Working Toward Evidence-Based End-of-Life Care

When the Affordable Care Act was proposed more than 6 years ago, the discussion about end-of-life care was limited by political fighting and inertia. Today, there is a greater openness to engaging in this important discussion, but the challenge is to avoid accepting non-evidence-based initiatives, writes Scott D. Halpern, MD, PhD, of the University of Pennsylvania, in The New England Journal of Medicine.

Dr Halpern laments the fact that policies and practices on end-of-life care as currently experienced by millions of dying Americans are backed by only a fraction of the evidence required by the FDA to approve ”even a relatively innocuous drug.” This is the case even though in the last months of life, 1 in 2 Medicare beneficiaries visits the emergency department, 1 in 3 is admitted to an intensive care unit, and 1 in 5 has inpatient surgery.

More than two-thirds of the states have implemented Physician (or Medical) Orders for Life-Sustaining Treatment (POLST/MOLST) programs despite the absence of compelling evidence that they actually improve patient outcomes, Dr Halpern wrote. There is even less evidence, he says, to support private end-of-life initiatives such as the widely used Five Wishes advance directive of the Aging with Dignity organization.

In July 2015, CMS announced plans to reimburse healthcare providers for engaging patients in end-of-life care planning discussions. This announcement, and the introduction of several bills in Congress over the past 3 years concerning end-of-life planning, signals a new federal willingness to intervene in a historically divisive healthcare arena, Dr Halpern writes. Unfortunately, the efforts made in both the public and private sectors to improve end-of-life care are not evidence based, he added.

“It is the responsibility of investigators and research sponsors to identify, develop, and rigorously test interventions so that they can offer guidance as growing political and cultural tolerance increasingly permits implementation of end-of-life care programs,” he writes.

There are at least 4 key developments that are required for this to happen, says Dr. Halpern.

1. More frequent and rapid conduct of large, randomized trials and quasi-experimental studies could help determine if current and new interventions improve outcomes that are important to patients and society.
2. It is important to reach consensus about which outcome measures to use in studies designed to quantify the effects of interventions near the end of life. Isolated measures of survival and costs of care are insufficient.
3. Developing interventions that are most likely to work requires clarification of how seriously ill patients, their family members, and clinicians actually make decisions about which care plan to follow.
4. Openness to experimentation by health systems, insurers, and other risk-bearing entities is key. The impulse to act quickly has led large health systems to implement interventions that have been studied only in narrow settings. If we approached end-of-life care policies as we approach policies for drug approval, every health system and insurance company would become a potential laboratory.

“Researchers, insurers, employers, and health systems can collaborate to advance knowledge about what works best for whom and the sooner they do so, the better—before the lack of demonstrable return on investment or the need to focus on other matters dampens policymakers current enthusiasm,” Dr Halpern concludes.