FormalPara Take-home message

Arguments put forward by doctors and families fell under one of ten main types and the types of arguments presented by families largely overlapped with those presented by doctors. A real exchange of arguments occurred in a minority of conversations and was generally quite brief in the sense that not all possible arguments were presented and then discussed together.

Introduction

In neonatal (N-ICU), pediatric (P-ICU), and adult (A-ICU) intensive care units (ICUs), decisions about continuation or discontinuation of life-sustaining treatment (LST) are made nearly every day [1, 2]. These decisions require good communication between doctors and families and involve careful consideration of the pros and cons of the (still) possible treatments [3,4,5]. As recent studies have underlined, argumentation in the sense of solid reasoning is one of the intrinsic parts of the decision-making process [6,7,8,9,10,11,12,13]. Professional guidelines not only recommend an open exchange of standpoints and underlying arguments within the medical team, but also between doctors and families to reach the most appropriate decision for individual patients [14,15,16,17,18,19,20,21,22,23,24,25]. Decisions about LST have major consequences for patients and their families. These decisions are particularly difficult to make when they take place in the gray zone, i.e., in situations in which there is incomplete knowledge about the relative harms and benefits of the remaining treatment options [27]. Decisions in the gray zone require an even more solid exchange of arguments by all parties involved. It is thus far not investigated how doctors and families argue in real-life conversations in the ICU. Studying these actual argumentation practices will provide insight into how decisions are reached, and into missed opportunities and best practices.

In this qualitative observational study, we thoroughly explored doctors’ and families’ argumentation in conversations regarding the decision to continue or discontinue LST in the N-ICU, P-ICU, and A-ICU [28]. We especially focused on decisions in the gray zone. We aimed to (1) identify which arguments doctors and families of critically ill patients use in support of standpoints to continue or discontinue LST (argumentation content), (2) investigate how doctors and families structure their arguments (argumentation structure), and (3) explore how doctors’ and families’ argumentative practices unfold during conversations (argumentation dynamics).

Methods

Design and setting

This study is part of the FamICom-project in which different aspects of doctor–family communication in the ICU are studied, such as family involvement, conflict management, and uncertainty [3,4,5, 28]. This qualitative explorative study focused on the real-life argumentative practices of both doctors and families regarding the decision to partly or fully continue or discontinue (i.e., withhold or withdraw) LST (henceforth: to continue or discontinue LST), using inductive thematic analysis. This approach was chosen, because argumentative practices within the ICU context have not yet been studied and guiding theories are therefore lacking [29, 30]. For the sake of clarity, we quantified the incidence of the main types of arguments. Data were derived from audio recordings of team–family conversations in the N-ICU, P-ICU, and A-ICU of the Amsterdam University Medical Center (UMC). ‘Families’ refers to the relatives or close friends who were present during these conversations.

Population and sampling

Seventy-one doctors and the families of 36 patients participated in our study. We strived for maximum variation regarding patients’ age, gender, diagnosis, disease progression, and course of treatment, families’ ethnic background, and doctors’ gender, medical specialty, and role.

Recruitment

Prior to data collection, all doctors and nurses from the participating units received oral and written information and were asked for their consent to participate.

Data collection

The inclusion period lasted from April 2018 to December 2019. Families of patients were eligible to participate from the moment when doubts were expressed by the medical team and/or the family whether continuing LST was in the patient’s best interests. Only conversations in Dutch or English were included. Conversations in other languages or with an interpreter were excluded.

The attending doctor or nurse introduced the study to eligible families. Families who were interested in participating were further informed and then asked for their oral and written consent by a member of the research team or the attending doctor.

From the moment of inclusion, all conversations between the medical team and families were audio-recorded until a final decision was reached to either continue or discontinue LST.

Data analysis

The audio recordings were transcribed verbatim and anonymized. Transcripts (n = 101) were then uploaded to MaxQDA 2020 [31]. Coding and analysis consisted of four phases, as illustrated in Fig. 1. In developing our coding scheme, we were inspired by the pragma-dialectical argumentation theory [32]. In line with this theory, we differentiated between standpoints, main arguments, and sub-arguments [32]. Standpoints are defined as the positions taken toward a specific proposition. Our study focused on two standpoints: the standpoint that LST should be continued and the standpoint that LST should be discontinued. Main arguments are defined as arguments directly substantiating a standpoint, whereas sub-arguments indirectly underpin a standpoint. According to the pragma-dialectical argumentation theory, in an ideal situation, all relevant standpoints should be presented in a discussion and substantiated with relevant arguments and sub-arguments [32].

