Emergency Department-Based Palliative Interventions: A Novel Approach to Palliative Care in the Emergency Department

Introduction

Up to 80% of cancer patients visit the emergency department (ED) during the last six months of life, with nearly two-thirds presenting more than once.^1–3 The ED is a critical point of care for patients with active cancer; oftentimes, these visits are for care that is palliative in nature, including visits for pain, dyspnea, nausea, vomiting, confusion, and anxiety.^1,4–6 Many of these visits are potentially avoidable if patients with active cancer had earlier access to outpatient palliative care. ⁷

Despite recommendations for early implementation of palliative care for patients with cancer by the American Society of Clinical Oncologists (ASCO) and the American Academy of Hospice and Palliative Medicine (AAHPM),^8,9 and evidence that palliative care improves patient and family member quality of life and decreases health care costs and hospital resource utilization, long delays between cancer diagnosis and palliative consultation are common.^10–18 On average, oncology patients at cancer centers in the United States in 2009 died within one week of inpatient palliative consultation. ¹⁹

Because patients with cancer frequently use the ED, and long delays exist between cancer diagnosis and palliative care consultation, the ED may be a key location to expand the reach of palliative care for patients with cancer. However, due to the fast-paced nature of emergency medicine practice, palliative care is not typically part of standard ED care. Previous emergency medicine-palliative care research has largely focused on the development of screening tools to help expedite palliative care consultation rather than address the needs directly.^20–22 In addition, the incidence of unmet palliative care needs in ED patients with cancer has not been well characterized.^2,20 Finally, there is limited research on ED-based palliative interventions (EPIs) focused on identifying and addressing palliative needs. The existing studies are limited by their retrospective design.^23,24 Given the clear benefits of early palliative engagement in oncology care, as well as the regularity with which cancer patients utilize the ED, we sought to enhance patient access to palliative services by delivering ED resources to address unmet palliative needs.

The primary aims of this study are to quantify the unmet palliative care needs of patients with active cancer presenting to an urban academic ED, and to assess the feasibility of using EPIs to address those needs. In addition, a secondary aim is to assess the impact of implementing EPIs on ED operational metrics such as length of stay (LOS) and return visits to the ED.

Methods

Participants and procedures

All adult (>17 years old) ED patients were screened for active cancer and inclusion in this intervention during the initial bedside nursing ED assessment. This screening was a routine part of care before the EPI project was initiated. For the purpose of this study, a patient with active cancer was defined as a patient who (1) was undergoing or in the past 12 months had undergone cancer-directed therapy (radiation/chemotherapy), (2) was known to or found by care providers in the ED to have metastatic disease, or (3) reported directly that he or she had symptoms related to known cancer (pain, fatigue, nausea, etc.).

The screen for palliative and end-of-life care needs in the ED (SPEED) Survey is the first published tool that was derived as an ED-specific palliative care screening tool.^20,27,28 SPEED focuses on five key areas of palliative care: pain management, home care, medication management, psychological support, and goals of care. It has since been adapted into a rapid five-question needs assessment named 5-SPEED. ²⁷ In >1500 cancer patients hospitalized through the ED, increasing 5-SPEED score was associated with increased likelihood of inpatient palliative care consultation. ²⁷ A custom 5-SPEED EHR screen was built for the EPI project and automatically inserted into the normal template of the bedside-nursing triage form. Both patients and nurses were able to opt out of screening if the clinical context did not allow for screening or the patient was not interested in answering the questions. Each domain of the 5-SPEED is scored on a 10-point Likert scale. Thresholds for screening “positive” in each domain were previously established by ED and palliative care experts.^27,29

For those patients who screened positive by 5-SPEED, an automatic EHR-based consult was placed to the appropriate ED resource service for each positive screen (Fig. 1). Therefore, if a patient had multiple palliative care needs, they would be eligible for multiple EPIs. Pharmacy consultations were initiated for patients with pain and medication reconciliation needs. Social work was consulted for those with unmet home care needs. The TCNs were consulted for those with unmet goals of care needs and an appointment referral was placed to supportive oncology for those with emotional distress.

OPEN IN VIEWER

FIG. 1.

The Emergency-Department-based Palliative Intervention Algorithm (ED, emergency department; 5-SPEED, Screen for Palliative and End-of-Life Care Needs in the Emergency Department-5; GEDI, geriatric emergency department innovations).

