Reconsidering Against Medical Advice Discharges: Embracing Patient-Centeredness to Promote High Quality Care and a Renewed Research Agenda

ABSTRACT

Hospital discharges against medical advice (AMA) are common, costly, stigmatizing to patients, and are associated with excess morbidity and mortality. Achieving better quality care for patients discharged AMA has been limited both by the sparse research illuminating how best to care for this challenging patient population, as well as a lack of standards regarding this clinical practice. This paper will review elements of the AMA literature and highlight the gaps, including the predictors of AMA discharge, challenges to high quality informed consent in AMA discharges, problematic aspects of AMA discharge forms, and the stigma associated with patients discharged AMA. These gaps in the evidence base collectively limit the ability to adequately and completely address AMA discharges and improve health care quality. This paper will recommend future directions to answer remaining questions for the field, and offer guidance for providing ethically sound and high quality care for the affected population. Applying the widely accepted principles of patient-centered care and shared decision making to AMA discharges offers the opportunity to improve quality of care and promote ethical health care practice.

Case: A 48-year-old man with hypertension, diabetes, and alcohol abuse presented to the emergency room (ER) with lower extremity pain and redness after 5 days of non-adherence with all his medication. He had been drinking more alcohol than usual. The patient was admitted for treatment of cellulitis and hyperglycemia. Three days after admission, with the patient’s glucose and blood pressure stabilized but the cellulitis not completely resolved, he angrily reported to the nurse that he needed to go home. The intern was called to the bedside to evaluate the patient further. After a lengthy discussion primarily focused on the risks of leaving the hospital, the patient persisted in requesting a discharge. The patient was then asked to sign a form indicating he would be discharged against medical advice (AMA). Follow-up care was not arranged. He was advised to return to the ER if he developed new symptoms.

Introduction

AMA discharges, in which a patient chooses to leave the hospital before their treating physician recommends discharge, constitute up to 2 % of all hospital discharges among medical inpatients (approximately 500,000 US discharges per year).^1–3 Compared to patients discharged conventionally, 30-day readmission rates for patients discharged AMA are 20–40 % higher,^3,4 and their adjusted relative risk of 30-day mortality may be increased as much as 10 %.² Despite the excess morbidity, health care costs, and associated stigma, AMA discharges are relatively unexamined.

The Patient Protection and Affordable Health Care Act of 2010 establishes Medicare payment reductions for hospitals that have higher readmission rates for certain conditions, and may exert enormous pressure to reduce avoidable hospital readmissions.⁵ Because patients discharged AMA are at increased risk of hospital readmission and the excess associated costs,⁶ studies that examine ways to decrease AMA discharges are needed.

Achieving better quality care and improving outcomes for these individuals has been limited in part due to the sparse research illuminating how best to care for and support this challenging patient population. This paper will describe the existing AMA literature and highlight its gaps, including the standards for designating an AMA discharge, predictors of AMA discharge, challenges to high quality informed consent in AMA discharges, problematic aspects of AMA discharge forms, and the stigma associated with patients discharged AMA. These gaps in the evidence base collectively limit the ability to adequately and completely address AMA discharges and improve health care quality. By highlighting these gaps, the paper will recommend future directions for answering critical questions for the field, and offer guidance for providing ethically sound and patient-centered care for the affected population.

Review of Existing Literature

Discharges Designated as AMA

Designating a discharge as AMA is a clinical practice that is not subject to professional standards. First, there is no clear medico-legal rationale for its designation among hospitalized inpatients. Legally, if a competent patient or their authorized surrogate chooses to decline further inpatient care, the physician must have an informed consent discussion and must document that discussion in the medical record. Medically, the actual designation of the discharge as AMA, and the associated formalized process sometimes associated with it (e.g., AMA discharge forms), does not have substantiated utility for advancing a patient’s care. In some cases, as described below, it may cause harm by stigmatizing the patient and reducing the patient’s likelihood of following up.^7,8

There is also no professional consensus on what constitutes a discharge as AMA. This promotes greater variability in the clinical use of the term, and provides less meaningful and generalizable data for research. This variability in clinical practices lacks transparency, impedes systematization, and may increase the probability that a physician’s personal values, rather than a medical standard, play a defining role in how these AMA determinations are made.

