When a patient shuns care, doctor checklist includes compromise

Scenario Mr. K, a homeless man, is admitted to the hospital with bacterial endocarditis and agrees to start intravenous antibiotics. On hospital day three, he decides that he “has to go,” and asks for a prescription for oral antibiotics. He also asks if you will give him taxi fare to a nearby shelter.

Reply About 1% of hospitalized patients nationwide request discharge despite medical advice to the contrary. Such requests create ethical and professional dilemmas for medical personnel, who strive to provide the best care possible for their patients.

In the U.S. health care system, patient autonomy is highly respected, and patients with sufficient decision-making capacity are entitled to make “bad” decisions that run contrary to the professional recommendations of their doctors. To ensure that patients are making informed decisions, as well as to satisfy administrative requirements that provide a modicum of protection from a risk management and medical liability standpoint, we typically ask such patients to acknowledge in writing that they are leaving “against medical advice.”

Many (but by no means all) patients who want to leave against medical advice are, like Mr. K, poor and disenfranchised. A disproportionate number have histories of mental illness or addiction to illicit substances and thus are even more vulnerable to receiving suboptimal care from a medical system that may not be able to meet their complex needs. In addition, such patients are easily stigmatized by medical professionals, which further alienates them from the care they need. As doctors, we are obligated not only to respect our patients' autonomy, but also to act as their advocates, even when they disagree with our recommendations.

While autonomy is a critical component of these difficult cases, it is not absolute. For patients to make any informed decision, whether it is to consent to or refuse medical recommendations, they must demonstrate sufficient decision-making capacity. In many senses, capacity is a situation-specific judgment call, made informally by clinicians every time a patient is asked to make a decision. Of course, when decisions carry higher consequences, such as the decision to refuse critical aspects of their care, a more formal capacity assessment is essential. A patient who has decision-making capacity is able to express personal wishes clearly, demonstrate understanding of his or her medical condition and the proposed treatment alternatives, appreciate the significance of his or her decision, and make a choice that is consistent with his or her own values.

Psychiatrists may be asked to assist in making capacity assessments, particularly for patients who may be impaired by underlying mental illness or when the patient's decision may have sufficiently dire consequences.

However, clinicians can and should gauge their patients' decision-making capacity independently. When patients lack capacity, surrogate decision-makers are asked to decide for the patient using the doctrine of substituted judgment — that is, making a decision in accordance with the patient's prior known wishes.

Only when patients are incapacitated and surrogates unavailable are clinicians empowered to make decisions that override their patients' stated objections. Fortunately, such situations are rare, and assuming there is not a medical emergency requiring immediate action, these cases typically require involvement of an ethics consultant, formal administrative approval or involvement of the courts.

Look at decision-making skills

Thus, the next step with Mr. K is to assess his decision-making capacity, and, if he lacks capacity, to try to identify surrogate decision-makers. In this case, Mr. K clearly understands the magnitude of his decision and indeed wants to leave despite the possibility of life-threatening complications.

Simply assessing capacity and allowing Mr. K to leave against medical advice does not satisfy our professional obligation … nor does it absolve us from looking out for his welfare. So, after assessing his capacity, we should explore his decision fully.

Is his choice due to incomplete or erroneous information? Does he think that outpatient treatment will be just as effective? Is he worried that further hospitalization will put his employment or access to social services at risk? An effort to understand his viewpoint may well pay dividends. Perhaps such an investment of time and expression of consideration for his well-being will alleviate his concerns and convince him to stay.

Next, we must clearly convey our disagreement with his choice and our rationale for disagreeing. It is quite reasonable, however, to continue to provide him with the same level of consideration and care (assuming he is agreeable to receiving it) that we routinely provide to all patients.

Does this include covering Mr. K's taxi fare? While our prior actions and deliberation have been fairly straightforward, we are now entering a gray zone. What is a clinician's responsibility when it comes to honoring the requests of patients who do not follow our recommendations? Would our facilitation of his ability to physically leave the hospital be tantamount to condoning or legitimizing his actions? In this case, if indigent patients who leave the hospital after consenting to treatment routinely have their transportation costs reimbursed, there is no compelling reason to deny Mr. K this service. I expect that many clinicians would offer reasonable dissenting opinions.

Further care still needs to be arranged and ensured, if possible. While we all recognize that a complete course of intravenous antibiotics with sufficient medical surveillance would be a more ideal treatment for bacterial endocarditis, we must realize that, perhaps, a compromise can be achieved.

Maybe Mr. K would agree to receive intravenous antibiotics as an outpatient if this would be safe and feasible (albeit unlikely given his homelessness). In addition, Mr. K's request for a prescription for oral antibiotics indicates his acknowledgment of his illness, as well as his willingness to undergo treatment on his own terms. Most doctors would agree that oral antibiotics would be better than nothing. Perhaps a mutually acceptable answer would be to provide oral antibiotics, arrange close outpatient follow-up and ensure that Mr. K knows the warning signs that would necessitate prompt return to the hospital. This plan, while admittedly not in accordance with accepted medical guidelines, may well be the best way to engage Mr. K in a trusting therapeutic partnership.

As all of us who have dealt with a Mr. K will attest, small victories and imperfect compromises are often the best that we can do. Clinicians who have the patience and fortitude to make concessions, but also to honor their professional obligations and judgment, are often best able to achieve better solutions amid challenging situations.

— Andrew G. Shuman, MD, Division of Medical Ethics, Weill Cornell Medical College, New York