Advance Care Planning in Dutch Nursing Homes During the First Wave of the COVID-19 Pandemic

Objectives To explore how physicians in Dutch nursing homes practiced advance care planning (ACP) during the first wave of the COVID-19 pandemic, and to explore whether and how ACP changed during the first wave of the pandemic. Design Qualitative analysis of an online, mainly open-ended questionnaire on ACP among physicians working in nursing homes in the Netherlands during the first wave of the COVID-19 pandemic. Setting and Participants Physicians in Dutch nursing homes. Methods Respondents were asked to describe a recent case in which they had a discussion on anticipatory medical care decisions and to indicate whether ACP was influenced by the COVID-19 pandemic in that specific case and in general. Answers were independently coded and a codebook was compiled in which the codes were ordered by themes that emerged from the data. Results A total of 129 questionnaires were filled out. Saturation was reached after analyzing 60 questionnaires. Four main themes evolved after coding the questionnaires: reasons for ACP discussion, discussing ACP, topics discussed in ACP, and decision making in ACP. COVID-19–specific changes in ACP indicated by respondents included (1) COVID-19 infection as a reason for initiating ACP, (2) a higher frequency of ACP discussions, (3) less face-to-face contact with surrogate decision makers, and (4) intensive care unit admission as an additional topic in anticipatory medical decision making. Conclusions and Implications ACP in Dutch nursing homes has changed because of the COVID-19 pandemic. Maintaining frequent and informal contact with surrogate decision makers fosters mutual understanding and aids the decision-making process in ACP.

Worldwide, nursing homes have been hotspots of COVID-19 mortality. 1 During the first wave of the COVID-19 pandemic, mortality rates in long-term care facilities in the Netherlands doubled, with an excess mortality of 5200 deaths between the 9th of March and the 10th of May, 2020. 2 Of nursing home residents who tested positive for the novel coronavirus (SARS-CoV-2) during the first wave, 42% died within 30 days. 3 The high morbidity and mortality rates among this population underline the importance of advance care planning (ACP) during the COVID-19 pandemic. 4,5 ACP enables individuals to align their future medical treatment and care with their personal goals and preferences. 6,7 ACP becomes increasingly important with increasing frailty. 8 Residents of nursing homes comprise a frail population, for whom ACP is of the utmost importance.
In Dutch nursing homes, proactive on-site medical care during office hours as well as on-call 24-hour medical care is provided by elderly care physicians, a medical specialty that encompasses generalist and specialist competencies. 9,10 The 3-year vocational training for elderly care physicians focuses mainly on primary and specialized care for frail older people including ACP with patients and their surrogate decision makers. 10 Hence, elderly care physicians have broad experience in practicing ACP with nursing home residents.
According to the Dutch association of elderly care physicians, typical occasions for ACP are admission to a nursing home, regular biannual multidisciplinary meetings and acute illness or decline in health including imminent death. 11 ACP is practiced by an elderly care physician or by other physicians or nurse practitioners under supervision of the elderly care physician.
An important part of ACP is discussing preferences and wishes regarding future medical treatment decisions. 7 Probably, the COVID-19 pandemic affects ACP discussions and specifically anticipatory medical care decision making in nursing homes. Physicians, residents, and their surrogates noticed the limited treatment options in severe cases and the poor prognosis of SARS-CoV-2 infections among residents of nursing homes. 8,12 Furthermore, as in many other countries, the lockdown in the Netherlands included a nationwide restriction for all visitors in long-term care facilities, including nursing homes, from the 19th of March to the 26th of May 2020, limiting possibilities for face-to-face ACP discussions. 1,13,14 The urgency of ACP on the one hand, and stress, time pressure, and limited possibilities for face-toface meetings on the other hand, presumably had great impact on daily practice for elderly care physicians, especially in their routine of practicing ACP. We hypothesized that elderly care physicians would have developed new routines to continue this important aspect of their work.
In this study, we explore how physicians working in Dutch nursing homes practiced ACP discussions during the first wave of the COVID-19 pandemic. Next, we explored whether and how ACP changed during the first wave of the pandemic.

Design
This qualitative study was conducted during the first wave of the COVID-19 pandemic. We designed a questionnaire on how physicians in Dutch nursing homes practiced ACP discussions during the first wave of the COVID-19 pandemic and whether and how this had changed by the pandemic. The questionnaire included open, categorical, and closed questions. The majority of the questions were open and respondents were invited to elaborate on their response.