Fig. 1
figure 1

Four phases of coding and analysis

Full size image

Ethics

The Amsterdam UMC institutional review board waived approval of this study (W17_475 # 17.548). All participants could withdraw their consent at any time.

Results

Seventy-one doctors and the families of 36 patients participated in our study. Table 1 lists patients’, relatives’, and doctors’ characteristics. All doctors and all but one nurse from the N-ICU gave their consent and participated. All but one of the approached families agreed to participate, the latter because they found participation too burdensome at that moment.

Table 1 Main characteristics of included patients, relatives, and doctors
Full size table

In most cases, doubts whether LST was still in the patient’s best interest were first uttered by the medical team. Only occasionally, families were the first to voice their doubts. In the A-ICU, it took between one and three conversations before a final decision to either continue or discontinue LST was reached, whereas in the N-ICU and P-ICU, this took between one and nine conversations.

In Fig. 2a, b, we present the number of cases and audio-recorded conversations in which doctors and/or families put forward a standpoint and/or arguments.

Fig. 2
figure 2

a Number of cases in which doctors and/or families put forward a standpoint and/or arguments in support of a standpoint. b Number of audio-recorded conversations in which doctors and/or families put forward a standpoint and/or arguments in support of a standpoint

Full size image

Argumentation content

Our analysis resulted in a complete overview of all arguments doctors and families of critically ill patients used in support of standpoints to continue or discontinue LST. These arguments could be classified under one of ten main types. These main types of arguments could be divided into four clusters, namely treatment-related (types 1–4), patient-related (types 5–7), family-related (types 8–9), and ‘deviating’ (type 10) types of arguments. Table 2 provides a description of the ten main types of arguments, complemented by illustrative quotes.

Table 2 Main types of arguments and illustrative quotes
Full size table

Argumentation structure

We further explored how doctors and families structured their argumentation. An argumentation structure refers to the relation between a standpoint, main arguments, and sub-arguments. Examples of such relations in our study are visualized in supplementary A and B.

The standpoints focused on in our study were the standpoint that LST should be continued and the standpoint that LST should be discontinued. Occasionally doctors presented both standpoints in the same conversation. In most conversations, they put forward only one of these standpoints.

Doctors in the N-ICU and P-ICU mainly presented the standpoint to discontinue LST, also in the initial conversations. Doctors in the A-ICU slightly more often presented the standpoint to continue LST in the initial conversations. In later conversations, they predominantly presented than the standpoint to discontinue LST.

Looking into the standpoints presented by families, parents in the N-ICU mainly appeared to advocate the standpoint to continue LST in the first and over subsequent conversations. In the P-ICU, in the initial conversations, most parents put forward the standpoint to continue LST, whereas in later conversations, parents largely presented the standpoint that LST should be discontinued. Families in the A-ICU mainly argued in support of the standpoint to discontinue LST in the first and over subsequent conversations.

Main arguments and sub-arguments

In the vast majority of conversations, both doctors and families substantiated their standpoints with one or more main arguments. These main arguments were underpinned with sub-arguments in a minority of conversations. Table 2 provides a description of the ten main types of arguments.

If doctors or families argued for continuation of LST, they mostly presented one main argument, for example: “We should continue LST, because there is still too much uncertainty to stop treatment”. In the conversations in which doctors presented more than one argument to continue LST, this mostly concerned arguments referring to generally accepted healthcare standards and uncertainty, as illustrated by the next quote: “The test that we normally do in this kind of situation, showed that she responded to the stimuli. We therefore can’t rule out that she will wake up”. In conversations in which families presented more than one argument in favor of continuing LST, these arguments often concerned the following types: family’s moral responsibility and psychological wellbeing of families. This is illustrated by the next quote: “She must have this last chance, because otherwise I will always keep wondering ‘what if’”.

Doctors often substantiated their standpoint that LST should be discontinued with two or more main arguments. They mostly combined arguments regarding the following four types: patient’s quality of life, medical (in)effectiveness, professional authority, and/or proportionality. For example: “We should not give CPR, because it will not result in a well-functioning heart. On the contrary, it will cause more damage and this extra damage will be too much for her”.

In the few conversations in which families presented more than one argument in favor of discontinuation, these arguments concerned the patient’s quality of life in combination with their own moral responsibility, to illustrate: “We [as her family] must take into account her remaining quality of life and happiness”.

