Does code status clarification for elderly patients being admitted from the emergency department make a difference?

Approximately half of all hospital admissions in the United States are initiated in the emergency department (ED), and this proportion is higher for elderly patients.1 Admission from the ED is common during the final stages of chronic, complex illnesses for the elderly, cancer patients, and those near the end of life.2, 3 A crucial component of caring for these patients is establishing goals of care (GOC), including clarification of patient code status preferences.4 GOC and code status conversations are often deferred in the ED, usually under the assumption the patient wants "everything done," and GOC are frequently never subsequently addressed during hospitalization.5, 6 Early GOC and code status conversations in appropriate patients are associated with a higher likelihood of honoring end-of-life wishes and less anxiety among surviving family members and result in less aggressive care at the end of life.7 Despite this, ED physicians often avoid these difficult conversations for multiple reasons including time constraints and lack of formal training.8 There is limited literature regarding preparing and encouraging ED physicians to conduct code status conversations with appropriate patients.9 As part of a system quality improvement initiative, we piloted an intervention to increase code status conversations for ED patients aged ≥75 years being admitted to the hospital. The goals of this study were to determine whether this intervention would increase documentation of code status discussions in the ED and whether this would be associated with a change in the proportion of such patients admitted under do not attempt resuscitation (DNAR) versus full code status. This before-and-after cohort study was designated exempt research by the local institutional review board. The setting was a suburban U.S. ED with an annual census of approximately 30,000 visits and 5000 hospital admissions through the ED per year. ED clinicians (physicians and nonphysician practitioners) were trained in code status discussion and documentation through an interactive session led by palliative care physicians. The 30-min session focused on code status options, documentation, and how to initiate code status discussions. ED clinicians were asked to discuss code status with any patient (or their surrogate) aged ≥75 being admitted to the hospital, document the discussion in the electronic medical record (EMR), and enter the appropriate code status order. Reminder flyers were put in the ED and email reminders were sent out weekly. An EMR autotext was introduced to help with efficient documentation of code status or GOC discussions. The 6-week pilot initiative lasted from January 3 to February 14, 2023, when email reminders were discontinued. This study includes all patients aged ≥75 admitted to the hospital from the ED, 6 weeks before and 6 weeks during the pilot initiative. After completion of the 6-week intervention period, an EMR search identified all patients meeting inclusion criteria between November 7 and December 18, 2022 (preintervention period), and between January 3 and February 14, 2023 (intervention period). A trained researcher extracted basic demographics (age, sex, race, preferred language, insurance), mortality, hospital length of stay (LOS), ventilator orders, code status orders, admission intensive care unit (ICU) versus ward status, and whether or not a code discussion was documented in the EMR. Any clarification of code status in the ED clinician's note was considered documentation of a discussion, but isolated code status orders were not. The primary outcome was the proportion of patients who had a code status discussion documented in the intervention group versus control group. The main secondary outcome was the proportion of patients with full code versus DNAR ordered on hospital admission in patients during the intervention versus control time periods. In our hospital system, code status options include: (1) full code; (2) DNAR comprehensive care, meaning the patient wants all lifesaving care, up until the point of cardiac arrest, at which point resuscitation will cease; (3) DNAR/do not intubate (DNI) comprehensive care, meaning that the patient wants all lifesaving care, up until the point of cardiorespiratory arrest or need for intubation; or (4) DNAR comfort measures, meaning the patient only desires treatments to alleviate suffering. Other predefined secondary outcomes were hospital LOS, ICU admission rate, mechanical ventilation, hospital mortality, and discharge disposition (home vs. rehabilitation or nursing facility). For binary measurements, simple comparative statistics were analyzed using a two-tailed t-test and are reported with p-values. Continuous variables (age, LOS) are reported as medians with interquartile range (IQR) and were compared with a two-tailed Mann-Whitney U-test. The EMR search identified 187 patients aged ≥75 admitted through the ED in the preintervention period and 188 during the intervention. All baseline demographics were similar between the two cohorts and are displayed in Table 1. Documented code status discussions were more common in the intervention period, 53.2% of cases versus 3.2% (p < 0.001). The intervention