Current Practice of Dutch Hospital-Based Palliative Care Teams: Advance Care Planning in Advanced Stages of Disease

Background: Specialist palliative care teams (PCT) are consulted during hospital admission for advice on complex palliative care. These consultations need to be timely to prevent symptom burden and maintain quality of life. Insight into specialist PCTs may help improve the outcomes of palliative care. Methods: In this retrospective observational study, we analyzed qualitative and quantitative data of palliative care consultations in a six-month period in four general hospitals in the northwestern part of the Netherlands. Data were obtained from electronic medical records. Results: We extracted data from 336 consultations. The most common diagnoses were cancer (54.8%) and organ failure (26.8%). 40.2% of patients were restricted to a chair or bed and 52.3% had an estimated life expectancy of less than three months. Within two weeks after consultation, 53.2% of the patients died, and the median time until death was 11 days (range 191) after consultation. Most patients died in hospital (49.4%) but only 7.5% preferred to die in hospital. Consultations were mostly requested for advance care planning (31.6%). End-of-life preferences focused on last wishes and maintaining quality of life. Conclusion: This study shows that palliative care consultations focus on terminal care and are more crisis-oriented than prevention-oriented. Death often occurs too quickly after consultation for end-of-life preferences to be met and these preferences tend to focus on dying. Educating healthcare professionals on when to initiate palliative care would promote a more prevention-oriented approach. Factors that indicate the need for timely PCT consultation should be dened.


Introduction
In 2018, over 150,000 people died in the Netherlands, of which 85% were ≥ 65 years old and 56% were ≥ 80 years old [1]. The most common reasons for death were dementia (12.7%), stroke (6.5%), lung cancer (5.9%), coronary diseases (5.6%), and chronic obstructive pulmonary diseases (4.7%). Whilst most people (60%) prefer to die at home, for 30% of these people with chronic diseases, this wish is not ful lled as these people die in hospital or in a nursing home [2].
Palliative care can bene t people with life-limiting chronic diseases. It can improve the quality of life for both the patient and their family by treating physical, phychosocial, and spiritual problems [3]. If palliative care is initiated during early stages of disease, it can prevent suffering by guiding the patient's decisions on their end-of-life preferences. These preferences can be discussed during advance care planning (ACP), which is de ned as 'the ability to enable individuals to de ne goals and preferences for future medical treatment and care, to discuss these goals and preferences with family and healthcare providers, and to record and review these preferences if appropriate' [4]. ACP is considered an essential part of good palliative care [5] and if it is not performed during early stage of disease, treatment decisions are often crisis-oriented and do not meet end-of-life preferences [6].
One crisis-oriented decision is acute hospital admission, which many older people experience in the last phase of life [7]. After being admitted to hospital in the Netherlands, these older patients are treated by healthcare professionals who often do not have specialist expertise in palliative care, such as junior doctors, medical specialists and nurses, according to the generalist plus specialist model [8]. In this model, general healthcare professionals are primarily responsible for providing palliative care. However, palliative care can be complex and ACP conversations can be di cult, and is only covered limited in the curriculum for healthcare professionals [9].
To help in these complex palliative care situations, a specialist palliative care team (PCT) can be consulted. Since 2017, every Dutch hospital providing oncology care must have a PCT [8]. These PCTs can be consulted for advice on treating inpatients and outpatients at different stages of disease and life, and these consultations have been shown to promote quality of care and satisfaction with care [10][11][12].
Most studies on PCTs have focused on patients with cancer; therefore, detailed insight into current practice of PCTs in the general population, including patients with organ failure, is lacking. Furthermore, the end-of-life preferences discussed during ACP conversations are currently unknown. Greater insight into the use, content, and timing of PCT consultations could help improve palliative care. In this restrospective mixed-method study, we assessed all PCT consultations performed in a six-month time period. Our aims were 1) to determine which patients received a PCT as well as the timing, content, and follow-up outcomes of the consultation and 2) to explore the wishes and end-of-life preferences of patients receiving a PCT consultation.

Methods
We performed a retrospective observational study in which data performed both quantitative and qualitative analysis. This study was conducted in four general hospitals in the northwestern part of the Netherlands. In each hospital, data on all PCT consultations were collected over a six-month period (January -June) in 2017 or 2018. All rst-time consultations for patients ≥18 years old were included. There were no exclusion criteria. The results are reported according to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines [13].

Ethical considerations
Informed consent was not required for this retrospective data collection since all data were analyzed and stored anonymously. Furthermore, obtaining informed consent was not feasible since most patients were already deceased at the time of data collection. In all participating hospitals, the study protocol was assessed and approved by the hospitals scienti c board before data were collected.

Data collection
Every participating hospital provided a list of patients who received a PCT consultation during the sixmonth study period. Data were collected from the electronic medical records (EMR). Since one participating hospital shifted from handwritten patient records to EMR in 2017, in this hospital data on consultations in 2018 were collected.

