Clarity or Confusion? Variability in Uses of “Allow Natural Death” in State POLST Forms

Dear Editors:

The importance of advance care planning is increasingly recognized, yet best practices are not well defined. Discussions about “do-not-resuscitate” (DNR) orders are particularly challenging. “Allow natural death” (AND) was proposed as an alternative to DNR in 2000. ¹ Limited research suggests that AND may be a more acceptable term than DNR and impact resuscitation decisions.^1,2 Still, critics question the clarity of its meaning.

AND terminology has gained traction in the Physician Orders for Life-Sustaining Treatment (POLST) program, established in 47 states. In December 2016, we analyzed POLST forms of states with “developing,” “regionally endorsed,” “endorsed,” or “mature” programs as designated by the National POLST website, and assessed them for use of the phrase “allow natural death.” Incidences were categorized by surrounding language, interventions specified, and clinical circumstances activating the order.

In total, 43 forms were analyzed. Three states designated as “nonconforming” were excluded; we were unable to obtain a form from one additional state. Twenty-eight states (65.1%) used “Allow Natural Death” in the title for a medical order. Of these, 24 (85.7%) coupled AND to DNR, Do Not Attempt Resuscitation, and/or No Code, whereas 4 states (14.2%) coupled AND to “Comfort Measures Only,” “Comfort Measures,” or “Comfort Care.” Among those using AND with DNR, 19 (79.1%) explicitly prohibited defibrillation in the order, whereas 5 (20.8%) did not specify. Twenty-one (87.5%) AND/DNR orders were applicable to patients both without pulse and without respirations, whereas the remaining 3 (12.5%) were applicable to patients without pulse and/or without respirations (Table 1). Among states using AND as a comfort order, three discussed “comfort” in terms of pain, wound care, dyspnea, and preferences for hospitalization and life-sustaining treatments. The remaining state addressed “comfort” with lesser specificity.

AND is establishing a foothold in the advance care planning lexicon, but is not uniform in definition or application. The divergence between its uses as a resuscitation-specific order versus a broader comfort order is concerning. Confusion could arise when patients receive care across state lines, or attempt to transfer their preferences to a state with a discordant POLST form. The potential for error may be even greater as AND becomes increasingly used outside the POLST setting, such as in verbal advance care planning conversations or non-POLST advance directive documents.

Other inconsistencies described are of uncertain significance. For example, if states do not explicitly prohibit defibrillation in AND/DNR orders, are they allowing it, or is its prohibition implicitly understood? Likewise, are there scenarios where initiating life support and resuscitation in a patient with cardiopulmonary arrest compared with a patient in respiratory arrest with a pulse, or vice versa, could yield different outcomes? The answer is debatable. We suspect these differences are carryovers inherent to the term DNR, lending support to the criticism that AND “merely replaces one problematic term with another.” ³ In our fragmented healthcare system, interpretive differences present significant risk for medical errors. Institutions and providers introducing AND into their practice should exercise caution. For all its merits, we believe that AND remains imperfect language.