period was also associated with a higher proportion of patients admitted as DNAR, 39% versus 23% (p < 0.001). Code status changes between admission and discharge were similar in the two cohorts (21% preintervention, 18% during the intervention). There were no statistically significant differences between cohorts in any secondary clinical outcomes (Table 1). We found that a simple 30-min teaching intervention with weekly email reminders was associated with significant increases in documented code status discussions and DNAR code status for patients aged ≥75 being admitted to the hospital from the ED. While this observational study cannot prove causation, it implies that routine ED clinician-initiated code status discussions, in this patient population, will identify a significant number of patients who wish to have a less aggressive code status. Early GOC conversations have been associated with favorable resource utilization, quality of life, and clinical outcomes in patients at the end of life,7 but emergency clinicians do not routinely have time to complete comprehensive GOC conversations for patients in whom it may be most beneficial. We chose code status conversations and clarification as our target and found that clinicians were willing to have and document a code status discussion more than 50% of target patients after a brief educational intervention. Previous research on code status discussions in the ED has been limited. A 2019 systematic review and meta-analysis by Becker et al.10 found 15 randomized trials of communication interventions to discuss code status, but none were performed in the ED. Interestingly, in meta-analysis, the interventions were associated with a 15% absolute increase in patients who did not want cardiopulmonary resuscitation, almost identical to our findings. In an ED based study, Miller and colleagues11 explored using the EMR system to mandate code status orders at the time of entering a bed request for hospital admission. The intervention was associated with increased likelihood of a DNAR code status in the ED (5.3% vs. 0.4% among all admitted patients, p ≤ 0.001), with no difference in mortality, LOS, or ICU admission. These results are in accordance with our findings. While our study population had a much higher overall DNAR code status rate, this is to be expected in an elderly population compared to the "all comers" population in the study by Miller et al. One limitation of the study by Miller et al. is that code status discussions were not evaluated, and there is evidence that physicians are willing to determine code status, including DNAR code status, without discussing it with patients.5 Our study has several limitations. This was an observational study, so we cannot determine whether the associations between the intervention, code status discussion documentation, and the increase in patients admitted as DNAR status was causative, but our results are consistent with those in other literature.10, 11 Our intervention was associated with a marked increase in documented code status discussions, but still only approximately half of eligible patients had a code status discussion documented. Opportunities may have been missed due to clinicians being uncomfortable discussing code status, not realizing their patients were ≥ 75, or simply forgetting to have or document the conversation on a busy shift. We were unable to reliably obtain some potential confounders that may have had an impact on secondary patient outcomes such as mortality and LOS. Although the obtained baseline demographics were nearly identical in the two cohorts, it is possible that there were unbalanced confounders that we were unable to account for. This was a pilot study and was underpowered to detect differences in most clinical outcomes. We did not evaluate the duration of effect of the intervention. It is unknown whether code status discussions and documentation fell after the weekly emails stopped, how frequently a "refresher" training session should be held, or what the ongoing impact of the intervention is likely to be. This study was carried out in a single, mid-volume ED in the United States, and the results may not be generalizable to other practice environments or patient populations. In conclusion, a brief educational training session paired with weekly email reminders was associated with a 50% absolute increase in documentation of code status discussions for patients aged ≥75 being admitted to the hospital through the ED and a 16% absolute increase in DNAR status. Further study is warranted to determine the impact on clinical outcomes, resource utilization, and quality of life. Carl Pafford—Concept and design, acquisition of the data, analysis and interpretation of the data, critical revision of the manuscript for important intellectual content. Amber R. Comer—Concept and design, critical revision of the manuscript for important intellectual content. Daniel Slubowski—Acquisition of the data, critical revision of the manuscript for important intellectual content. Laurae Rettig—Acquisition of the data, critical revision of the manuscript for important intellectual content. Benton R. Hunter—Concept and design, analysis and interpretation of the data, drafting of the manuscript, statistical expertise. The authors declare no conflicts of interest.