While more children die around the time of birth than at any other age, advances in fetal medicine and neonatology have improved the diagnostic accuracy of neonatal life-threatening conditions, and in many cases, have resulted in increased survival. For those cases where death is unavoidable and imminent, there is an urgent need to initiate discussion with families in order to formulate a consensus on end of life care. This consensus should encompass the ethic questions, processes and procedures that will be encountered.

There is increasing awareness of the need for guidance surrounding the decision to either initiate care and treatment with palliative intent or to begin with intensive care followed by transition to palliative care^{[1, 2]}. In the modern era of neonatal intensive care it is quite possible to prolong life using advanced technologies in situations where death is both imminent and eventually unavoidable. In such cases prolongation of life may increase pain and suffering for the child and often creates unreasonable hope and expectations for parents and families. Since the process of death in neonates is increasingly related to the decision when to initiate palliative care rather than failure of medical therapy, there is a need for increased awareness and attention from caregivers to the ethical principles that guide such decisions. Medical staff must strive for effective relationships with parents and the establishment of trust in order to guide parents through these difficult times^{[3, 4]}. As such, this topic requires well-defined policies and procedures derived from consensus, ensuring that all components of excellent palliative care are included, thus improving the experience for families and caregivers.

The Netherlands and United Kingdom have adopted general guidelines for palliative care^{[5, 6]}. For the most part they are non-specific, lack details, and have been heavily criticized^[7] for being too vague. It has also been suggested that they ignore important prognostic factors such as timing of death and details of diagnosis, and on rare occasions, cases ending in unexpected prolonged survival. In North America the neonatal end of life care practices vary between institutions^{[8, 9]}. A 2014 study from Saskatchewan and Manitoba found that there were variations in the end of life processes and protocols between the 3 sites studied and that none of the sites had every element of what has been described in the literature as ideal^[4]. Another report from Canada highlights the challenges that face physician trainees^[10]. The challenges faced by all neonatal caregivers emphasized the need for participation and input from several key parties such as family members, medical teams, social workers, and ethics boards. An important part of this goal is to evaluate the parents’ experiences in order to include their values in the consensus. It has been suggested that parents experience less guilt when the physician starts the discussion on when to initiate palliative care^[11]. It has also been found that parents with a higher educational status are able to come to a decision on the matter more quickly^[11]. A cross sectional survey among US neonatologists and trainees revealed that perspectives of palliative care and end of life care are important and that exploration of differing values is needed^{[12, 13]}. A review of literature revealed that development of consensus was difficult, suggesting that more research and models are needed which emphasize shared decision-making , avoiding acting on labels derived from gestational age categories alone along with improved competency, knowledge and confidence^{[12, 13, 14, 15, 16, 17]}. Recently, experts in palliative care treatment have been involved in the end of life care of neonates^[18]. While more and more preterm infants at the margin of extreme prematurity survive, Janvier et al^[19] have reminded us of the need for advanced ethical knowledge and sound decisionmaking. As a preliminary step towards the goal of developing a national widely consultative and approved guideline we felt it was timely to report our experience. We also believed that our experiences and results obtained from our study may help other hospitals and neonatal care staff.

Ethical approval was obtained from the University of Saskatchewan Biomedical Research Ethics Board. A retrospective review was conducted of the health records of neonates who died within a 5-year period (2008-2012). Study eligibility included all perinatal, neonatal and infant deaths; the author (Koravangattu Sankaran) was involved in the care and discussion with family; palliative care was offered; follow-up information was available; and in the absence of pediatric palliative care consultant, a neonatologist offered palliative care.

The Division of Neonatology at Royal University Hospital in Saskatoon SK follows local, consensusdeveloped processes and procedures for neonatal end of life care (1990 reviewed in 2000), which included the following:

(1) Objective confirmation of diagnosis of a condition where intact survival is negligible or impossible (for example, severe grade IV intraventricular hemorrhage, irreversible multi-organ failure, prematurity at the margin of survival, severe irreversible brain damage, multiple severe congenital anomalies, genetic, metabolic, and chromosomal disorders).

(2) Initiation of end of life care discussion with the family following confirmation of diagnosis.

(3) The attending neonatologist, along with the neonatal nurse caring for the infant, meet with the parents formally and explain the situation. The first meeting focuses on explaining the condition of the infant, futility of further intensive care, and the options for the family. Further meetings are scheduled as and when requested by both parties until the final decision is reached.

(4) A senior neonatologist is consulted in order to review the diagnosis and confirm the decision.

(5) Appropriate orders and detailed progress notes are written in the chart as and when necessary.

(6) Senior clergy from the religion of the infant’s parents and or elders in the family are included in the discussions.