Results

For the seven-month study period, 1278 patients with active cancer presented to the study site ED. Of these, 817 completed the 5-SPEED screen. Demographic characteristics such as age, gender, ethnicity, and insurance coverage were largely similar between cancer patients not screened with 5-SPEED, 5-SPEED negative patients, and 5-SPEED positive patients. Patients with an ESI triage level of 1 and patients with public insurance (Medicare and Medicaid) were less likely to be screened. 5-SPEED negative patients had a slightly higher likelihood of being discharged from the ED (30.9%) compared with patients with one or more palliative care needs identified by 5-SPEED (p = 0.004) (Table 1). Palliative care needs were identified in 422 (51.7%) by the 5-SPEED screen. Many patients had more than one unmet palliative care need: 163 (19.6%) had one unmet need, 107 (13.1%) patients had two unmet needs, 85 (10.4%) had three unmet needs, 27 (3.3%) had four unmet needs, and 35 (4.3%) had five unmet needs. Among cancer patients screened, 289 (35.4%) had an unmet pain need, 130 (15.9%) had home care needs, 131 (16%) had medication challenges, 235 (28.8%) had emotional distress, and 130 (15.9%) had issues with their goals of care (Table 2).

Of the 422 patients with an unmet palliative care need, 167 (39.6%) received an EPI during their ED stay. Of these patients in need, 75 (26.0%—pain) and 36 (27.5%—medication reconciliation) received a pharmacy consultation, 32 (24.6%) had a social work consultation to address home care needs, and 32 (26.2%) received a TCN consultation to address issues with goals of care (Table 2). Of the 167 patients receiving an intervention, 97 (58.1%) received one intervention, 54 (32.3%) received two, 15 (9.0%) received three, and 1 (0.5%) received four. There was no statistically significant difference in LOS between these groups. The initial plan for an automated referral to supportive oncology through the in-ED Physician Referral Service (PRS) staff was unsuccessful due to high PRS staff turnover, and the increased complexity with performing this referral compared with usual referrals in the ED. There was no statistical difference in ED LOS and 10-day ED revisit rates between cancer patients not screened, 5-SPEED negative patients, 5-SPEED positive patients who did not receive an intervention, and 5-SPEED positive patients who received one or more interventions (Table 3).

Discussion

This article describes a novel care pathway, the EPI project, which identified cancer patients with unmet palliative care needs and addressed those needs through existing ED resources at an urban academic medical center. The care pathway utilized EHR-embedded screening and automated EHR-triggered consultation that delivered palliative services to ED cancer patients while not affecting ED LOS.

This is the first study to our knowledge to detail the rate of palliative care needs for patients with cancer in the ED. We found that >50% of active cancer patients presenting to the ED during the study period had one or more unmet palliative care needs. In a recent systematic review, Wong et al. were unable to determine the incidence of patients with palliative care needs in the ED due to limitations in available data. ² Our study shows not only the feasibility of screening in the ED for palliative care needs but also demonstrates high prevalence of unmet care needs in the population sampled. The data from this study may also serve to further support recent calls for prospective interdisciplinary study of ED-based palliative services given the high prevalence of unmet need that was found in this sample.^31,32

Using existing ED resources, we were able to deliver specialized care to nearly 40% of the cancer patients with unmet palliative care needs. The inability to address a higher proportion of patients in need was multifactorial. In part, this rate of EPI delivery was due to staffing; as noted earlier, many of the ED consultation resources are unavailable during evening and overnight hours. During weekday, daytime hours, the workload already placed on these ED services is extensive and limited their ability to evaluate every cancer patient with a palliative need. Furthermore, it was also a central goal of developing the pathway such that it did not affect ED operational metrics negatively, namely ED LOS. As such, normal care delivery was not delayed for EPI consultation, and many patients with unmet needs may have been already admitted or discharged before an EPI team member could address their palliative care needs. Furthermore, the intervention rate calculation included patients with emotional distress, and although the initial PRS-driven intervention was unsuccessful, the chart review approach captured all patients with emotional distress after June 2017 and forwarded their information to supportive oncology. Unfortunately, we were unable to capture this data in our EMR data query. As such, the overall intervention rates appear lower than the true proportion of patients receiving an EPI for their palliative care need. Finally, while completing chart reviews we found consults often contained improper documentation resulting in an inability of our data query to capture these EPI consults.