Establishing uniform standards for AMA discharge will allow for rigorous development of rational interventions to prevent or reduce AMA discharges. A proposed standard definition of an AMA discharge is: A patient request for a discharge treatment plan that is outside the range of medically acceptable options. If the patient’s preference for care is to decline further inpatient treatment, and if the physician’s agreement with that plan would be outside of accepted medical standards (e.g., based on risk, etc.), the discharge would be consistent with the above definition of AMA.

Predictors of AMA Discharge

The majority of research related to AMA discharges over the last 4 decades has focused on determining patient-related factors that predict AMA discharges. The patient factors associated with AMA discharge have been reasonably consistent over time, and include lower socioeconomic class, male sex, younger age, Medicaid or no insurance, and a history of substance abuse.¹ Commonly reported patient reasons for leaving AMA include personal or financial obligations, feeling better or receipt of social assistance payments.^9,10 The stated aim of this research was to generate potential interventions that could mitigate the number of such discharges. Despite this intent, interventions have not yet been systematically studied and published in the literature. All of these studies examined patient predictors of AMA discharges and some examined hospital-related predictors. No studies to date were designed to study the characteristics of physicians who designated a discharge as AMA. Because the decision to leave the hospital lies with the competent patient, but the choice to designate a discharge as AMA lies with the physician, a full understanding of AMA predictors requires data about physicians who discharge patients AMA.

Stigma and AMA Discharges

Although stigma in the AMA discharge population has not been systematically examined, a review of studies that have included AMA discharge data suggest a need for further research examining stigma in this population.¹¹ Haywood et al.’s study¹² was designed to test a video intervention on clinician attitudes towards patients with sickle-cell disease by using a scale of clinician attitudes towards positive and negative patient behaviors. To create the scale, clinicians rated examples of a range of patient behaviors; a history of AMA discharges was considered as negative “red-flag” behavior. This conclusion about a patient’s prior behavior suggests subjective assumptions about this population of patients. Although leaving the hospital AMA is legally and ethically consistent with a competent patient’s individual choice to decline an episode of inpatient care without stigma, clinician perceptions on this process may differ. As demonstrated in the health disparities literature, clinician attitudes towards patients can vary based on patient characteristics. This variability can in turn be associated with varying health care quality. Similarly, clinician attitudes towards patients wishing to leave AMA or previously discharged AMA may be associated with stigma and lower quality care.

Another study illustrated a clinical effect on patients if they were discharged AMA. Jerrard et al. examined the perceptions of patients discharged AMA from an ER, and found that 25 % did not return for care because they felt they had angered the staff in their decision to leave, and that they “anticipated derision” if they returned.⁷ Although it is not possible to conclude whether patients’ perceptions were due to staff behavior or other unstudied variables, perceptions that potentially lead to worse access to care after AMA discharge are noteworthy.

Documentation of AMA Discharges—the Use of AMA Discharge Forms

Although the use of specialized AMA discharge forms appears to be a relatively common practice,^13,14 understanding their perceived use and purpose has received comparatively little attention. The form, typically pre-printed and distinct from general hospital discharge forms, often includes some combination of the following elements: documentation of the patient’s choice to leave the hospital against their physician’s advice, generic associated risks, a release of the hospital and its staff from liability, and a section for the patient’s and witness’s signature.

The medical literature provides no clear history of why the forms began to be used or how they were used to advance patient care. From a legal perspective, the use of the form does not appear to be mandatory for hospital inpatients leaving AMA,^15,16 but rather assists clinicians only in ensuring adequate disclosure as part of the informed consent process for an AMA discharge. Although physicians often believe the form is useful in managing liability, the aforementioned literature suggests it provides little protection from that possibility. Moreover, some legal scholars have suggested that the use of the form in attempting to release the hospital from legal liability may violate public policy in some states.¹³

There remain unanswered questions about the optimal use and characteristics of such a form. Review of current forms indicates that some are not written at an acceptable reading level for all patients, and some include confusing and potentially coercive legal language. It is not clear, nor has it been evaluated, if these forms affect patient understanding or voluntariness as part of informed consent for an AMA discharge.