Because doctors and families often used only one main argument and infrequently substantiated their main arguments with sub-arguments, their line of arguing appeared to be rather abstract and vague. For example, doctors regularly argued that LST should be discontinued because “the treatment is not proportional anymore” without elaborating or explaining this (dis)proportionality. Similarly, families for instance argued that LST should be discontinued, because the patient “suffered”, without elaborating on their observations and without being invited to do so by the doctor. Despite the vagueness of the argumentation, doctors seldom asked open questions to explore what families meant. Families in turn did also seldom ask for further clarifications.

Argumentation dynamics

In almost all conversations, doctors were the first to argue and families followed, thereby either countering the doctor’s line of argumentation or substantiating it. Often, families did not immediately respond to the arguments presented by the doctor but did so much later in the conversation. Only in a minority of conversations, a true exchange of arguments (i.e., doctors and families responding to each other’s arguments by complementing or contradicting the other’s arguments) took place. This was mainly the case within the N-ICU. The exchange of arguments was generally quite brief in the sense that not all possible types of arguments were presented and then discussed together.

In response to doctors’ standpoint to discontinue LST, most parents in the N-ICU plead for continuation of LST and brought in new arguments. In their subsequent response, most doctors further substantiated their standpoint to discontinue LST with additional arguments. Still, the outcome of most exchanges was that LST would be continued for the time being. In the P-ICU, families argued as often in line with doctors’ previously presented standpoint as in the opposite direction. When they agreed with the doctor’s standpoint, they often broadened or deepened this standpoint by providing new or additional arguments. For example, a doctor argued that LST should not be continued, “because it is not in your child’s best interest”. In response, the parents put forward a new argument, namely that their child “suffered visibly, because she was often gasping for air and was not smiling as often anymore”. Families in the A-ICU mainly argued in line with the doctor’s standpoint to discontinue LST. They less often provided additional arguments than parents in the N-ICU and P-ICU.

In the conversations in which doctors and families did not exchange arguments, they presented their arguments as stand-alone arguments instead of being part of an overall argumentation structure. These conversations had two different outcomes. First, the treatment option that the doctor suggested was followed. Second, if families fiercely (i.e., repeatedly and/or extensively) argued for continuation of LST, a decision was postponed and a follow-up conversation was planned.

Almost none of the conversations ended with a kind of meta-discussion in which all exchanged arguments were summed up and weighed before an overall conclusion was drawn.

Discussion

We thoroughly explored which arguments doctors and families used to either argue that LST should be continued or discontinued, how they structured their arguments, and how their argumentative practices unfolded during conversations. In almost all conversations, doctors were the first to argue and families followed, thereby either countering the doctor’s line of argumentation or substantiating it. Arguments put forward by doctors and families in the conversations we analyzed fell under one of ten main types. The types of arguments presented by families largely overlapped with the types of arguments presented by doctors. A real exchange of arguments occurred in a minority of conversations and was generally quite brief. If families did respond to the arguments presented by the doctor, they often did so much later in the conversation.

The arguments presented by doctors in our study appeared to be largely consistent with recent professional guidelines in their content [15,16,17,18,19,20,21,22,23,24,25,26]. This consistency may be related to the highly protocolized care for critically ill patients over the last decades, including end-of-life decision-making. Then again, guidelines are adapted in response to actual (practice) developments within critical care. In that regard, it is of interest to see which of the arguments we identified are not represented in the current guidelines. We found two novel arguments, namely those referring to uncertainty and to professional authority.

The first novel argument type, referring to uncertainty, was frequently put forward by both doctors and families in all ICUs to plea for continuation of LST. This can partly be explained by the fact that technical innovations increasingly enable ICU teams to keep patients alive, in theory for as long as the patient or the patient’s legal representatives wish for—a few exceptions notwithstanding. Yet, the patient often pays a high price for his survival, including lasting severe disabilities and a permanent loss of independency. In this light, the patient’s quality of life in the short and long term has become a more and more decisive factor in making decisions about the continuation or discontinuation of LST. However, the assessment of the remaining quality of life is much more prone to uncertainty. Recent studies in which families of critically ill patients were interviewed also show that uncertainty is an important argument for families to ask for continuation of LST, at least until there is more certainty about prognosis and treatment effects [4, 33,34,35,36,37,38,39,40,41,42]. Studies also reveal that adequately discussing uncertainties improves both the decision-making process and the wellbeing of patients and their families [33,34,35,36,37,38,39,40,41,42]. Based on these findings, we recommend including uncertainty as a topic for discussion in professional guidelines concerning end-of-life decision-making in the ICU. This will not reduce the amount of uncertainty itself. Yet, addressing this issue in conversations with families may help them to better understand the recommended line of action and to come to term with their decision [4, 43].