Data Collection
Questionnaires were sent by e-mail on the 29th of May 2020 to physicians working in nursing homes associated with the vocational training of residents in Elderly Care Medicine at the Department of Medicine for Older People of Amsterdam UMC in Amsterdam, the Netherlands. Also, a link to the questionnaire was added to the newsletter of Verenso, the Dutch association of elderly care physicians.
First The questionnaire consisted of 3 parts (Supplementary Material 1). First, sociodemographic characteristics were assessed. Second, we asked to describe the last case in which the respondent had an ACP discussion. We asked whether the treatment policy before the ACP discussion was curative (the treatment goal is to cure disease), palliative (the treatment goal is to optimize the quality of life, prolonging life is acceptable), or symptomatic (the treatment goal is to optimize quality of life, prolonging life is unacceptable). Next, we asked whether or not the ACP discussion resulted in change of physician treatment orders. We asked who participated in this discussion and in which way the discussion had technically taken place. Next, we asked whether the following factors played a role in the ACP discussion: medical situation, preference of the resident, preference of the surrogate, professional view of other professionals, and COVID-19erelated factors. Finally, we asked respondents whether they judged that ACP had been influenced by the COVID-19 pandemic in the specific case they described, and whether the pandemic was of influence on ACP in general.

Data Analysis
All questionnaires were distributed among 5 researchers, who were all practicing physicians in nursing homes at the time of the data collection. The analysis encompassed 5 steps: (1)  Discrepancies in the codes were discussed and solved in consensus. Whenever changes in codes were introduced, all previously coded questionnaires were checked again. An iterative approach (the process of going back and forth between the data, the codes, and themes) was followed across the different steps to ensure a systematic analysis. This approach was followed until saturation had been reached. After saturation had been reached at 60 questionnaires, 10 more questionnaires were checked to make sure no new information was missed. After that, all remaining questionnaires were screened for new information to ensure saturation.
Descriptive statistics were used to report characteristics of the respondents. We used Microsoft Word and Microsoft Excel to create the codebook and themes.

Respondents
We received 129 questionnaires from practitioners. One practitioner did not work in a nursing home at the time of the questionnaire and was therefore excluded. One respondent was a medical student and therefore excluded. Respondent characteristics (N ¼ 127) are shown in Table 1. Answers to the multiple-choice questions regarding the cases described by the respondents (N ¼ 127) are shown in Table 2.

Advance Care Planning
The respondents' answers to the questionnaire could be arranged in 4 themes. The 4 themes are as follows: reasons for ACP discussions, discussing ACP, topics discussed in ACP, and decision making in ACP. Within these themes, we distinguished how ACP was practiced and what had changed due to the COVID-19 pandemic. We included all cases, without adjudicating whether the case referred to an ACP discussion (regarding a medical problem that might evolve in the future), or to a discussion on actual decision making (regarding a medical problem that is already there). In the following paragraphs, we will elaborate on the respondents' answers on the 4 themes.

Reasons for ACP Discussions
In some cases, the COVID-19 pandemic served as a trigger for an ACP discussion. In other cases, the reasons for the ACP discussion were not related to the COVID-19 pandemic. Generally, ACP was started at an earlier stage because of the pandemic.
Respondents were clear about their reasons for starting an ACP discussion.
Some described a case in which the reason to start this discussion was not affected by the COVID-19 pandemic. Those reasons included admission to a nursing home, a biannual multidisciplinary meeting, acute illness, or decline in health, including imminent death.
[The reason was a] biannually planned multidisciplinary meeting/care plan discussion. [CID 154] Additionally, respondents mentioned 2 reasons for starting an ACP discussion directly related to the COVID-19 pandemic.
First, a (suspected) COVID-19 infection was a reason for starting an ACP discussion.
A discussion about [treatment] policy was held with every COVID-suspected/-positive resident. [CID 57] Second, in case of a patient infected with COVID-19, transfer to a specific COVID-19 unit in the nursing home, imminent death, and lack of agreement between family members concerning treatment policies were reasons for a (renewed) ACP discussion.
She would have to transfer to a COVID-19 unit because of a positive test. Before admission we always discuss the treatment policy with a patient, and we also discuss hospitalization. [CID 143] At the time of the positive test results, Mr.
[X] was not gravely ill yet, but his condition worsened in the following days leading to an expected upcoming death. In the discussion about the treatment policy [we] checked whether the spouse still supported previously made agreements. [CID 207] Generally, there was a tendency to start an ACP discussion in an earlier stage during the COVID-19 pandemic.
Quicker policy agreements have been made with clients and family. [CID 175] However, some respondents explicitly mentioned that their decision making did not change at all, for example, as they were very keen on early anticipatory medical care decision making beforehand.
[The anticipatory decision making is generally] not [influenced by the COVID-19 pandemic], because I already am a very cautious physician and make agreements on policy as restrictive as possible. [CID 234]