Challenges to High Quality Informed Consent in AMA Discharges

When a patient wishes to leave the hospital prior to a medically specified endpoint, the informed consent process is critical to ensure quality care. Informed consent aims to ensure that patients participate in decision making about their health care. Never simply a signature on a form, informed consent is the process that includes physician disclosure, patient understanding, and patient voluntariness.¹⁷ Some authors label an AMA discharge as an “informed refusal,” rather than an informed consent discussion that results in refusal of care.

The limited research on inpatients declining care suggests that poor quality informed consent practices contribute to the problem. Appelbaum et al. demonstrated that the most common reason patients decline inpatient treatment is lack of information,¹⁸ suggesting that informed consent practices may contribute to these conflicts. Two recent studies identified that house officers and attendings believe and incorrectly inform patients that if they leave the hospital against medical advice, that their insurance company may not pay for the hospitalization.^14,19 Both studies found no evidence of denial of payment after AMA discharges by either private insurance companies or the Center for Medicare and Medicaid Services. The larger of these studies found that attendings who believed that insurance denied payment were significantly more likely to tell patients that they might be financially responsible if they left AMA. Eighty-five percent of residents and 67 % of attendings in this study reported that they told patients about insurance denial so that patients would reconsider staying in the hospital.

These findings raise concern with regards to the quality of informed consent and AMA discharges. While a physician’s recommendation and rationale to remain hospitalized is ethically acceptable, the threat of a penalty for refusing care (e.g., “If you leave AMA, your insurance may not pay for this hospitalization”) is considered coercive and potentially undermines the patient’s voluntary choice. Similarly, the possibility that physicians limit patients’ access to treatments and services so a reluctant or ambivalent patient will agree to remain in the hospital has not been evaluated systematically. For example, do physicians attempt to coerce patients to remain in the hospital and not leave AMA, by limiting their access to medically indicated services (e.g., discharge medications, transportation, or follow-up appointments)?

Discussion

Review of the existing literature on AMA discharges reveals multiple opportunities for gathering additional data that can inform critical questions for the field to reduce and mitigate the impact on patients (Fig. 1). There are opportunities to better understand physician and nurse attitudes about patients with a history of AMA discharges. Understanding factors associated with a physician’s willingness to label a discharge as AMA can assess how they apply the label in varied settings and situations, and help to inform the creation of a formalized standard. Examining the AMA form content in general, as well as evaluating physician and patient attitudes and beliefs about the use of these forms, will help to guide ethically appropriate use of these forms. Qualitative studies that evaluate AMA informed consent discussions and their chart documentation can identify strengths and inadequacies in the process, and illuminate how patients and doctors talk about these complicated decisions.

Figure 1.

Potential research opportunities in AMA discharges.

While further research is needed to inform best practices, highlighting patient-centered care as a response to AMA discharges can help promote high quality care for this population of patients. The 2001 Institute of Medicine report “Crossing the Quality Chasm” established six aims for quality improvement to address key areas in need of progress,²⁰ one of which was patient-centeredness. The report defines patient centeredness as, “providing care that is respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions.”

The means by which patient-centeredness can be achieved is operationalized as shared decision-making (SDM).²¹ Although SDM has an accepted place in value-laden clinical decisions like prostate-specific antigen (PSA) and mammography screening, the principles of SDM are intended to apply to a very broad range of health care decisions. Some authors have suggested that SDM should be applied to most clinical discussions around care, except when the decision is value-neutral²² (i.e., unrelated to a patient’s individual values, like which size endotracheal tube to use), or there is virtual unanimity about the “right” decision for care.²¹

During hospitalization, health care decisions are still sensitive to patient preference; that is, they are value-laden, and therefore appropriate for SDM. Decisions about choosing to leave the hospital before a physician recommends can be related to a patient’s individual values (i.e., they are preference sensitive), and should be part of a SDM discussion. AMA discharges can risk minimizing the role of patients’ values in decision making by suggesting a patient is defying their doctor if they choose an option other than what the physician recommended.

