forms with the same list of treatment options are presented the same way to people with very different levels of disease and markedly different prognoses.

    In addition to the advance directive (or POLST), I think it is our job to also tell/translate the patients’ story into the medical record. Information about what the marked check box means to a patient and why they chose it may add much more helpful information about the patient when myriad other complex and difficult medical decisions need to be made in addition to CPR.

  5. Rebecca

    Everything is a default. We currently have a default to doing more. We do not make sure people who 'want' CPR know what they are getting into

    And I am not a form advocate tho I would say we have defaults in our conversations too. There is no neutrality.

    I also wonder how we upscale if everyone is different


  6. Really cool behavioral economics article. Echos the organ donation default data. People don't like to spend time making decisions so we accept the default…

    Bob's point that there already is a default is spot on. We will never get away from that.

    I agree with Diane and Rebecca that the person, the story, there experience is paramount. But, Alex said it well "But let's not kid ourselves here, folks…the forms matter."

    The reasons to not change the default are purely political, otherwise, it's a great idea. I just wouldn't want to be the one to propose it.

    I remember my 97 year-old declining patient saying "They wouldn't resusciate me!" Oh yes they would! We filled out all the appropriate forms.

  7. I do like the DNR default with perhaps improving the form and continuing to education providers (of course we know it isn’t always education , it is changing practice). Several years ago, a palliative care physician I know was working on a Masters (I believe in Ethics) and had discussions with three groups,; physicians, nurses, and general public (older adults). In the nurses group we had about 20. During the discussion the physician writing a medically futile DNR came up. Although most agreed it was wrong to do CPR on someone who was end stage anything there were only two of us who supported Medically Futile DNR. One nurse said-who do you think you are, God? Isn’t it rather doing CPR on someone inappropriate for it playing God? As the physician pointed out doing a code on someone is the only intervention you do not have to have informed consent for. I am not saying we shouldn’t code most people but by defaulting to DNR I think I would empower some to make that choice. I have been in a code where the person was actively dying-91 yo, end stage cardiac but he wanted to be coded. Urine output had already stopped; agonal breathing was taking place; you could hear the ribs break, someone handed me the defibulator pads because they didn’t want to do it. It is about 15 years ago and it still makes me sick to my stomach when I think about. There are all sorts of horror stories those of us who have worked in hospitals have. Default DNR might be one way to help.

  8. Bob’s point is well taken…our default is to “code.” I just had a conversation yesterday afternoon that was similar to Dan’s comment. My 90 yr old frail patient said “If I go down, no one would do CPR on me. I am 90 years old.” When I described to him and his daughter what would likely happen, they were able to make an informed choice.

    I think one way to upscale for everyone who is different is to make it mandatory that the patient’s story is included on the form along with the check boxes. To document whether the patient understand their prognosis, what a do not resuscitate means to them, etc. I had a patient who had marked the DNR box and then when I asked told me that he marked it “because I only wanted to be on machines for a week or so.”

    I think the addition of the patient’s story along with the documentation that the patient understands the ramification of their choice (even if that choice is to the default full code) can make the checkboxes come alive and help the forms become more individualized.

  9. One of our medical students, Annie Lu, just presented findings at ATS this week that show how doctors inadvertently communicate the "default" in medicine with their rhetoric.

    She did a qualitative analysis of 114 encounters between attending hospital physicians and a standardized patient and his surrogate. The patient was in his late 70s with advanced metastatic gastric cancer who presented from a nursing home with hypoxia due to cancer progression and, perhaps, infection.

    When physicians discussed intubation as a treatment option, they used words like "have to" "need to" (e.g., "When the oxygen is this low we need to intubate.")

    When they discussed not intubating, and instead providing comfort measures only in anticipation of death, they used more discretionary words like "could" "an option" (e.g., "Or we could just keep you comfortable…"). Also, this treatment pathway was always mentioned after intubation (if at all).

  10. I agree with Bob–everything is a default. Our entire society, including our health care system, is defined by "choice architecture"–read Thaler and Sunstein's book, "Nudge," for more on this (

    Indeed, much of public health policy making turns on the idea of setting up defaults in our health care system that have the greatest benefit for the most people–a utilitarian goal (e.g. vaccination policies and various types of screening–see Halpern SD et al NEJM 2007). So why not do so when we we observe the current default– CPR for all regardless of context/benefit–ends up harming those who are unlikely to benefit from CPR? Survival rates for those patients with end stage illness are terrible. We certainly wouldn't recommend interventions with success rates so low. Also, changing the default doesn't take away the possibility that patients could "opt out" and choose CPR. Vaccination policies do not change the possibility of parents with strong convictions about vaccines to opt out of vaccinating their children (though there may be other societal repercussions of not doing so, like not being able to register for school, camp, etc). If we were to change the default for patients with end stage disease to 'no CPR/comfort', they would still have the option of choosing to receive CPR if that were consistent with their goals, beliefs, etc. For example, I imagine an orthodox Jewish rabbi choosing CPR–and opting out of the 'no CPR/comfort' default–on the grounds that any additional life is sacred regardless of the impairment to neurological/physical functioning should he survive CPR. The reality of changing the default is that choosing CPR when it is thought to not be beneficial will be the minority of patients, thus we would be mitigating potential harms by simply changing the default, or at a minimum changing the approach to the conversation.