For all infants included in the study families were contacted 8 to 12 weeks after death for a face-toface semi-structured interview where the neonatologist responded to the parents/families concerns and questions. On exit they were asked to offer a critique and opinion about the process, procedures and care provided by staff and on overall hospital stay. A liaison nurse, who is an expert in family relations, arranged the follow up meetings, offered bereavement services as necessary and attended the meetings along with the neonatologist. The bereavement services included, a follow up phone conversation, a home visit if necessary and arrangement of hospital chapel visit and prayer with hospital staff and family in selected cases. In cases with a single mother, a grandparent from the maternal or paternal side was asked join in on the discussion with mother’s consent. After the joint meeting, the liaison nurse met with the family to document feedback, wrote down salient points, and then shared this information with the neonatologist. Verbal consent was obtained prior to collecting feedback.

The primary author collated that data for this study. A clinical data sheet was prepared and reviewed with the following information: final diagnosis; initiator of end of life discussion (family member, staff, or both); number of discussions held; final decision (initiate palliative care, switch from intensive to palliative care or other); where death occurred; palliative care details; and follow-up data. Descriptive statistics were used to explain the study results where applicable.

There were 40 infants who died of all causes during the time period. Thirty infants met inclusion criteria (16 males and 14 females). Gestational age at birth ranged from 22 to 41 weeks.

Eleven (37%) out of 30 were diagnosed with genetic and metabolic disorders (including severe congenital lactic acidosis, severe non-ketotic hyperglycenemia, severe harlequin ichthyosis, alpha thalassemia, severe hyperammonemia and severe hepatic cytochrome deficiency). Eight (27%) out of 30 suffered severe asphyxia with irreversible brain damage and multi-organ failure (MOF), 7 (23%) were diagnosed with extreme prematurity with intraventricular hemorrhage and MOF, and 4 (13%) with severe lethal congenital anomalies and chromosomal disorders (including Pena-Shokeir syndrome, Wolfe-Hirschon syndrome with multiple congenital heart defects, Ives-Houston syndrome, Meckel-Gruber syndrome, CHARGE association with congenital heart disease, trisomy 21 with severe necrotizing enterocolitis, multi-organ failure and congenital heart disease, trisomy 13 and trisomy 18).

In 24 (80%) out of 30 cases the physician or staff team member initiated the discussion. One (3%) discussion was family-initiated, and in 5 (17%) cases, both family and physician started the conversation for compassionate care around the same time.

The mean number of discussions needed per case was 7, with a range from 1 to 30. However, one case had 30 discussions, which was atypical and never actually resulted in a final decision. When the outlier was excluded from the analysis, a mean of 4 discussions were required to achieve consensus. In some cases the discussions were started prenatally. In general, single, first Nation and Catholic parents needed about twice as many discussions compared to other groups prior to reaching a final decision.

In 18 (60%) out of 30 of cases the decision was made to switch to palliative care after intensive care had already begun. Of the remaining cases, 11 (37%) resulted in a decision to only initiate palliative care and in 1 (3%) case the parents could not come to an end of life decision-in this case intermittent life supporting intensive care was provided as and when necessary.

seventeen (57%) out of 30 infants passed away in the NICU, including one in the PICU. Nine (30%) infants died in the delivery room area and 3 (10%) infants died at home: one case with severe hypoxicischemic brain injury following perinatal asphyxia and two cases with severe multiple congenital anomalies. All deaths in hospital occurred within minutes to hours after stopping aggressive life-sustaining therapy. In the one case where an end of life decision could not be reached, one parent insisted on life-sustaining care and the other insisted on palliative care. Eventually the infant passed away in the PICU.

Morphine and fentanyl were the intravenous drugs used to ensure comfort, while acetaminophen and chloral hydrate were the oral drugs used. In cases where there a definite antenatal diagnosis had been made, discussions were initiated prior to delivery, and most deaths occurred in the delivery room area after comfort care was initiated. However, two deaths occurred in the NICU even though parents chose to proceed with comfort care in lieu of life-sustaining care. These infants had multiple congenital anomalies and were moved from the delivery room area to the NICU. Parents and family members were uniformly concerned that their child would feel pain during death and needed reassurance that their infant passed away peacefully. This was again raised in followup interviews. Parents and families were reassured that their loved ones would not feel pain as they lost consciousness immediately after stopping aggressive therapy. Twenty out of 30 infants received specialized palliative care.