In reconsidering the patient described in the introduction and in light of the described gaps in the literature, it’s possible to envision physicians intervening with patients in different and better ways. A new approach seeks to be more patient-centered and productive, and can support patient autonomy, promote harm reduction, and attempt to maintain the patient’s therapeutic alliance with the physician. This approach respects the role of a physician’s recommendation and a patient’s right to make decisions about their care. Ideally, it will maintain patients’ willingness to return for care when they are ready, while not creating barriers to follow-up.

In responding to a patient’s wish to leave the hospital, the physician should first make a determination about what constitutes a medically safe treatment plan, which potentially includes more than just the single treatment option of remaining hospitalized (e.g., outpatient follow-up). Then, the physician should empathically elicit the patient’s preferences and values regarding the decision to leave the hospital, and discuss any evidence-based risks of leaving, along with the risks and benefits of remaining. Having clear information about a patient’s values, and thus their motivating behavior, may allow a physician to target counseling more appropriately and to potentially negotiate a discharge treatment plan that is more agreeable to both the patient and physician. But this technique is not always successful and the patient may prove difficult to sway. By using SDM, and incorporating the patient’s medical status as well as their values and preferences for care, the physician should evaluate if agreeing with the patient to leave now is within the range of reasonable medical options. If that is the case, a suggestion to the patient like the following might promote the therapeutic alliance and ongoing care, if not this particular episode of care: “I can appreciate that you are eager to return home to manage your unfinished business. As we have discussed, remaining in the hospital will likely lead to faster healing, safer monitoring of your serious infection and will allow us to address your alcohol use. That is my primary recommendation for your care now. However, leaving the hospital now and getting follow-up as an outpatient is a reasonable option. As I have said, it’s not as ideal medically, but your care should fit with your preferences and I will support you however you choose. I’m here to help.”

Having obtained informed consent for the patient’s choice to leave and determined that it is not an AMA discharge, there are still additional options that can promote harm reduction. The physician could then provide a standard discharge plan and arrange for follow-up for the cellulitis and alcohol abuse in 1–2 days as an outpatient. The physician would then fully document in the medical record this informed consent discussion, including the patient’s preferences and his understanding of the decision, the physician’s primary recommendation to remain hospitalized that was declined, and the physician’s secondary recommendation that became the agreed upon discharge plan. This avoids a potentially punitive, non-productive AMA designation in the chart, while ensuring an accurate description of events is recorded.

Alternatively, what if the physician assessed that the patient’s choice to leave now would result in a treatment plan outside the range of medically acceptable options? The physician would still create a discharge plan for the patient and document the full informed consent discussion in the chart. The subsequent decision to designate the discharge as AMA (and use an AMA discharge form) would be left to the physician, but choosing to do so would not advance the patient’s care in an evidence-based way. Indeed, the available data suggests that an AMA discharge is more likely to harm the patient by reducing the likelihood of follow-up.

A physician’s choice to formally designate a discharge as AMA is partly a reflection of the disagreement about a patient’s informed choice to decline recommended care. But the AMA discharge is an inconsistent and potentially stigmatizing exception to the therapeutic alliance that physicians promote in other areas of patient care when a disagreement arises. Patient-centered care supports patients in their choices,²³ even if their choice is for a care plan that a physician does not recommend. Therefore, accepting a patient’s preferences for care, even when they do not appear to coincide with commonly accepted notions of good decisions about health, can still be embraced by physicians. This approach provides for the trust and the needed endorsement of the physician–patient relationship, if not the specific choice the patient has made.

Conclusions

Ensuring higher quality care for patients discharged AMA will depend, in part, on an approach that incorporates the ethical considerations of patient-centered care. SDM, which aims to provide care that is consistent with a patient’s values, is an important part of this process. Physicians and care teams need tools and training to engage in SDM and reduce stigma for conflicts over discharge. Further research on AMA discharges is needed to inform best practices. There should be a call for clear professional standards for AMA discharges, most importantly by defining it to systematize research and clinical practice. By setting these standards, systematic and rational interventions can then begin to be developed. The introduction of the Patient-Centered Outcomes Research Institute Funding Announcements represents an opportunity for researchers to pursue these and other questions related to improving patient-centeredness in AMA discharges.²⁴ By better addressing AMA discharges now, physicians can fulfill their dedication to deliver patient-centered care and honor their commitment to all of their patients.