Discussing ACP
The theme discussing ACP concerns contact with surrogates, mode of conversation with surrogates, and the decision-making process.
Visiting restrictions were implemented to minimize the traffic on nursing home units and thus reduce the introduction of COVID-19.
Owing to the visiting restrictions, frequent face-to-face contact with surrogates was impossible. On top of that, respondents indicated that they themselves visited nursing home units only when this was strictly necessary. Our respondents indicated these restrictions had 2 implications.
The first implication was that ACP discussions with surrogate decision makers more commonly took place by phone or video calling. When having the ACP discussion with the patient themselves, it was a bedside discussion frequently using full personal protective equipment. Sometimes, surrogates and other family members joined that discussion by phone. Respondents mentioned upsides and downsides of this way of communicating. The upsides of having ACP discussions by (video)phone were the ability to speak to more family members simultaneously, and the convenience of scheduling calls, as compared with face-to-face meetings. Over time, respondents became more experienced in making decisions by phone or online.
[Contact by telephone] is relatively easy and quick, so easy to do 'in-between.'" [.] [CID 26] Important downsides of contact by (video)phone, mentioned by respondents, were fragmentation of the decision-making process, lack in nonverbal communication, and difficulty in dealing with strong emotions.
Nonverbal communication is missed during contact by telephone. The fact that you are unable to communicate with the family face-toface and unable to communicate with the resident at the same time creates a barrier. Therefore, I wait a little longer before I start the discussion. The natural moments to discuss policy, such as during an evaluation, are not there at the moment. [CID 48] The daughter was emotional. That is easier to support in a faceto-face discussion. [CID 48] Another mentioned downside of having ACP conversations by (video)phone was the difficulty to assess a patients' situation by physicians and family members, as they could not visit the patient. The nursing staff still had a day-to-day interaction with the patient, so  The second implication was that because of the visiting restrictions, all multidisciplinary meetingsdin which ACP was routinely performeddwere postponed. Our respondents mentioned this as an impeding factor, as ACP could not be performed regularly as usual.
We are further removed. . . . By not holding the multidisciplinary meetings it is also less talked about. [CID 72]

Topics Discussed in ACP
For an overview of the topics discussed in ACP discussions, see Table 2.
Our respondents mentioned 2 topics that were added to their conversation because of the COVID-19 pandemic: intensive care unit admission (with or without mechanical ventilation) and treatment preferences in case of a COVID-19 infection. Most respondents mentioned to be restrictive in admitting COVID-19 infected nursing home residents to a hospital.

Factors Influencing ACP
Five factors were mentioned as influencing ACP. The first 2 factors were unaffected by the COVID-19 pandemic. The last 3 factors were influenced by the COVID-19 pandemic.
The first factor of influence was the patients' frailty and prognosis.
Resident was already gravely ill, chance of recovery was minimal in a patient who was in a fragile state beforehand (advanced dementia, moderate intake of fluids and nutrition, fully ADLdependent, passive lift transfers, passive wheelchair) etc. Third, visiting allowances during terminal illness were more lenient in nursing homes than in hospitals during the first wave of the COVID-19 pandemic in the Netherlands.
What also played a role was that patients could not receive visitors there [eg, the hospital] and that we were allowed that in case of terminal illness. [CID 25] Fourth, ACP was influenced by (ideas regarding) treatment options, availability, and risks of medical care in nursing homes and hospitals. Most respondents were more reserved regarding hospitalization in COVID-19 infections, but some surrogates felt treatment was being withheld.
[Daughter felt that] her mother was withheld treatment, she would have preferred that chloroquine had been tried, despite limited evidence and risks concerning comorbidity. Daughter considered hospital an option. [ If patients need to be hospitalised, this still happens. You do notice that patients themselves are more cautious concerning hospitalization because they are scared of contagion. Previously, [treatment] policy was completely active, just no resuscitation or ICU admission. There was a complete understanding for no [hospital] admission: spouse had seen the misery after the "recovery" from COVID. Administering fluids by use of hypodermoclysis in the nursing home was discussed. [CID145]