    My colleagues, Millie Solomon and Eric Krakauer, and I argued for a more nuanced, differentiated approach to CPR determination ( We suggested that for those not thought to benefit from CPR (e.g. patients with end stage dementia or metastatic cancer) the default approach to the discussion should be 'not to recommend CPR' but still be open to the individual patient's values and goals. For the patient who is imminently dying–and these conversations have not been had already–then we suggest that the approach should be to 'not offer CPR' as a default, but again be open for unique (and unusual) values and goals that would make an attempt at CPR appropriate. This sort of "liberal paternalism"–allowing for choice, but setting up the defaults to influence what, medically speaking, is the "better option" seems not only reasonable from a public health perspective, but a more ethical policy than the one currently in practice. It's nice to see some empiric research showing the results changing the default has on patient's end-of-life decision making. I agree with the authors, this should be studied further with advance directives to look at the impact this would have on important patient centered outcomes… My guess is we will see significant changes in the number of patients who are harmed by CPR, and less unwanted care.

  11. I've been thinking about the idea of using a form with default code status for a little while now. And I have a couple of thoughts.

    1. The really interesting view of the article is that it shows that how we present CPR makes a big difference. The devil’s advocate side of me says that all it shows is that we can trick people into signing things the way we want, which is more concerning than interesting. As with video’s on CPR, yes we can scare people into being DNR. Great.

    2. As a primary care geriatrician in the office and nursing home, the use of a default code status form would not fit the nature of the relationship I have with my patients. Whether it is the decision to exercise, pursue cancer screening or adhere to medications, overall, I take a very collaborative approach with my patients. I talk a lot, I explain things, I try to ensure understanding. Using something like this would introduce an element of paternalism that, on the face of it, makes me uncomfortable. I don’t even know if it would stick when my patients get hospitalized. Maybe this makes sense when the doctor has no relationship with the patient.

    3. I don’t trust doctors to exercise paternalism for older adults. Often young physicians will see a patient with the diagnosis of dementia, from a NH who is DNR and assume they are comfort care. Or assume my patients have a worse quality of life than my patients would feel. Or that living in a nursing home is automatically a terrible quality of life. With so much ageism rampant, I don’t trust other doctors.

    4. It’s not clear to me what the goal is with this form. It’s not like all older adults should be DNR any more than all of us should be DNR. The chances of any of us coming through a cardiac arrest in a shopping mall with intact neuro function is only about 3%. When do we pull out the default DNR form? To me, the goal should be to make sure the code status reflects a patient’s values in a way that is medically efficacious. Shouldn't we be measuring fit instead of the number of older adults that we can trick into being DNR? A robust 85 year old with life prolonging goals and a high tolerance of pain is different than a frail 70 year old patient who has comfort goals.

    5. What is the angst in older adults being full code? (I do feel it too). But my point is shouldn't we think about each situation and figure out a better way of communicating it? Instead of giving up on trying to communicate it.

    6. As a pcp, my code discussions are longitudinal and they are not brought up in a vacuum. I disagree that we are always using a default. Defaults only matter if the discussion starts with the doctor talking. If it starts with a doctor listening and knowing the patient, then the discussion can be anchored in what a patient says about themselves or decisions they have already made or experiences they have had. So if I say to a patient, “you were miserable in the ICU last time and said you hated it and everyone after CPR ends up in the ICU for days to weeks” there is no random, impersonal default. Then the patient has to weigh their decision to pursue CPR against their desire to avoid the ICU.
    Anyway, I rambled, but we all approach these discussion differently and in our own contexts. My ER buddies do these discussion while doing chest compressions. Obviously it is different. But as a PCP, I would have a hard time using a form like this.

  12. joshuy. Those are good points. As a PCP who takes care of elderly and nursing home patients, I frequently see patients who have 'end-stage' chronic illnesses who would likely not survive CPR. I wonder why we feel obligated to offer CPR as an option to them at all, default or not. If we knew an antibiotic would be ineffective for an infection would we ask the patient if he wanted it anyway?

  13. I agree pvel. I don't offer PEG tubes for wound care. I don't even bring it up. There are still several issues I have with defaults though.

    1. It focuses on the procedure instead of the goals of care. That always bugs me.

    2. It is still offering a procedure that hopefully the pt/family won't pick. What happens if the form "doesn't work" and the family chooses full code anyway? Now how is one going to "advance" their code status. Now will one say, well we offered it but it doesn't work? Implicit in the offering is that it is a viable option.

    3. It still feels way too manipulative for me. I feel like I'm at a used car dealer with this type of technique.

    4. I think it is unnecessary. There are better options.
    a. A facility policy on futile CPR (i.e. unwitnessed arrests in a NH setting)
    b. Being up front and recommending against it and taking the time to explain in real words why.
    c. Refusing to do it and explaining why
    d. Asking people what goal they hope to achieve and being clear as to why CPR won't get them to their goal (not therapeutic, no functional improvement, not a comfort oriented therapy, makes people worse off, robs people of a peaceful death without any benefit, not great for those who are unwilling to gamble etc etc).

    All these options are more honest, deepen relationships between doctor and patient (the last option especially). Usually if I press enough, patients have their own value/goal/risk tolerance that will limit their desire for CPR. I just need to listen carefully and point it out. It takes time. Forms are easier. Difficult to upscale what I do I suppose, (and I enjoy what I do). But seriously, how hard is it to listen?

    5 Questions to ask
    1. Is longevity still a goal?
    2. What is your ideal/minimal quality of life? Are you willing to live with a worse level of function?
    3. Are you willing to tolerate the trauma of CPR?
    4 How much are you willing to gamble that CPR usually won't work, will prolong your dying, or leave you impaired?
    5. Have you undergone medical treatments you regret?

    If the answer isn't yes to all of the above, then the physician has their "in" to discuss a DNR status. As long as we are focused on goals, there are a million reasons to explain to people why CPR won't meet their goals.

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top