INTRODUCTION
The benefits of inpatient palliative care consultation have become increasingly documented. The impact of palliative consultations, as well as targeted palliative care units (PCU) on intensive care unit (ICU) decompression and modification of site of death for critically ill patients has not been as extensively reported. The importance of appropriate palliative care consultation is highlighted by the fact that ~50% of all hospital deaths occur after failed ICU care and approximately one-third of all patients who die in the hospital, do so after spending a minimum of 10 days in the ICU during their final days of life (The SUPPORT Principal Investigators, 1995). Data from Dartmouth Atlas confirm that nearly 50% of cancer patients die in the hospital, and, in addition, state that ~25% are in the ICU in their final month of life (Goodman, et al., Reference Goodman, Fisher and Chang2010).
Although public perception views palliative medicine and critical care at opposite ends of the medical spectrum, it may be that merging these extremes is vital to optimizing patient care. In the United States, about 20% of patients who die do so in an intensive care unit (ICU) or following ICU discharge. This translates into 540,000 ICU deaths annually; comparable to the number of patients dying from cancer each year. Nearly one third of Medicare patients with advanced cancer die in hospitals, often in ICUs (Goodman, Reference Goodman, Esty and Fisher2011). For most hospitalized patients, death occurs following the withdrawal or withholding of life-sustaining treatment (Prendergast, Reference Prendergast and Luce1997). For these critically ill patients, integration of palliative care practices and principles is essential for high-quality critical care.
Transferring dying patients from the ICU to a PCU may improve the treatment of pain and suffering associated with critical illness. Undertreatment of pain remains a problem in ICUs, with 77% of patients reporting moderate-to-severe pain (Gelinas, Reference Gelinas2007). The creation of aggressive inpatient PCUs capable of caring for chronically critically ill patients allows for decompression of ICUs while ensuring delivery of treatments consistent with patient or family preferences in an environment conducive to comprehensive treatment and palliative care.
METHOD
A retrospective electronic database review of all MICU discharges (5,035 cases) from January 1, 2006 through December 31, 2009, in a 873-bed university hospital operating in the New York metropolitan area, was performed to evaluate the impact of opening a new dedicated 10-bed PCU on January 1, 2008. This PCU was created with the goal of decompressing overly crowded ICUs and providing care to patients likely to die despite aggressive treatments. The 10-bed PCU was developed to accommodate all patients in need of palliative care including those receiving mechanical ventilation, renal replacement therapy, vasoactive medications, and other life-sustaining treatments. The 10-bed unmonitored unit was created within a newly renovated space formerly used as an ICU. The admission criteria for the PCU (see Table 1) were inclusive. Patients could be receiving mechanical ventilation, intravenous infusion of vasoactive medications, dialysis, or antibiotics, and did not have to have a Do Not Resuscitate (DNR) designation upon transfer. Ventilator patients were deemed unlikely to wean from mechanical ventilation by their ICU physician. The goal of the PCU was to provide care to all patients who could benefit from palliative care regardless of treatment preferences.
The palliative care unit
MICU, medical intensive care unit; SICU, CTU, NSICU, CCU, RCU. PCU, palliative care unit; NSUH, DNR, do not resuscitate.
The newly created PCU was staffed with one nurse for each five patients. A board-certified hospice and palliative medicine medical director was in charge of the unit. A nurse manager, social worker, chaplain, hospice and palliative medicine fellow, and attending physician visited all patients daily. Ventilator patients were managed by respiratory therapy in conjunction with the hospice and palliative medicine attending physician. As the unit was not monitored, vital signs were recorded as ordered, usually only once daily, or were discontinued altogether. Standardized palliative care orders were used to facilitate care in this unit (Appendix A).
Data from a 17-bed medical intensive care unit (MICU) was reviewed with the hypothesis that the opening of the PCU would have beneficial impact on overall MICU mortality and MICU length of stay (LOS). The Wilcoxon two-sample test was used to determine whether quarterly ICU mortality rates and LOS significantly decreased over time (across quarters) (Zar, Reference Zar1984). The working (two-tailed) hypothesis was that mortality rates, as well as LOS, would decrease after the introduction of the PCU. A two-tailed exact test with α = 0.05 was used.
RESULTS
Table 2 shows the results of our retrospective electronic database review of the 5,035 MICU discharges from January 1, 2006 through December 31, 2009. Our results indicate that since the opening of the PCU on January 1, 2008, MICU crude mortality significantly decreased (21% to 16%, Wilcoxon two-sided exact test, p = 0.003). During this time, 155 MICU patients died in the PCU after it opened. There was also a significant reduction in MICU LOS (4.6 to 4.0 days, Wilcoxon two-sided exact test, p = 0.014). During the study period, the mean patient age remained unchanged, holding consistent at 68 years (Wilcoxon two-sided exact test, n.s.).
Key Characteristics Before and After the Opening of the Palliative Care Unit
1 = Medical Intensive Care Unit, 2 = Length of Stay, 3 = Acute Physiology Age and Chronic Health Evaluation
As seen in Figure 1, most of the mortality rates after the opening of the PCU in 2008 were significantly lower than the mortality rates in 2006 and 2007. Additionally, after the opening of the PCU, the highest mortality rate was in the fourth quarter of 2009, which was still significantly lower than the combined average of the rates prior to 2008. The same can be said for LOS, as LOS was significantly lower after opening the PCU. Furthermore, as seen in Figure 2, LOS has continued and remained on a downward trend after the opening of the PCU.
Medical intensive care unit mortality rate across quarters. Black solid line on Q1 of 2008 denotes the opening of the palliative care unit.