Twenty-seven (90%) out of 30 families were satisfied with the processes and procedures that were followed in reaching the end of life decision. Twenty families returned to the NICU with thank you cards and gifts. Fifteen (50%) parents and/or family members received bereavement services from the liaison nurse. Two (10%) families felt that the communication and support could be improved. For example, one family was upset by the fact that the neonatologist was not freely available to speak to the family during weekend hours; subsequently this was corrected. Further comments included being thankful for having been provided with a private, quiet place for grieving and for being able to hold the infant during the time of death. All parents provided positive comments for the empathy, support and comfort offered by the liaison nurse.

In this study there was a large proportion of infants with multiple congenital anomalies and rare genetic and metabolic diseases. We believe that it is the pattern we have been observing in our centre over the years and reflects the ethnic population mix in our province. Our perinatologists routinely perform ultrasound in an antenatal clinic and have an excellent referral system; therefore, in many cases clinical diagnoses and the decision-making process were started during the antenatal visits. Neonatologists, geneticist, palliative care specialist, pediatric cardiologist and social workers meet with parents as and when necessary for antenatal consultation. This highlights the importance of multi-specialty consultations, specialized targeted education, and regular combined perinatal/neonatal rounds. The involvement of our palliative care specialist was particularly helpful for both staff and families. She was essential in improving communication between parties and provided us with specialized end of life care expertise.

The need for an average of 4 discussions prior to achieving neonatal end of life decision and additional discussions in special situations is interesting and important for all caregivers and planners alike. To our knowledge this has not been reported elsewhere in this manner. This would be relevant during planning of personnel allocation and preparation of future guidelines.

During the follow-up meeting with families 8-12 weeks after death, we were able to discuss their needs, what processes they found helpful, and what their fears and apprehensions were. The services of an empathetic well-trained liaison nurse were invaluable. She acted as a connecting link with the staff and families and helped in the free flow of communication between staff and families and acted as a catalyst. The additional benefits included the following: (1) To confirm the diagnosis and reassure the family and parents in general. (2) To provide support in cases where the results of special diagnostic or pathologic tests were only available later. (3) To answer any lingering questions. (4) To provide bereavement counseling and closure. (5) To receive and deal with important family satisfaction details.

Again it is important and relevant to realize that all parents and family members were thankful for the private quiet space provided to them during grieving. The care, support and free availability of the liaison nurse were also relevant because all of these highlighted the need for the provision of adequate space and staff.

Ethically speaking, beginning care with palliative intent and switching from intensive care to palliative care are not different-although the processes and procedures involved are different. We believe the perception exists that beginning initial treatment with palliative intent may be easier and causes less suffering and pain to the infant than switching from intensive to palliative care. This did not appear to the case in our cohort. We do, however, recognize that palliation of pain and suffering is paramount in both scenarios.

We also observed that ethnicity and religion played a role in reaching end of life decisions. First Nations families in general needed more time and roughly twice as many consultations and discussions, which is similar to a previous report^[20]. This may indicate the need for more enhanced, culturallysensitive discussions and educational sessions. Our findings are similar to a previous published review report^[21] where timings of death after decision were discussed. Ethical issues arising from racial, ethnic, and religious perspectives are indeed very difficult to deal with for families and staff alike and ideally should be addressed without inciting wounds or legal repercussions. Appropriate, advanced training and education is crucial in this area.

A report called for the development of proper criteria for initiating life-sustaining treatment and asked for less medical paternalism and more informed parental involvement^[22]. Recently there was a dialogue among the members of the Saskatchewan Medical Association regarding this issue initiated by the Royal College of Physicians and Surgeons of Canada, however the authors are unaware whether this discussions included infants and neonates. In Norway the neonatologists follow somewhat similar processes and procedures, but the second opinion was obtained from a staff neonatologist belonging to a separate NICU. The American Academy guidelines on palliative care deals mainly with somewhat older children and not as detailed in this paper on neonates^[23]. The palliative care details are somewhat similar, however we believe our approach, attention to details, meticulous, consistent and effective delivery of process and procedures are unique. It is important to recognize that this study followed care practices, ethical principles, religious beliefs that are prevalent in western societies and may not be applicable to all hospitals and societies in the world.

Weakness of this report include its retrospective nature, the semi-structured collection of follow-up data and a single center experience. However the strengths are: excellent family/parent satisfaction reports, consistent and effective delivery of process and procedures, and inclusion of a trained and experienced liaison nurse in the process and delivery of care.

In summary, an average of 4 discussions were needed prior to reaching end of life decision. Switching from aggressive care to comfort care was the process preferred by parents and families. Feedback from parents and the support of a trained liaison nurse were extremely helpful. The centuries-old words of Hippocrates still ring pure and true today:

“First I will define what I conceive medicine to be in general terms, it is to do away with the sufferings of the sick, to lessen the violence of their diseases, and to refuse to treat those who are overmastered by their disease, realizing that in such cases medicine is powerless.” - The Hippocratic Corpus.