Patient characteristics
We extracted patient characteristics from the EMR, including demographic characteristics, main lifelimiting diagnosis, prescribed medication, estimated life expectancy, WHO/ECOG performance status [14], and healthcare use in the six months before consultation. Healthcare use included emergency room (ER) visits and unplanned hospital admission.

Mortality and place of death
We collected data on the patients' preferred place of death and the eventual place and date of deaths.

Consultation characteristics
Data on consultation content included the referring specialist, reason for and type of consultation, timing of consultation, and the number of visits and follow-up contacts by PCT members.

End-of-life preferences and wishes before death
We collected data on end-of-life preferences and wishes before death from reports written by the PCT consultant in the EMR. Quotes were taken literatim from the reports.

Analysis
Quantitative data Standard descriptive statistics were used to describe patient and consultation characteristics. Data were analyzed using IBM SPSS Statistics (Version 26). Quantitative data included all patient and consultation characteristics except for the end-of-life preferences and wishes.
Qualitative data End-of-life preferences and wishes before death reported by the PCT consultant were coded using open, thematic, and attribute codes [15] and themes were identi ed [16]. We used MaxQDA 2020 for this analysis.

Patient characteristics
We collected data from 336 PCT consultations. The number of consultations per hospital over the sixmonth period varied from 27 to 183. Patients had a mean age of 75.1 years (SD 11.6) and 54.8% were male. The most common main diagnosis was cancer (54.8%), followed by organ failure (26.8%) and surgical problems (10.7%). In the six months prior to the PCT consultation, 47.9% of patients had visited the ER and 46.1% had been acutely admitted to hospital. Most patients (40.2%) were restricted to their bed or chair at the time of consultation, and 52.3% had an estimated life expectancy of less than three months (Table 1).    End-of-life preferences and wishes before death were discussed and reported in 145 (43%) consultations. In 22.1% of consultations, patients noted they had no preferences or wishes before death, half of these cases were because the patient did not want to talk or think about the end of their life.
We extracted several themes from the PCT reports about end-of-life preferences and wishes before death.

Future treatment, organization, and transition of care
For one patient, preventing more hospital admissions was discussed. The most common wish regarding transition of care was to go home. Some patients did not explain this wish while others said they wanted to go home to be with family or to die at home. Some patients said they wanted to go to a hospice.
Patients also wanted clarity about their treatment options, the opportunity to discontinue treatment if they wanted to, and clear statements about why they wanted to stop treatment.
"Wants to undergo treatment, mild side effects are not a problem for him but in case of serious side effects and especially severe pain, he wants euthanasia."

Quality of life
For quality of life, patients wanted to maintain control, enjoy life, and not suffer from symptoms, particularly pain and dyspnea.
"The most important thing is to be as comfortable as possible, and to experience as little symptom burden as possible"

Last wishes
Last wishes included activities, seeing and saying goodbye to family, and euthanasia. Reported activities ranged from speci c one-time activities to more general long-term activities. These long-term activities mainly included holidays and being able to perform activities in general. In many cases, the patient or PCT consultant questioned the feasibility of these wishes.
"He says he still wants to do many things. He has many chores in and around the house waiting to be done. He mentioned so many chores, I indicated I thought that was not realistic and that it's best to focus on doing the things that bring joy and peace." Enjoying life was a general, non-concrete wish reported by patients. Keeping control of life and dying was also an important topic.
Patients mentioned family, friends, and pets in their last wishes. For many, seeing and being with a loved one before death, sometimes to say goodbye, was important. Others mentioned important days and events they wanted to attend, such as anniversaries and birthdays.
"The biggest wish is to be able to see grandchildren grow up and experience the upcoming birth of a grandchild." Lastly, many patients mentioned euthanasia in their last wishes; for some this was a short-term wish and for others it was something they wanted in the future.
"Wants to say goodbye to his children before euthanasia starts."

Discussion
In this mixed-method study, we analyzed all PCT consultations performed in four hospitals in the Netherlands within a six-month period. Our ndings suggest that palliative care consultations usually take place during the terminal phase of illness, since most patients died within three months after their consultation. A quarter of patients died during the hospital admission at which the PCT was consulted, and half died within one or two weeks after consultation. This meant that ACP, which was often the reason for consultation, could not be arranged in time. This could have contributed to many patients not dying in their prefered place of death. In addition, many reported end-of-life wishes focused on the nal days and weeks before death and rarely had to do with treatment. The feasibility of these wishes was sometimes questioned.