Outcomes of ACP
Respondents considered anticipatory medical care decisions as ultimately made by physicians. Respondents acted more directive regarding physician treatment orders concerning COVID-19 and put emphasis on the best interest in their regard.
I could not see myself admitting this woman. The fuss of the hospital, she would not understand anything because of her deafness and face covering PPE [personal protective equipment], I could see her getting scared and she is someone of whom we were convinced that it made medically no sense to admit her. I could already hear the hospital doctors complaining and asking why I am admitting her. I also considered it irresponsible to have this vulnerable woman occupy a bed when it was needed for people with a higher chance of survival due to the increasing number of COVID patients and the imminent overcrowding of hospital and ICU beds. [ Our respondents indicated that the majority of long-term care patients already had a do not resuscitate (DNR) order, which remained unchanged. Advance care decisions concerning curative, palliative, or symptomatic care were generally made in a similar fashion as before the COVID-19 pandemic.
The resident had dementia. This played a large role in the anticipatory medical decision making. Hospitalization would have been very confusing for her, and add very little. She already had a no resuscitation policy, and admission to the ICU was not desired anyway. Because other than that she was relatively healthy, it was decided that treatment possibilities within the nursing home would be used, such as medication, oxygen or temporarily administering fluids. [CID 184] However, in many instances, physician treatment orders regarding a potential COVID-19 infection were added. For example, in patients with a curative care policy including hospitalization, there was an exception for hospitalization and ICU admission in COVID-19erelated health problems.
Because of COVID-19, it was agreed upon that hospitalization would not happen. While for other diagnoses hospitalization could still be indicated. [CID 143] Symptomatic while experiencing COVID, after that a new discussion. [CID 49]

Discussion
In this study, we explored how physicians working in Dutch nursing homes practiced ACP during the COVID-19 pandemic, and whether this had changed compared to before the COVID-19 pandemic. Four themes emerged from the data: (1) reasons for ACP discussion, (2) discussing ACP, (3) topics discussed in ACP, and (4) decision making in ACP (factors influencing ACP, outcomes of ACP), in which both continuity and change due to the COVID-19 pandemic were mentioned. In summary, additional reasons for ACP discussions emerged, including COVID-19 infection and the formation of COVID-19 units. ACP discussions took place at an earlier stage, whereas social distancing measures affected the frequency and mode of communication between physicians, patients, and their surrogates in both positive and negative ways. The topics discussed in ACP were expanded with ICU admission and specific policies in case of COVID-19 infection. The decision-making process was influenced by visiting policies within nursing homes and hospitals and media coverage, whereas physicians were more directive in anticipatory medical care decisions.
Changes in ACP during the first wave of the COVID-19 pandemic are the result of both the medical implications of COVID-19 for nursing home residents and nationwide visitor restrictions in nursing homes. 5,15 The strict visitor restrictions imposed limitations on the assessment of the medical situation and quality of life of the nursing home resident by the surrogate decision maker. 13,14 Our study shows that this also impacted ACP and challenged medical decision making. Because of public discontent regarding the loss of quality of life caused by the strict visitor limitations, visitors were allowed during the second wave of the COVID-19 pandemic. 16,17 These measures might play a lesser role in ACP in the future course of the COVID-19 pandemic.
The media coverage on the COVID-19 pandemic facilitated ACP discussions regarding anticipatory medical care decisions in several ways. First, our study shows that surrogate decision makers were generally well informed concerning the poor prognosis of COVID-19 for frail older people and receptive for ACP discussions. Also, our respondents, elderly care physicians, noted they started discussing ICU admission more regularly during the COVID-19 pandemic. Other studies have shown that before the COVID-19 pandemic, elderly care physicians hesitated to discuss ICU admission in advance and that during the COVID-19 pandemic, anticipatory medical care was discussed more often. 18,19 Our respondents indicated that anticipatory medical care decisions were made in the patients' best interest. This is corroborated by a study, which also points out the difficulties in assessing the best interest and risk of making decisions that are not in the patients' best interest during a pandemic. 20 Factors of influence were as follows: the patients' frailty and prognosis, the patients' personal convictions and behavior, and perceived individual benefits and risks of different treatment options and treatment settings. These factors also played a role before the COVID-19 pandemic. 21 The COVID-19 pandemic increased surrogate's awareness of the possibility to forego burdensome life-prolonging treatments with little potential benefit. Also, the more lenient visitor restriction policies in nursing homes during terminal illness were taken into consideration. Some respondents mentioned pressure on hospitals and ICU as a factor, probably because they became more aware of their gatekeeper function during the pandemic. 22 The results from this study underline the importance and fluidity of ACP discussions in nursing homes so as to facilitate discussions on the right care for the individual patient. This is an endorsement for the advancement of ACP worldwide in care for older patients and nursing home residents.

