Care Redesign

The Value of ICU Care at the End of Life

Article · September 28, 2016

As a physician who practiced critical care medicine for over 20 years, I find myself frequently reflecting back on the most onerous part of that responsibility: dealing with death. In an age when we are trying to understand what really constitutes value for patients, I remain most troubled by health care’s failures to match what patients want and what too often happens to them when life ends.

The value approach is a balance between what outcome a patient wants relative to the cost — both financial and nonfinancial — to achieve that outcome. An ICU stay should improve value for the patient, with the outcome being to survive this awful experience and, hopefully, to live many more years. Unfortunately, that positive outcome does not happen as often as we all would like. The costs of such end-of-life experiences in the ICU are immense.

For far too many patients, their critical care experience is of very low value. When asked how they would prefer to die, about two-thirds of people say they want to die at home unless it is too much of a burden to their loved ones. Yet since the 1980s, less than 20% of patients die at home, and over a fifth of all deaths in the United States today happen in an ICU or shortly after an ICU stay. So we in health care are not doing a great job matching what people want with the outcomes they get.

For thousands of years, people died at home surrounded and supported by their family and friends. While not always pleasant, it was the way it was. The dying process for mankind was transformed by the emergence of hospitals in the 19th century and the development of ICUs in the 20th. With hospitals, we introduced a new place to die outside the home. With ICUs, we intervened with life support and stayed the natural processes of respiratory and circulatory failure — the common pathways to death. The advent of the ICU transformed dying from a natural process to one where care stopped only when death was imminent.

On the cost side of the value equation, 25% of U.S. health care spending goes to the 6% of people who die every year. ICUs account for 20% of all health care costs. Thus dying in the ICU is both more expensive than dying at home and less desirable in the eyes of most people. So why does it continue to happen?

One explanation is that death is a retrospective diagnosis. If you knew for sure you were going to die, then you might make different choices about care. Another problem is that most ICU patients lose their ability to make informed choices once they arrive, because of illness and treatment. Decisions are then delegated to families and significant others, who often neither understand the patient’s wishes nor have sufficient confidence to make a decision to end life support. While advance directives are an attractive concept, the reality is that they are not legally binding documents and are frequently put aside by well-meaning families.

Deciding What Is Right — With the Patient

The problem is both humanitarian and economic. We want to do what patients want, and we do not want to burden the health care system with expensive, unwanted life support at end of life.

Yet it is a difficult problem with many facets. Since physicians can rarely predict when someone will die, and since we have been trained to try to prevent death from illness, the default position is to help. Even with advanced, recalcitrant malignancies, progressive neurological disorders, and many other diseases, it is difficult to know when death is imminent. Often it does not become clear that death is inevitable until patients are well into life support ICU care.

So how do we transform end-of-life care in this difficult situation? We need patient engagement. Not just when they are ill, but when they are well. The best time to have a conversation about the end of life is not when the patient is in impending respiratory failure, but when everyone is well and clear-headed. Patients need to know what ICU care really means. How does it feel to be intubated and receiving cardio-respiratory support? What would death in an ICU look like? In a small study about do not resuscitate / no code orders, when patients saw a video depicting CPR, many chose not to have CPR. The same should happen about the realities of death in an ICU. We should plan for dying the way we plan for births. We should feel comfortable talking about death with our loved ones.

Physicians need the time and support to have frank conversations with their patients about their wishes. The fast pace of American medicine often prevents well-meaning clinicians from having these important conversations in their offices — instead postponing it for a hospital ward when the patient is short of breath or as the patient is being transferred to an ICU with impending respiratory failure. Treating physicians need to develop better relationships with palliative care teams, who have tremendous things to offer patients with advanced diseases.

While physicians should not be rationing care, there need to be limits to what we do at the end of life. Just because we can do something in an ICU does not mean we should. Providing comfort to a patient at the end of life is also an important role for the health care team. Value for the patient means providing the outcome they want based on their unique situations. Unless we ask, we will not know.

New Call for Submissions ­to NEJM Catalyst

Connect

A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »

More From Care Redesign
Responding to Opioids Epidemic - chronic pain, opioids epidemic data

The Physicians’ Quandary with Opioids: Pain versus Addiction

The national addiction crisis has led to a crackdown on physician prescribing abilities for opioids. Frontline physicians and clinical leaders are torn between wanting to prevent addiction and wanting to treat chronic pain.

2013 Jefferson Emergency Department Safe Opioid Prescribing Guidelines Poster

A Grassroots Effort Led by Emergency Physicians to Mitigate the Escalating Opioid Epidemic circa 2012

Collaboration among Emergency Departments in addressing a public health crisis can have a broad-reaching impact that addresses the unique needs of an at-risk population.

Patient Outcomes Stratified by Low, Moderate, and High-Risk HEART Score

Improving Emergency Department Care for Low-Risk Chest Pain

How Kaiser Permanente Southern California improved ED standard of practice for possible acute coronary syndrome by adopting an evidence-based clinical decision tool.

Academic Innovations Collaborative Colorectal Cancer Driver Diagram - Leading Transformational Change in Primary Care Delivery

Five Keys to Leading Transformational Change in Primary Care

Lessons learned from a Harvard Medical School intervention to strengthen primary care delivery that emphasized teamwork, reflection, and leadership engagement.

Providing Free Food as a Treatment for Diabetes Yields Improved Outcomes for Patients While Reducing Cost of Care - Fresh Food Farmacy - Geisinger program for addressing food insecurity in patients with diabetes

Prescribing Food as a Specialty Drug

Geisinger’s program of providing free food as a treatment for diabetes yields improved outcomes for patients while reducing the cost of care.

Care Design - The H-Evolution of Intensive Care Units - ICU and Human Care Bundle Human-Centered Care Model

My Favorite Slide: The ICU and the Human Care Bundle

Eight aspects of care to improve ICU management from a human-centered care model.

Ring01_pullquote_ Specialty Musculoskeletal Care Services for the Underserved

Specialty Care for the Underserved

The opening of a new medical school provided impetus to restructure musculoskeletal services to give the safety-net population timely access to needed specialty care.

Levels of Independent Empowered Self-Dialysis

Is “Empowered Dialysis” the Key to Better Outcomes?

Self-dialysis patients enjoy far lower mortality and hospitalization rates and far better attitudes about treatment. But there are hard limits to its adoption.

cognitive load theory implications for health care pullquote

Cognitive Load and Its Implications for Health Care

If we recognize and address the limits of brainpower, we can become better caregivers.

Integrated Care - Most effective means of integrating mental and behavioral health services

Survey Snapshot: Integrated Care Makes Sense — but Barriers Remain

NEJM Catalyst Insights Council members call for integration of mental and behavioral health services with primary care.

Connect

A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »

Topics

Primary Care

150 Articles

The Physicians’ Quandary with Opioids: Pain…

The national addiction crisis has led to a crackdown on physician prescribing abilities for opioids.…

Social Needs

62 Articles

What If We Treated Mental Health…

How would we react if only 1 out of 10 people with cancer saw a…

Mental Health

19 Articles

What If We Treated Mental Health…

How would we react if only 1 out of 10 people with cancer saw a…

Insights Council

Have a voice. Join other health care leaders effecting change, shaping tomorrow.

Apply Now