Length of stay across quarters. Black solid line on Q1 of 2008 denotes the opening of the palliative care unit.
Although Acute Physiology Age and Chronic Health Evaluation (APACHE) III scores were available, they were not statistically analyzed, as data were not available for the last two quarters of 2009. Despite missing data, and as seen in Table 2, there was a general decline in the APACHE score as time progressed.
DISCUSSION
An ICU death is often associated with aggressive interventions, burdensome treatments, and emotional isolation. The value of the palliative care consultation has been well documented, and the ability to more effectively communicate prognosis and goals of care near the end of life can lead to a more appropriate level of care, as proposed in this article. Protecting the patient from an ICU death allows more liberal visiting hours for family members, fewer burdensome procedures, less intense monitoring, and more effective pain and symptom management. Critically ill patients have been shown to have undertreated symptoms, high caregiver burden, and low patient and family satisfaction regarding their medical care (Teno et al., Reference Teno, Clarridge and Casey2004). These data have also supported the hypothesis that hospital palliative care programs are linked to more effective symptom management and improved family satisfaction with care.
Our data show that the creation of a PCU had a significant impact on reducing MICU mortality. Our dedicated 10-bed PCU is located in close proximity to all 87 ICU beds, allowing for a seamless transition of ICU patients to the PCU. Unique to the few other units of its type, our PCU frequently accepts patients receiving mechanical ventilation, along with those on vasopressor support. This may appear to be contradictory to the traditional thinking of palliative care. However, transitioning patients from an ICU to a PCU must be done in concert with patient preferences and family's wishes. The PCU 24/7 open visiting hour access and nursing expertise in end-of-life clinical care promotes an environment aimed at improved family support and patient care.
The benefits of palliative care consultation in ICUs appear to extend beyond the decreased MICU mortality and MICU LOS documented in our chart review. The quality improvement in patient care in our retrospective data review appears to be manyfold. Helping patients and their families to obtain the most appropriate location of death can provide many clinical and non-clinical benefits. It has been previously documented that the ICU is one of the places in a hospital where family members suffer most (Pochard et al., Reference Pochard, Azoulay and Chevret2001; Azoulay et al., Reference Azoulay, Pochard and Kentish–Barnes2005). The ICU is also the location where family members are most often asked to make critical decisions about their loved ones. Combine these stressful circumstances with tightly restricted ICU visiting hours, numerous subspecialty consultants, invasive procedures, and unclear prognosis, and the burden on families of terminally ill ICU patients can have lasting effects. One study reported that post-traumatic stress reactions were observed in one-third of family members 90 days after an ICU discharge or death of a relative (Azoulay et al., Reference Azoulay, Pochard and Kentish–Barnes2005). In addition, perceived incompleteness of the information received in the ICU was also independently associated with post-traumatic stress reactions. These findings reinforce previous data regarding the specific needs of family members of ICU patients. The availability of palliative care consultants who are experienced in communication and bereavement counseling coordination could serve as an instrumental component in improving patient and family satisfaction with the end-of-life decision-making process.
As previously reported, palliative care consultation has the opportunity to influence resource utilization and cost of clinical and non-clinical aspects of care (Morrison et al., Reference Morrison, Penrod and Cassel2008). It has also been previously shown, in a prospective matched-control study, that patients receiving palliative care had statistically significantly lower costs per day (Smith et al., Reference Smith, Coyne and Cassel2003; Hanson et al., Reference Hanson, Usher and Spragens2008). Although reviewing cost savings was not an initial goal of our review, the data were made available and were seemingly significant enough to include here. It is important to note that there is little literature currently available on specialized PCUs. There are very few operating in the United States, with only a few more having published data internationally. Our article hopes to underscore their potential in providing optimal patient care and to raise awareness of their potential niche in today's medical landscape.
Our results reflect a decrease in MICU mortality and MICU LOS during the creation and the implementation of a PCU. A study limitation is that APACHE III scores were not included in the analysis, even though a general decline in APACHE scores over time is evident in Table 2. The reason for excluding APACHE scores was simply lack of data: APACHE scores for the last two quarters of 2009 were not available. Given that higher scores imply a more severe disease and higher risk of death, it is indeed possible that severity of illness may partially explain the decrease in LOS, despite the fact that the age of our sample remained relatively consistent throughout the study period. It should be noted, however, that the APACHE III score ranges from 0–299. Since the score range is so wide, a change of at least five points is required to significantly impact hospital mortality rate. Our observed mean scores were within this range and therefore had no impact on observed mortality differences, which in our study were attributable to the opening of the PCU, creating a more appropriate location for terminal care of critically ill patients.
Previous data have reported shortened hospital and ICU LOS for patients with global cerebral ischemia after cardiopulmonary arrest and multiple organ system failure (Campbell & Guzman, Reference Campbell and Guzman2003). This was demonstrated after implementing a proactive palliative care approach to patients with these admitting diagnoses (Campbell & Guzman, Reference Campbell and Guzman2003). However, this type of data was not generalized across all MICU discharges as in our study. Barriers to palliative care in the ICU and criteria to identify those patients who would likely benefit from palliative interventions have both been identified in previously published literature (Nelson, Reference Nelson2006; Bradley & Brasel, Reference Bradley and Brasel2009). Limitations to our study are inherent to it being a retrospective chart review. Physician turnover and management strategies may have varied in our MICU over the study period. Prospective, randomized trials are beginning to be published in the literature at present, and address both the clinical and non-clinical influences of palliative care consultation. Only with continued research will the true impact of these services be recognized.
Appendix A