Strengths and Limitations
Our study has several strengths. First, by distributing the questionnaire via the professional newsletter of Verenso, we were able to potentially reach nearly all of the elderly care physicians in the Netherlands. We surveyed many elderly care physicians simultaneously, despite social distancing measures and the high workload of elderly care physicians during the COVID-19 pandemic.
Second, because the questionnaire was sent out at the end of the first wave of the COVID-19 pandemic, the described ACP discussions for nursing home residents in the Netherlands took place during the COVID-19 pandemic. Third, because of the timing (during the pandemic) and the study design (description of the most recent ACP discussion), there was a limited recall bias. Finally, saturation had been reached after analyzing half of the questionnaires, indicating that all relevant topics were explored.
There are also some limitations. First, because of the use of a questionnaire, we were not able to respond to the answers of our respondents and ask further questions, although the questionnaire was open-ended with ample possibility to elaborate on the topics. Second, the results are based on self-reported answers with the risk of limited self-reflection. Third, despite pilot-testing the questionnaire, some respondents had misunderstood the questions and we could see that theirs answers did not relate to the question that we meant to ask.

Conclusions and Implications
This study shows that ACP in Dutch nursing homes has changed as a result of the COVID-19 pandemic. In ACP discussions by elderly care physicians, ICU admission has become a regular topic of discussion. Awareness of the impact of ICU admission for nursing home residents might have a lasting effect on ACP. There was less face-to-face contact with surrogate decision makers owing to visitor restrictions and social distancing measures. This prompted the use of technologies for frequent, low-threshold, informal contact between physicians and surrogate decision makers. Maintaining these practices after the COVID-19 pandemic may foster mutual understanding and aid the decision-making process. You indicated that you were not a practitioner in a nursing home during the COVID-19 pandemic. You are not required to fill in this questionnaire. Thank you for your interest.

Section 4
We would like to ask you to look back on the way you made anticipatory medical decisions concerning residents of nursing homes during the COVID-19 pandemic.
Imagine the last case in which you intentionally started the conversation on anticipatory medical decisions.
What was the treatment policy in this case? , Decisions regarding resuscitation , Decisions regarding submission to the hospital , Decisions regarding submission to the ICU , Decisions regarding use of antibiotics , Decisions regarding the artificial administration of nutrition and fluids , Other decisions, such as_____ , None of the above Explanation_____ On which ward did the patient in this case reside?
B Primary care temporary stay for recovery B Geriatric rehabilitation B Geriatric psychiatric B Hospice B Psychogeriatric B Somatic B Other, namely_____ Who were the participants of the conversation? Also name the participants who were there remotely (via videocall, or phone call, etc).
, Resident , Resident's next of kin Specify_____ , Carer/nurse , Other practitioners Specify_____ , Others Specify_____ Describe the last case in which you intentionally started the discussion on anticipatory medical decisions.
What was the reason for the discussion? How was the discussion conducted 'technique wise'? Did communication happen via telephone, or, for example, a videocall?
How did the means of communication mentioned above affect the process of anticipatory medical decision making?
There will now follow a few questions about the factors that possibly played a role in making anticipatory medical decisions in the case.
In what manner did the medical situation of a resident play a role in the anticipatory medical decision making?
Please provide an explanation for your answer In what manner did the personal preference of the resident play a role in the anticipatory medical decision making?
Please provide an explanation for your answer In what manner did the personal preference of the resident's next of kin play a role in the anticipatory medical decision making?
Please provide an explanation for your answer In what manner did other professionals (care practitioners, etc) play a role in the anticipatory medical decision making?
Please provide an explanation for your answer In what manner did COVID-19 related factors play a role in the anticipatory medical decision making?
Please provide an explanation for your answer Section 5 In the questions above, we asked you to describe one case.
To what extent do you think the anticipatory medical decision making in this case was affected by the COVID-19 pandemic?
Please provide an explanation for your answer To what extent do you think the general anticipatory medical decision making in your daily practice was affected by the COVID-19 pandemic?
Please provide an explanation for your answer

Section 6
You have now reached the end of this questionnaire. We thank you kindly for your answers.