This study analysed treatments, symptoms and clinical management of adult inpatients at the end of their life. Notably, the majority of patients died in internal medicine wards where patients presenting with exacerbations of their chronic diseases are frequently hospitalized.[17]
Considering the differences in death expectation found in the present study, many patients underwent several diagnostic and therapeutic procedures in the last hours of life. Routine blood tests were the major diagnostic procedure performed in patients whose death was highly expected. Many patients underwent a considerable volume of blood samples for laboratory tests during the last hours,[18] phenomenon already described in previous studies, where at least 50% of patients whose death was highly expected performed blood tests before dying.[14, 19] Literature widely describes that also radiological procedures for diagnostic purposes are frequently performed in patients with poor prognosis admitted to hospice,[20] even if in the present study the diagnostic procedures (especially chest radiography) have been performed mainly in patients whose death was not expected.
The intensivist was the most requested consultant and the main responsibility was to declare the do-not-resuscitate order. This could explain the very small number of activations of the medical emergency team to perform cardiopulmonary resuscitation. Thus, the main activity of the intensivist was contributing to the decisions making over the appropriate intensity of care, both for patients suffering from end-stage chronic-degenerative and neoplastic diseases, as well as for patients with acute clinical conditions and poor prognosis.[21, 22] However, not all intensivists can make decisions about the appropriateness of treatments (i.e., administration of new antibiotics) or about the suitability of escalation to high dependency or intensive care. This is also acknowledged by the Italian Society of Anaesthesia, Analgesia, Resuscitation and Intensive care (SIAARTI), which has published specific recommendations to guide EOL decision-making for patients outside the intensive care unit.[23]
Performing diagnostic and therapeutic procedures clashes with the problem of symptoms control in the last hours of patients’ life. Symptoms like pain, agitation and dyspnea were only partially controlled in the present sample. In a large proportion of patients, the level of consciousness was severely compromised, thus they were not able to communicate appropriately the characteristics of pain. For this reason, it is recommended using clinical scoring tools for pain evaluation which may include patients’ behavior and not only professionals’ clinical judgment.[24] Nevertheless, reducing peri-mortem to a complex and demanding measuring act may be an obstacle for the final decision to care and not to cure.
Amidst symptoms, an effective pain control management is expected in most patients, and physicians should be comfortable in prescribing repeatable dose of analgesic drugs to reach the analgesic peak effect.[25] In this study, all patients with cancer received opioids. As documented in the last national report,[26] their usage seems to be constantly increasing. However, despite the administration of analgesic drugs, pain control was not always achieved and different studies show that an important percentage of patients without cancer (25–40%) does not receive an adequate pain-relieving treatment.[27] In addition to pain, the presence of respiratory symptoms (i.e., dyspnea), which mainly characterized the present cohort, may be of interest of the consultant in palliative medicine,[28] whose competencies were requested in a small number of patients, just because the hospital does not have a palliative care service.
As regards nursing care, patients continued to receive body hygiene care, active mobilization and vascular access management until the moment of death. Patients with a lower Norton score were at high risk of developing pressure ulcers, and this might explain the great number of nursing interventions received, mostly related to skin care and prevention of pressure ulcers onset, such as anti-bedsore mattress positioning.[29] Body and oral hygiene care are very important nursing interventions to provide comfort to the bedridden patients.[30] In particular, poor oral hygiene is the most common cause of mouth problem,[31] especially in weakened and fatigued patients. Nurses perceived a good EOL quality when they carried out a high number of interventions to old and frail patients, as summarised by MCA. Despite different cultural aspects could contribute to the concept of good death, dying without discomfort and suffering is considered a good way of dying in any culture.[32] In order to improve the quality of EOL phase, treatments and nursing interventions are not always enough to relieve symptoms but it is also necessary to timely involving the family to support patients emotionally.[33] On the whole, the present findings show how interventions to ensure comfort and intimacy for patients whose death was highly expected were frequently carried out. Correct information regarding the severity of clinical conditions and the presence of family members, without any time constraints, were ensured to almost the totality of patients. These aspects were also included in the Liverpool Care Pathway model, a protocol that has been adopted in the United Kingdom for a decade, which help physicians and nurses to increase the quality of treatments in the EOL care.[34]
Despite the correct information about the clinical conditions, it is rare to address the theme of organs and fabrics for transplantation as a subject in the EOL care discussions.[35] Cornea transplant is the most common form of organ transplantation performed globally.[36] Unfortunately, it records a high refusal rate during organ donation request, as reported in the present study as well. As a result, the number of patients who need transplant exceed the cornea tissues availability.[37] Nurses are the caregivers most involved in EOL care and they could positively affect the organ donation request.[38] However, without an appropriate EOL care with patients’ families, donation, and the opportunity to help a large number of recipients, will not occur.[39]
The clinical judgment of nurses and physicians showed a moderate agreement about death expectation. This is not surprising, considering that identification of the EOL phase is incredibly difficult for the healthcare staff. Defining when a patient is in a phase of stability or instability, worsening or in a terminal phase (phase illness) can be very challenging,[40] particularly in patients without cancer.[41] However, prognostic information remains necessary not only for patients and their families, but also for healthcare providers, in order to guide their action and offer necessary interventions. Unfortunately, survival predictions made by clinicians suffer from their subjectivity and sometimes are overly optimistic and not always reliable.[42, 43] On the contrary, nurses could be the first to recognize patients’ end of life, potentially making a significant contribution to the EOL quality.[44] In fact, nurses and physicians showed also different rates of agreement about the quality of EOL in the present study. Nurses gave a lower rating compared to physicians, suggesting that they may equally expect death but perceived a different quality.
These divergencies can have various origins. First, training on EOL issues is very inhomogeneous and the most part of healthcare providers feel uncomfortable in facilitating EOL discussions.[45] Also, nurses often have conflicting feelings when providing care to dying patients and experience distress when disagreeing with the appropriateness of the medical treatments.[46, 47] On the whole, the different professional background and professional aims may be responsible for such differences.
Study limits
The study has several limitations. The sampling was limited to our Institution and generalization should be made with cautious to smaller Italian hospitals. The expectation of death was asked only after the patient had died, and the subjectivity of the answers provided by nurses and physicians during the interviews could have overestimated the quality of EOL management. This is an unavoidable limitation for assessment of the dying experience. However, the limited time-window used to assess quality of dying in the present study should have minimized recall bias. Furthermore, we used only nurses’ judgment to stratify our analyses, being available other studies already showing lower accuracy to predicting death by physicians compared to nurses.[48, 49]