Although we had modest success in addressing the needs of those for whom screening was completed, only 64% of cancer patients presenting during the study period received a 5-SPEED screen. In discussing the screening process with nurses, there are several reasons for this suboptimal rate of screening. First, some patients are too ill, too distressed, or unwilling to participate in the 5-SPEED screen. As seen in Table 1, those patients categorized as ESI level one or two, those thought to need the most expedient emergency care in the ED, were the vast majority of those cancer patients not screened. This finding suggests the acuity of presentation may have been a limiting factor in screening rates within this group in comparison with those with less urgent ESI levels. Although the 5-SPEED screen was rarely used in these critical situations, emergency physicians frequently perform a brief informal evaluation of goals of care and treatment preferences for critically ill patients.

In addition, discussing topics such as emotional distress and goals of care are difficult issues to address and quantify in a bedside screen, especially one taking place in the ED with nurses who have just met the patient and their families for the first time. Nurses commented that when they first attempted screening, they encountered tearful and anxious patients when discussing emotional distress and goals of care and felt uncomfortable asking these questions. These negative experiences made the nurses reticent to perform the screening with future patients for fear of being in similarly emotion-laden conversations. Despite frequent nursing education, it was difficult to reach a high rate of successful screening due to these challenging issues.

To our knowledge, this is the first published work attempting to address palliative care needs with ED-specific interventions.^23,24 To date, emergency medicine/palliative care research has largely focused on the development of screening tools to help expedite palliative care consultation.^20,21 A recent study by Gruzden and colleagues in 2016 showed that ED-initiated palliative care consultation improved quality of life in patients with advanced cancer and had no detrimental effect on one year survival. The finding that using ED-based resources did not negatively impact patient LOS may support other programs in garnering administrative support for similar pathways. The EPI care pathway ultimately seeks to extend the reach of palliative care services to communities that currently do not have sufficient access, including minority populations, smaller communities, and rural centers by utilizing the ED as both a novel arena for the delivery of multidimensional palliative interventions and a palliative care safety net. We note that EPI may require the addition of ED resources, which smaller EDs may not currently have; however, access to these services are all recommended in the geriatric ED guidelines. ³³ As EDs work to meet the GED guidelines and geriatric ED accreditation, access to these services for patients with palliative care needs should be considered. ³⁴

This initial description of a unique EPI project has some limitations. The outcomes reported are process and operational outcomes, rather than clinical and patient-centered outcomes, which were not measured. Future work will focus on quantifying the impact of the intervention on patient's quality of life, reassessment of the 5-SPEED palliative care needs, the rates of subsequent palliative care involvement, patient survival, ongoing functional status, and days spent in hospital and ICU. Only two-thirds of patients with active cancer were screened. If nurses were selectively screening patients in whom they anticipated a need, the prevalence of need estimate may be falsely elevated. However, even if all unscreened patients were 5-SPEED negative, the unmet need in this sample would still be 33% (422/1278).

This was a single site study and external generalizability is limited as many resources may not be readily available at smaller, nonacademic institutions. Specifically, the TCNs, who have received specialized training through the Education in Palliative and End-of-Life Care for Emergency Medicine (EPEC™-EM) course, are a unique resource in the NMH ED. However, the EPEC-EM course is widely available and nurses and ED staff could be trained to conduct these challenging conversations with cancer patients. The supportive oncology department is also a highly specialized resource dedicated to addressing the well-being of cancer patients and their families, and may not be available at other hospitals. Nevertheless, many oncology departments have social workers who provide similar support for cancer patients and many psychologists who help to manage anxiety and depression within the cancer populations, making this intervention potentially feasible at a wide variety of health care systems. In addition, supportive oncology follow-up was unsuccessful as initially planned; therefore, in June 2017 the EPI Team personally forwarded patients' contact information for supportive oncology follow-up. The initial PRS process had a low rate of success with only two total patients having their information forwarded in the first five months of the study period. After the change in process all patients with emotional distress had their information forwarded; however, this is unlikely to be a sustainable long-term solution.

Conclusions

The EPI pathway provides a novel approach to delivering palliative care services to oncology patients by utilizing the unconventional setting of the ED. These data demonstrate that the need for such services is great with a majority of screened patients with active cancer reporting unmet needs. This EHR-enabled solution streamlined nursing screening and auto-triggered consults thereby requiring no manual consult entry or extra work from ED physicians. By leveraging existing ED resources, we not only effectively screened for palliative needs, but also enhanced patients' access to palliative services without increasing ED LOS. The EPI pathway holds promise for extending the reach of palliative care services in many diverse ED settings.