Treatment escalation in the intensive care unit among patients with preexisting treatment limitations: best-laid plans gone awry?

Abstract

In this issue of JAMA Internal Medicine, Hart and colleagues 1 provide, to our knowledge, the first known epidemiologic data on treatment patterns among patients admitted to the intensive care unit (ICU) with preexisting treatment limitations. Treatment limitations areexplicit statementsby thepatient,documented in themedical record, refusing certain treatments, as is their right according to the Patient Self-determination Act of 1990.2 In a hospital context, treatment limitationsmostoften includedonot resuscitate (DNR) or do not intubate orders. UsingProject IMPACT, anonrepresentative sampleof ICUs across the United States with detailed clinical data on all ICU admissions, Hart et al1 document that 5% of all ICU admissionswere amongpatientswith preexisting treatment limitations. Most treatment limitations were DNR orders. Many of thesepatients experiencedanescalationof treatment, including some form of life-supporting treatment, while in the ICU, and 1 in 4 of these patients received cardiopulmonary resuscitation (CPR), seemingly in violation of their wishes. Overall, 65% survived to discharge, but 35% died, corresponding to theobservation that 1 in 6patients admitted to the ICUwith a preexisting DNR order had received CPR. The authors further document substantial between-unit variation inmanagementof ICUpatientswithpreexisting treatment limitations. Although the 95%CIs arewide and unmeasured confounding cannot be ruled out, this variation nonetheless leads us to ask: What are the “right” rates? First,what is the right proportionof admissions to the ICU that should be made up of patients with preexisting treatment limitations?There isa lotof controversyabout thisamong clinicians, but guidelines promulgated by the Society of Critical Care Medicine3 regarding ICU admission criteria indicate that apreexistingDNRorder is not a criterion for refusal of ICU admission. Patients with DNR orders frequently desire curativemeasures shortofCPR, including treatmentsavailableonly in an ICU setting, such as hemodynamic support and noninvasivemechanical ventilation,4 andpatientsmaybenefit from such treatment. If there is a potential for meaningful clinical benefit from an escalation of treatment (suggested by a 65% survival rate) andpatientswould accept treatment escalation short of CPR, then the right rate of ICU admission will be directly proportional to the prevalence of preexisting treatment limitations in the critically ill population. The overall prevalence of treatment-limiting advance directives among Medicare decedents is 39%, and this varies significantly byUS region.5 Given a lack of information about the prevalence of preexisting treatment limitations inpatients admitted to each of the study hospitals, we cannot know whether the observed rates are too high, too low, or just right. But what about CPR, which most patients with DNR orders say theywish to forgo?What is the right rateofCPRamong ICUpatientswithpreexisting treatment limitations?Receiptof CPR in this population requires a reversal (ie, revocation) or reflects a violation (ie, ignoring) of the patient’s DNRorder. A reversalmight occur in 2ways: first, the patient and/or their surrogatemightreversetheDNRorderafterreasonedconversations with themedical care teambased on expected treatment benefit and thepatient’sunderlyinggoals for treatment.This could be because the patient’s treatment preference changed. Studies show that patients’ treatment preferences are unstable. In one study, competentpatientswillingly acceptedburdensome Related article page 1019 Research Original Investigation ICU Care and Preexisting Limits on Sustaining Life