Field Notes

Getting right with guidelines

Near the beginning of our three-year project to revise and update the 1987 Guidelines on the Termination of Life-Sustaining Treatment and the Care of the Dying—a.k.a. "The Hastings Center Guidelines"—one of our working group members, a physician, described our audience as "good docs in bad systems." I don't think this physician was referring to "systems" as institutions, or that nebulous entity, "the health care system." All good docs (and good nurses, and so on) work inside systems, all the time. A patient care team is a system. A process for conducting ethics consultations, or for handing off responsibility for patient care at the end of a shift, is a system. Theorists and scientists alike remind us that a system will deliver the results it has been designed to deliver. And if a system consistently delivers bad results—or inconsistently delivers good results—it's a bad system.

I was reminded of that physician's remark when I read an essay published last August 8th in The Lancet, entitled "Reality Check for Checklists." The essay notes how quickly the complex patient safety initiative developed by Peter J. Pronovost to reduce line infections in the intensive care unit had been reduced to the myth of the "simple checklist." According to it, the checklist is the perfect system: just pull it out of your pocket, in any setting, follow it step by step, and, presto, no infections. There is a grain of truth in this myth, as Pronovost has the evidence to support the intervention. However, the authors of this essay—who include Pronovost himself, as well as medical sociologist Charles L. Bosk—point out that swallowing the myth pushes aside the real problem of how human beings, at every level in a hierarchy, adapt to, or reject, the use of tools such as checklists. According to the authors, understanding why this is so is a bigger challenge than getting the items on the checklist right.

As I'm immersed in final revisions to our new Guidelines, it is somewhat unnerving to note that Bosk and Pronovost extend their cautions about "checklists" to "guidelines," too: neglect the human factors in the system, and your guidelines will stay on the shelf. (Right next to the disaster planning protocols.) Checklists and guidelines alone do not create good systems, and following them to the letter does not ensure that readers will think critically about the systems in which they are already embedded. Efforts to improve the quality of care near the end of life cannot be limited to encouraging individual professionals to use our Guidelines at the bedside without regard for those bad and so-so systems that fail to protect dying patients and their interests. We must also aim to persuade health care leaders, inside and outside of clinical settings, that good end-of-life care is part of good health care and should be sustained by policy, education, research, investments, and institutional culture. The care of persons near the end of life is an opportunity—and an obligation—for quality improvement in health care, and for the design of organizational systems that can support and reward good practice by those Guidelines-reading good docs.