Interpretation of ndings
Many studies have focused on palliative care in oncology patients [17][18][19][20]. In our sample, only half of the patients were diagnosed with cancer. As in previous research [10], we found that most consultations were requested for patients in the last three months of life. This suggests that palliative care continues to focus on terminal care, as previously reported by Dalgaard et al [21]. Patients who received a PCT consultation were often acutely hospitalized and severely restricted in daily activities, suggesting a more ad-hoc approach to palliative care that does not recognize the need for palliative care timely enough [4].
Although most patients in our study were diagnosed with a life-limiting chronic disease and had been acutely admitted to hospital in the six months before the consultation, identifying the need for palliative care still appears to be di cult. Previously reported barriers to identifying the need for palliative care include an uncertain prognosis, lack of knowledge about palliative care, and lack of consensus between healthcare professionals (van Doorne et al, submitted) [21,22].
Although palliative care was initiated during advanced stages of disease, ACP was the most common reason for PCT consultation in our study. This high proportion of consultations for ACP suggest that general physicians have di culty discussing end-of-life matters and are reluctant to start these conversations themselves. In addition, the increasing attention placed on ACP [23] could explain this high proportion of consultations. ACP is a complex process, for which often multiple conversations are needed for patients to fully explore their end-of-life preferences and make end-of-life treatment decisions [4]. It is questionable if the ACP conversations reported here were su cient to serve this purpose.
One of the goals of ACP is to reduce overtreatment [24]. For many patients in our study, the hospital admission at which the PCT was consulted was their last care transition before death. If the PCT had been consulted during an earlier hospital admission (half of the patients had been admitted to hospital in the previous six months), then perhaps proactive communication could have avoided this last admission for some patients [25]. This would possibly also reduce the number of in-hospital deaths, which is often a goal of ACP [26]. In our study, 62% of patients wanted to die at home, which is in line with previous studies [2,27,28]. However, in our study, most patients actually died in the hospital, although very few patients wanted this. The congruence between the preferred place of death and the actual place of death was lower in our study than reported in previous studies [2]. This could be related to the more advanced stages of disease and the limited time until death in our patients, which cannot be compared to patients who did not receive a PCT consultation.
The end-of-life preferences reported by the patients in our study re ected the severrity and advanced stages of their disease. In all themes, end-of-life wishes included things the patient wanted to do before death and were often one-time and nal wishes. Reconsidering these wishes was not possible in most cases.
The underlying goals of ACP were described by Fleuren et al. [24] in a systematic review. These goals include improving quality of care, preparing for the end-of-life, and respecting patients' autonomy. These goals could be achieved by preparing patients for 'in-the-moment decisions'. Patients in our study were already 'in-the-moment', and were not able to make treatment decisions anymore. As a result, the themes 'future treatment' and 'wishes regarding dying and death' were focused on preparation for the end-of-life limitedly and often did not include treatment decision making. The theme 'quality of life' is comparable to previously described end-of-life wishes of terminal patients [29], in which symptom management was important.

Strenghts and limitations
A strength of our study is that we included all PCT consultations rather than focusing on a speci c patient population. This is in contrast to previous studies that often focused on patients with cancer or other speci c diseases. Therefore, we provide insight into PCT consulation in a broader more general population. Furthermore, including four hospitals gave a sample that is representative of the northwestern part of the Netherlands, re ecting urban as well as rural parts. The detailed data collection provided useful information on the characteristics of patients who receive PCT consultations and a deeper insight into the content of PCT consultations, including the end-of-life preferences and wishes of patients.
However, our study also has limitations. First, all data were obtained from the patients' EMR and focused on reports written by the PCT consultant, so the completeness of our qualitative data depended on completeness of the healthcare providers' report. Some topics may have been discussed with the patient before the consultation or may not have been included in the EMR. In addition, the patients' preferences we collected were described by members of the PCT and may have been misinterpreted. Second, we could not track down the place of death for all patients, particularly those who did not die in the hospital, however, we do know the place of death for 72% of all patients, which we consider to be su cient.
Third, the number of consultations varied widely between the included hospitals so our ndings might be mostly contributable to hospitals that conducted the most consultations. To improve the generalizability of our ndings, this study should be performed in all parts of the Netherlands or in more countries.
In conclusion, our study shows that timely identi cation of palliatiave care needs remain to be lacking. Palliative care consultations are still focused on terminal care and are more crisis-oriented than prevention-oriented. Because death occurs soon after consultation, end-of-life preferences are often not met, such as dying in the preferred place of death. In addition, patients' end-of-life preferences are mainly focused on the last days and weeks before death and do not focus on future care, which ACP aims to discuss to ensure that end-of-life care meets the patients' preferences. Since most consultations are requested to perform ACP, healthcare professionals could bene t from education in timely identifying palliative care needs and training in initiating palliative care. In addition, patient-related factors that might identify a need for PCT consultation (for instance an acute hospital admission because of foreseeable symptom burden for patients with life-limiting diseases) need to be de ned to ensure a more preventionoriented approach to palliative care that allows crisis situations to be discussed in advance so that patients' end-of-life preferences can be met. Informed consent was not required for this retrospective data collection since all data were analyzed and stored anonymously. Furthermore, obtaining informed consent was not feasible since most patients were already deceased at the time of data collection. In all participating hospitals, the study protocol was assessed and approved by the hospitals scienti c board before data were collected.

Consent for publication
All authors gave consent for publication Availability of data and material The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
There are no competing interests