Changes in Clinical Course Before and After Do-Not-Resuscitate Order Placement in a Pediatric Intensive Care Unit Setting

Introduction

The time surrounding end-of-life care can be extremely challenging for those in the decision-making role, including care providers, patients, and their families. Many patients at the end of life require intensive care with uncertain outcomes, taking a large emotional toll on patients and families. ¹ Adult data have shown that early do-not-resuscitate (DNR) order placement at admission can improve quality of life, decrease invasive care utilization, and reduce cost in patients at the end of life.^2–4

Although the relationship between clinical course and DNR status is well described in the adult intensive care unit (ICU) population, ⁵ this relationship is not well studied in the pediatric intensive care unit (PICU) setting. ⁶ To our knowledge, no studies specifically address the timing of a DNR order in the pediatric ICU and the subsequent changes in a child's clinical course.

Studying the direct relationship between DNR status and clinical course in the acute care setting at the end of life can help medical providers better understand the effects of a DNR order on patient's clinical course and potentially guide treatment toward care that holistically incorporates the patient's and family's wishes. Our primary objective was to describe the relationship between DNR order placement and the clinical course of children who died in the PICU of a large, tertiary academic hospital for an eight-year period, from June 2008 to 2016.

Methods

Results

One-hundred sixty-one children died within the time period studied. Forty-eight percent were male with a median (IQR) age of 3 years (0–12). Eighteen percent (29/161) of patients were classified as early DNR, 33% (53/161) as late DNR, and 49% (79/161) as no DNR. The demographic and clinical characteristics of our cohort at the time of admission are given in Table 1. There were no significant differences between the three groups with respect to age, gender, prior ICU admissions, admission diagnoses, and severity of illness at admission.

The clinical course of the patients is shown in Figures 1 and 2. Patients with early DNR order placement had longer LOS but decreased number of procedures, consults, and imaging studies done per day when compared with the late and no DNR groups (p < 0.001) (Fig. 2).

OPEN IN VIEWER

FIG. 1.

PICU length of stay in days by DNR group, p-value <0.001. DNR, do-not-resuscitate; PICU, pediatric intensive care unit.

OPEN IN VIEWER

FIG. 2.

Number of procedures per day, consults per day, and imaging studies per day by DNR group, p-value <0.001. *Y-axes have been truncated.

To account for the possibility that a longer PICU LOS increases the chance that a patient will have an early DNR placed, we performed a sensitivity analysis and only included children whose LOS was >14 days and ≤60 days (Table 2). In these 25 children, there was no significant difference in patient characteristics between the early DNR and late DNR groups (Table 2). Both groups had a clear decrease in the number of invasive procedures, imaging studies, and consults performed after DNR order placement (Table 2).

Discussion

Our findings represent the first description of the timing of DNR order placement and the clinical course in a cohort of critically ill children. In our patient cohort, there was a clear decrease in the number of invasive procedures, imaging studies, and consults performed per day in the early DNR group and overall once a DNR order was placed.

These results are corroborated in the adult literature in which DNR placement is associated with decreased aggressive care utilization^8–10 and an early DNR, defined as DNR order placed within 24 hours of final admission, is associated with a decrease in aggressive care interventions. ⁴ For our patients, this was similarly true despite differences in the patient population and the definition of early DNR, suggesting that early DNR is, indeed, associated with fewer aggressive interventions.

Although there was a significant decrease in invasive procedures, imaging studies, and consults performed once a DNR order was placed in our study cohort, the actual decrease (from 1.73 to 1.53, for example) may not represent a change in the perceived experience of a critically ill child and/or his or her family. Further studies should be done to examine the experience from the patient/family perspective.

In the pediatric world, many clinicians agree that DNR discussions happen later in a patient's hospital course than is best for the patient. ¹¹ Many pediatric intensivists feel that they are inadequately trained to discuss DNR order implementation and that they are ill equipped to face parental expectations and limitations to end-of-life care. ¹¹ Garros et al. reported that there is frequently a discrepancy between physician and parent motives for DNR placement, with physicians' motives centering on “lack of benefit from further therapy” and parents focusing on quality of life and parental “perception of their child's pain.” ¹² DNR discussions should happen within the context of larger discussions regarding goals of care, allowing families to pursue treatment measures that are most consistent with their beliefs and their child's needs. Specialty palliative care teams, working in conjunction with primary intensive care teams, can help patients and families verbalize those beliefs and their desired goals of care.

One unexpected finding in our data is that early DNR placement was associated with increased PICU LOS. In a study by Burns et al., the authors also found an association between shorter final PICU stays and lower utilization of DNR placement and hypothesized that a DNR order becomes less relevant in a rapidly decompensating patient. ¹³ We propose two alternative potential explanations for our findings. First, given that all three groups in our study had similar predictors of mortality based on severity of illness scoring, early DNR placement may allow for less aggressive care and limit futile interventions while not ultimately affecting the natural course of a disease process. The second explanation is that increased length of hospitalization may allow additional time for providers both to help improve patient/family medical understanding and prognostic awareness and to hear more clearly patient/family values that inform goals of care including wishes around resuscitative efforts.

Our study was not without limitations. A retrospective single-center study with a small sample size and no defined pediatric palliative care team may mean that our results are not generalizable and limit our ability to detect other potential confounders, including the occurrence and content of family meetings.

This study is also the first pediatric study to compare the clinical course of patients with early verses late DNR order placement. There are several important implications. Critical care providers should consider initiating DNR discussions earlier in the course of hospital stay in critically ill children. This may help improve the emotional and physical well-being of patients and families during the acutely stressful days of a terminal admission. These critically ill children deserve further research into changing goals of care, clinical course, and quality of life, with and without specialty palliative care consultation.