The present work demonstrates that when patients with terminal illnesses are evaluated by an early CPE in the emergency room, compared to later evaluations during hospitalization, they have a shorter hospital stay but return more often after discharge.
Analyze the impact of inserting an ECP in the ED as a strategy to optimize the flow in health services, adding value to care, ensuring that the role of the emergency is not focused on prolonging life (healing, saving) but also on guaranteeing the quality of care (caring, welcoming) is challenging. It is essential to consider the nature of the intervention, the study population, and the outcomes studied.
In a recent systematic analysis, Wilson JG et al. observed that despite increasing the impact of implementing palliative care in the ED is variable, but it can improve patients' quality of life and does not change survival (11). One of these factors is the nature of the intervention, and in most studies, it was represented only by establishing a flow between the ED and hospitals dedicated to palliative care. Other studies developed more elaborate strategies, such as consultation with the institution's Palliative Care Service, but few had a more elaborate intervention, such as creating a PCT dedicated to the ED or an exclusive intensive care bed. Considering this diversity of strategies analyzed by the studies, the present work presents a PCT intervention dedicated to the ED. In addition, the PCT is composed of a multidisciplinary team, has protocols dedicated to the emergency environment, and carries out training for other teams, characterizing it as a highly complex intervention.
The ECP developed in the present work was similar to the study by Weng et al.(12), which also involved hiring a professional specialist to implement the intervention plan in the initial phase, training the team, and defining a flowchart with the phases of care for patients with a palliative profile to facilitate visualization and guide the teams. To our knowledge, this is the only work that approaches the complexity of the intervention performed in the present study, but it does not allow an adequate comparison due to the studied population, which was composed of younger patients and trauma victims (13).
Regarding the population in our study, chronic-degenerative conditions, especially oncological ones, were the most prevalent. Thus, populations with severe acute conditions, such as multiple trauma patients, are underrepresented, as in other studies(14). More than 70% of patients had severe functional impairment with a PPS of 10 and an ECOG 4 above 85% (15, 16). The referral of these patients to ED is because most patients had no access to palliative care before. Most studies rarely use the PPS and ECOG scales to characterize populations in different studies. These scales could provide a more objective evaluation of the populations, adjust for confounders, and there could be a potential for directing more customized care plans.
A recent systematic review points out the possible "triggers" used to identify patients who need palliative care, but there is much divergence, and the authors recommend more rigorous and systematic measurements such as those presented in this paper (17).
The outcomes used to assess the early intervention of palliative care in an emergency are different and incipient. We can group into those related to patients and their families, health professionals, and the use of health system resources. Quantitative patient-related outcomes include mortality, time to detect palliation needs, and length of stay. Qualitative outcomes are poorly documented and expressed by the quality of care assessed by family members, which is subject to several biases. The assessment of the impact on health professionals is poorly studied. Regarding the impacts on the use of the health system, the indicators are quantitative and represented by the length of stay, transfers to other hospitals, and admission to the intensive care unit.
Measuring mortality is easy but difficult to interpret. It is expected in the population of patients with palliative care, especially in the emergency setting, as observed in the present study, when in-hospital mortality represented a percentage above 65% regardless of the intervention. Unlike other studies, this study analyzed the time to death, seeking to assess whether the intervention could reduce dysthanasia represented by several undesirable procedures to which these patients are submitted and that do not imply an improvement in the quality of life. As expected, early identification in the ED reduced the time for hospitalized patients to die. Although it was not possible to determine the reasons more precisely, it is reasonable to consider that this could have occurred due to a better definition of the individual treatment plan for patients, in which the prevention of futile interventions and better preparation of the team and the family for death has occurred. We can infer the same for the behavior of out-of-hospital mortality in Strategy II, which can be influenced by the definition of the care plan and the definition of goals early, according to the patients' preferences. It is important to emphasize that we transferred more than 75% of the patients who died outside the hospital to hospices. Time to death can be a more objective indicator than the mortality rate, especially for the profile of the population in this study.
The Strategy II perception index demonstrates faster identification of patients in the ED, both in the total number of patients and when excluding in-hospital death. These data agree with the literature (12) and suggest that investment in team training can effectively improve sensitivity in recognizing patients in need of palliative care. Regarding the follow-up time, there was no significant change. For the length of hospital stay, there was a significant drop when patients were evaluated earlier, thus optimizing the use of beds. These data suggest that the early assessment of patients in the ED by a PCT is advantageous in guaranteeing patient flow.
Thus, the earlier the recognition of palliative care needs, the shorter the hospital stay. Other studies have shown a reduction in the length of hospital stay by an average of 4 days and improving the quality of life and improving patient satisfaction(17, 18). Wu et al. reported that the mean length of stay of patients who received palliative care at the ED reduced by 3.6 days compared to those who only received palliative care after hospitalization (19).
From the point of view of the health service organization, the objective indicators presented for the characterization of patients and time to event (perception and duration of hospitalization) point to better use of available resources while ensuring the quality of care for the patient and family members.
The present work did not objectively assess the intervention's qualitative indicators. However, the implementation of PCT in an ED was adequate. It organized the need for a palliative care approach, with a multidisciplinary structure, developing and incorporating work tools and care protocols that ensured the incorporation of the philosophy of palliative care. Findings confirm that the ED has a role in identifying unmet palliative care needs(17).
Although there is no consensus on how to introduce it, it is undisputed that palliative care is being incorporated into the training of emergency professionals and considered an essential component of care in other countries(20). Additionally, by providing shared decision-making for difficult situations, PCT ensures emotional support for the professionals involved and is a continuing education strategy.
Study Limitations
It was impossible to quantify all the patients who benefited from the palliative care approach in the ED. However, teams' training during the implementation process may have had a much more significant impact than the figures presented. In this observational study in which strategy II followed Strategy I in the same institution, we cannot exclude that temporal changes in the care team or the team's training may have influenced these results. Furthermore, our multivariate regression model did not include other potential confounding variables.