The other novel argument, professional authority, can be regarded as deviating in the sense that one could question whether it is an argument or a persuasion technique [44]. That is, presenting a treatment option as an authorized decision suggests that a group of experts regards a certain option as the best option, and families of patients might not feel comfortable to disagree with the perceived best option [44]. In that sense, professional authority arguments may limit family participation in the decision-making process [44]. Arguments based on generally accepted healthcare standards and on comparisons can also function as participation limiting due to their steering character. That is, families may fear that they will not be able to forgive themselves in case of a negative outcome if they went against generally accepted healthcare standards [44]. Comparisons can steer toward a particular treatment decision, because they may impede the evaluation of the pros and cons of each option available [44]. Despite these drawbacks, references to professional authority, generally accepted healthcare standards, and comparisons may have an important function in conversations with families. These three types of arguments can underline that what the doctors do and say is based on professional, evidence-based standards. If worded clearly and empathically, this may well help families to understand and accept the suggested course of action. On the long term, it will reduce families’ feelings of doubt and guilt that treatment has either been withdrawn too early or too late [4]. These three argument types show that there is a thin dividing line between argumentation and persuasion [5, 44, 45]. Doctors should be aware of this and provide room for weighing of arguments. Doing so enables open, meaningful, and insightful discussions which stimulates well-considered and appropriate treatment decisions.

Regarding the argumentation structure and dynamics, doctors presented only the standpoint to continue LST or the standpoint to discontinue LST in most conversations. This was unexpected, since most of the conversations concerned decisions in the gray zone. These decisions are characterized by the fact that there is no clearly best treatment option, because all remaining treatment options have their pros and—often very serious—cons. In that light, it is important to explicitly discuss both options, including these pros and cons [27, 32]. It is important that doctors guide this process well and to maintain the right balance between inviting families to participate in the decision-making process and not overburdening them with too much responsibility.

We frequently observed a lack of concrete explanation and deepening of arguments and of summarizing all exchanged arguments. This raises the question whether doctors and families really understood each other’s arguments, especially since both parties seldom asked clarifying questions. This may add to mutual misunderstandings which in turn may well result in conflicting views between doctors and families [5, 46, 47]. Former studies have shown that true deliberation and a complete and extensive exchange of arguments and information will enhance both family satisfaction and the quality of the argumentation process, and by effect the treatment decision that is reached [6,7,8,9,10,11,12,13,14, 44, 48, 49]. Nevertheless, communication is much more than argumentation alone: it implies the building of true relationships, too. Communication requires synchronizing (non-)verbal messages between healthcare staff and families for the benefit of the patient and the family as a whole [3,4,5]. As each patient and family has their own communication needs and preferences, it is important for healthcare professionals to tailor their communication to those needs and preferences [3].

Limitations and strengths

We have pushed for maximum variation regarding the participating patients, their families, and their doctors. Still, selection bias may well have occurred. This limits the generalizability of our study. Another limitation is that we only audio-recorded conversations to minimize the intrusiveness of our data collection, which precluded the analysis of non-verbal communication. Moreover, we cannot rule out that the Hawthorne effect may have occurred due to the audio-recording of conversations [50]. Finally, this study describes the argumentative practices of doctors and families within one medical center only, which was unavoidable given the logistical demands of the study. It would be relevant to expand this study to other ICUs in the Netherlands and, preferably, to ICUs in other countries, thereby also multiplying the number of cases. This will enable investigating contextual and cultural influences on the argumentation practices of healthcare staff and families.

The main strength of this study is that we collected a large dataset of transcripts of real-life, highly sensitive conversations which we meticulously analyzed. Furthermore, we investigated the actual argumentation practices in three different ICUs.

Practice recommendations

Doctors’ awareness of argumentation in end-of-life conversations may add to open and insightful discussions and to well-considered and appropriate treatment decisions. We present some hands-on recommendations in Fig. 3. Future research is needed to investigate how and to what extent best practice recommendations improve families’ and doctors’ experiences of the decision-making process, including their satisfaction with the communication.

Fig. 3
figure 3

Recommendations for doctors

Full size image