Culture and Sanity at the End-of-Life

“To the well-organized mind, death is but the next great adventure.” – J.K. Rowling (via Albus Dumbledore)

In this issue of JAGS, we gain more insight into an interesting question: do people who worked in health care use health care differently when they become ill compared to persons who did not work in health care? The general hypothesis is that they must; of course, doctors and nurses realize that as you get sick, the benefits of modern medicine slowly begin to decrease and can even become harmful. However, the growing body of research seems to suggest the opposite.^1–5 In reality, populations of people have pretty similar utilization patterns regardless of whether or not they have worked in health care. In this issue, Bynum et al. cleverly link the Nurses Health Study (NHS), a cohort study with a population of over 120,000 female nurses with 42 years of follow up, to the Medicare claims data to explore utilization patterns among people with Alzheimer’s dementia and related disorders (ADRD) and heart failure (HF).⁶ In sum, their findings were consistent with prior studies, nurses and non-nurses use health care similarly.

Even though we are about to discuss how this article highlights that the culture of health care needs an upheaval, there are some things to be excited about in this paper. First, the fact that nurses and non-nurses use health care similarly means that the NHS dataset may provide a very rich source to study questions of utilization that may indeed be generalizable. The NHS is one of those beautifully conceived and executed endeavors that will continue to provide many returns on the initial investment. Prior to this analysis, a study using the NHS exploring utilization would have been criticized based on the belief that nurses are different and not generalizable to the larger population. At least for health care utilization questions around ADRS and HF, this paper shows that this is not true. Researchers should eagerly begin answering important questions about utilization that claims data alone are not able to answer. Second, some of the trends around use of care at the end of life are encouraging. The facts that ~55% of people with ADRD and ~45% of people with HF are using hospice in the last 30 days of life are evidence of improved use of hospice.⁷ Given that hospice is associated with quality,⁸ these increasing numbers are encouraging.

Yet the differences Bynum et al. found between the nurse cohort and non-nurse cohort for the outcomes invasive interventions, death in the hospital, ICU stay during terminal hospitalization, and hospice use only ranged from 2-4%. Unquestionably, the differences in EOL care patterns are statistically significant and clinically meaningful but they are unremarkable. The study nurse population is composed of highly educated women, with high health literacy and experience with illness. In our opinion, the differences observed are modest and not of great importance. In the use of acute care hospitalizations and post-acute care at the end of life in the last 6 months of life, there were no differences in the populations, similar to previous findings of physician utilization patterns in the last months of life.⁵ Beyond education in general, bedside nurses arguably spend more time with patients and families facing serious illness than any other provider in health care. If the sum effect of this education and experience results in only modest differences in outcomes at the end of life, how can we expect any behavioral or educational intervention to have a consequential impact for end of life care.

This leads us to the next most obvious question, if education and experience with serious illness lead to only small differences in end of life outcomes, what would move the needle? If we want to get serious about addressing the issue of high cost, low value care at the end of life (e.g. terminal hospitalizations for patients with chronic progressive illness, invasive interventions at the end of life that are unlikely to improve survival or quality of life), we need a cultural shift and a system overhaul. Over twenty-five years ago the Robert Wood Johnson foundation funded the Project on Death in America as part of a national effort to change culture and the experience of dying in America. And over the ensuing decades, we have seen a change in culture. This includes the growth of inpatient, outpatient, and community based palliative care programs. Greater percentages of persons facing the end of life utilize hospice care. There is also some level of acceptance of our own mortality seen in mainstream society through the popularity of the Conversation Project, Death cafes, and Atul Gwande’s book Being Mortal spending over 68 weeks on the bestseller list. Yet when only about half the people dying from ADRD and HF are accessing hospice care, we have more work to do.

Death is apolitical and a way towards a solution must also be apolitical. Our leaders and our society must come together and have a real dialogue about what low value care really means. Medicine pushes the boundaries of what we can do, and we can lose sight of what we should do. Previous attempts to engage in a national discussion related to end of life care have resulted in saber rattling and cries of “rationing of care”. One such example is when advance care planning was added to the Affordable Care Act legislation, giving birth to “death panels”, a term that persists in our lexicon. The argument resurfaced when the Obama administration added advance care planning to the annual Medicare visit. George W. Bush had already included advance care planning as part of the annual welcome to Medicare visit. This generated a flurry of politically fueled backlash that may have further derailed a sober national dialogue about when we should say no to high cost, low value care. Bynum, et al. included a composite outcome of invasive interventions at the end of life which included feeding tubes, dialysis, and mechanical ventilation. Through the Choosing Wisely campaign there is an active recommendation against the use of feeding tubes and artificial nutrition for patients with dementia. Yet we can conclude that if this was included in the composite outcome, there remains a measurable number of patients receiving this invasive intervention that demonstrates no benefit and has evidence of harm.

At a policy level what can be done? We fully support the Palliative Care and Hospice Education Act and acknowledge the work of national organizations such as Center to Advance Palliative Care and the advocacy of the American Academy of Hospice and Palliative Medicine and Hospice and Palliative Nursing Association. However, work is also needed in health care reform. We need to look closely at some of the perverse incentives that drive care. For instance, with the implementation of penalties for 30-day readmission, there has been an increase in the use of Subacute Nursing Facilities for rehabilitation. A study of Medicare decedents demonstrated that older adults who use these SNFs are more likely to die in a nursing home facility, often while receiving rehab care, and less likely to use hospice care, leading to the phenomenon of “rehabbing to death”.^9,10 We need to put policies in place that ensure high-quality, goals of care conversations take place, not just policies that incentivize unwanted and low value rehabilitation care.

Our society now faces the challenge of more people living longer with more interventions available than ever before. This is not a problem we can study in the history books. If informed doctors and nurses do not experience different outcomes, then it appears the real work to improve care at the end-of-life is one of counter-culture work. We have a philosophy that technology can fix everything, and that death is failure. To really see the improvement in end of life care, we need to continue to address the culture within which we all exist. When the culture is insane, sanity is working counter to that culture.