Article Text

Download PDFPDF
Advancing the science of discharges against medical advice: taking a deeper dive
  1. David Alfandre
  1. VHA National Center for Ethics in Health Care, NYU School of Medicine, Department of Medicine and Population Health, New York NY 10010, USA
  1. Correspondence to Dr David Alfandre, VHA National Center for Ethics in Health Care, NYU School of Medicine, Department of Medicine and Public Health, New York NY 10010, USA; david.alfandre{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Against medical advice (AMA) discharges continue to confound. These discharges, when patients leave the hospital prior to a clinically specified and physician-recommended endpoint, are associated with worse health and health services outcomes. Although their prevalence has remained between 1% and 2%, the rate is markedly higher in vulnerable populations with mental health and substance use disorders, HIV and the uninsured.1 2 Compared with patients discharged conventionally, 30-day mortality is higher, and 30-day readmissions are more likely after AMA discharges.3

Contemporary research on AMA discharges has primarily been directed towards elucidating the patient-related factors that are associated with its adverse outcomes. Identifying the demographic variables and patient comorbidities that are more likely with AMA discharges is intended to uncover what drives this phenomenon and to point towards what interventions can mitigate their impact. Also, because readmissions are more common after AMA discharges, and acute myocardial infarction (AMI) is a targeted condition for Medicare’s Readmission Reduction Program, reducing readmissions after AMI is an important goal for improving healthcare quality and resource utilisation. Although AMA discharges are excluded from Medicare’s readmission penalty calculation for patients admitted with AMI, attempts to reduce AMA discharges are welcomed.

The report by Kwok and coauthors in their Heart paper,4 focuses on the high-risk cohort of patients admitted with AMI. The investigators drew on 2010–2014 hospital data from the Nationwide Readmission Database and included adults hospitalised with an admission diagnosis of AMI who were either discharged home or AMA. The authors examined independent demographic, comorbidity, procedural and hospital variables associated with AMA discharge as well as the influence of AMA discharge on 30-day readmissions.

The study confirms many findings that have been well established in the literature. The study’s AMA discharge prevalence of 1.5% has been demonstrated previously as have the common independent patient-related predictors of AMA discharge, including younger age, male gender, uninsured or Medicaid insurance, and a history of substance use disorder.5 6 In their analysis, patients discharged AMA had a twofold higher 30-day readmission rate which is also consistent with previous findings.3 7

The study also demonstrated numerous associations between the intensity of inpatient care and AMA discharge, which represent important new findings that advance the literature. Specifically, the odds of AMA discharge were significantly reduced among patients who had coronary artery bypass grafting, implantable cardiac defibrillator insertion, drug-eluting stent (DES) placement, angiograms or percutaneous coronary interventions (PCI). As a result of this study, we also have additional data about readmission risk after AMA discharge. Time to readmission was shorter (10.6 days vs 14.1), and total costs were lower (US$26 676 vs US$31 536) among patients discharged AMA, with a longer length of stay (LOS) on readmission (5.1 days vs 4.5). Patients with AMI who were discharged AMA were significantly more likely to be readmitted with AMI (OR 3.65, 95% CI 3.31 to 4.03).

Evidence of the reduced intensity of care for AMI associated with discharge disposition is important information for providing better care to this population. We do not yet know the reason for these disparities, but evidence of them is noteworthy. Why are these patients receiving fewer DES, angiograms or PCI? A common hypothesis is that a patient’s shortened LOS might make provision of these treatments less likely. Analysing discharge disposition and receipt of emergent PCI in patients with ST Elevation Myocardial Infarction (STEMI) might help to elucidate the issue. But other hypotheses deserve critical evaluation. Are clinicians not offering some patients inpatient cardiac treatments because they believe patients who are eventually discharged AMA won’t comply with the necessary follow-up for these inpatient procedures? Do patients who decline certain inpatient cardiac treatments also eventually decline further hospitalisation, and thus are discharged AMA? We cannot yet answer these questions definitively.

There are novel research strategies that are within reach for making meaningful improvements to patient care in this area. First, to better clarify why patient’s leave AMA, there should be an expanded consideration of what non-patient factors potentially contribute to the process. Kwok and coauthors add to this inquiry, as investigators have done previously,1 5 by demonstrating that AMA discharges are more likely with hospitals that are larger, in urban locations, and that are non-teaching. Previous research has also shown the relationship between hospital quality and AMA discharges thus providing further evidence that hospital-level factors may play a role in the AMA process.8

Identifying treatment team (ie, physicians, nurses, social workers, etc) and health system-level factors that may impact AMA discharges can lead to a broader range of interventions designed to mitigate the adverse outcomes of such discharges. The focus primarily on patient-level factors has not yet yielded an appreciable impact on decreasing AMA discharges, and misses the opportunity to intervene in the healthcare system. Yes, competent patients make the ultimate decision about whether to remain hospitalised, but health systems can identify if there are any elements within their control that influence that process. Along those lines, more qualitative research about patient’s and healthcare provider’s perspectives on AMA discharges, which provide valuable insights,9 have the potential to open new avenues of inquiry for engaging important stakeholders in this process of quality improvement.

What can clinicians do at the bedside in the meantime to improve the quality of care for patients admitted with AMI who are at risk for an AMA discharge? The primary goal should be to help the patient stay hospitalised till their acute condition is adequately treated. Clinicians should maintain a therapeutic alliance by approaching a patient’s request to leave the hospital with openness and empathy and by eliciting the patient’s preferences and competing priorities and addressing any unmet needs. They can also engage patients in the decision-making process for discharge by providing them with an expanded medically reasonable range of options instead of simply a choice to stay or leave.10 However, competent patients have the right to decline recommended care, so if remaining hospitalised becomes untenable, physicians should provide a medically acceptable and safe alternative treatment plan that attempts to reduce the harm associated with the patient’s decision to leave the hospital. Because substance use disorders are so prevalent among AMA discharges, clinicians can also work closely with addiction specialists to ensure the patient receives treatment for pain and withdrawal that can help to keep the patient hospitalised.11 12 Finally, clinicians who are well informed about the ethical and legal considerations of AMA discharges are better positioned (and more comfortable) focusing on the patient’s needs.13

AMA discharges remain a continuing challenge to improving the quality of care for vulnerable populations. Both more research and deeper analyses are warranted. This means addressing identified disparities in care and understanding not just patient-related factors associated with AMA discharges, but hospital, healthcare provider and health system variables that affect the quality of care. We have begun to uncover promising results by examining the surface of this problem. It’s time to dive down and discover more.



  • Contributors The author listed has contributed sufficiently to the project in order to be included as author and only those who are qualified to be authors are listed in the byline. The manuscript has not been previously published and is not under consideration in the same or substantially similar form in any other peer-reviewed media.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Disclaimer The views expressed in this article are those of the author and do not necessarily reflect the position or policy of the US Department of Veterans Affairs, the US Government or the VA National Center for Ethics in Health Care.

  • Competing interests None declared.

  • Patient consent Not requried.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles