The clinical practice guideline palliative care for children and other strategies to enhance shared decision-making in pediatric palliative care; pediatricians’ critical reflections

Abstract

Background

Because of practice variation and new developments in palliative pediatric care, the Dutch Association of Pediatrics decided to develop the clinical practice guideline (CPG) palliative care for children. With this guideline, the association also wanted to precipitate an attitude shift towards shared decision-making (SDM) and therefore integrated SDM in the CPG Palliative care for children. The aim was to gain insight if integrating SDM in CPGs can potentially encourage pediatricians to practice SDM. Its objectives were to explore pediatricians’ attitudes and thoughts regarding (1) recommendations on SDM in CPGs in general and the guideline Palliative care for children specifically; (2) other SDM enhancing strategies or tools linked to CPGs.

Methods

Semi-structured face-to-face interviews. Pediatricians (15) were recruited through purposive sampling in three university-based pediatric centers in the Netherlands. The interviews were audio-recorded and transcribed verbatim, coded by at least two authors and analyzed with NVivo.

Results

Some pediatricians considered SDM a skill or attitude that cannot be addressed by clinical practice guidelines. According to others, however, clinical practice guidelines could enhance SDM. In case of the guideline Palliative care for children, the recommendations needed to focus more on how to practice SDM, and offer more detailed recommendations, preferring a recommendation stating multiple options. Most interviewed pediatricians felt that patient decisions aids were beneficial to patients, and could ensure that all topics relevant to the patient are covered, even topics the pediatrician might not consider him or herself, or deems less important. Regardless of the perceived benefit, some pediatricians preferred providing the information themselves instead of using a patient decision aid.

Conclusions

For clinical practice guidelines to potentially enhance SDM, guideline developers should avoid blanket recommendations in the case of preference sensitive choices, and SDM should not be limited to recommendations on non-treatment decisions. Furthermore, preference sensitive recommendations are preferably linked with patient decision aids.

Background

It is estimated that 750–1000 children with a life-threatening illness die each year in the Netherlands and 5000 children are diagnosed with a life-threatening illness [1]. Of the 1200 Dutch pediatricians, 60 pediatricians (being part of pediatric palliative teams in university medical centers) regularly encounter children with a life-threatening disease. Most Dutch pediatricians – and other healthcare professionals – infrequently encounter children who are in need of palliative care, making them less sure how to proceed in these cases. Moreover, a life-threatening illness has an enormous impact on both the parents and the child. The pediatrician can help in coping, and counseling on treatment related issues as well psycho-social aspects should start as early as possible. Furthermore, new developments in pediatric palliative care and limited knowledge thereof among pediatricians led to unwarranted practice variation [1, 2]. The Dutch Association of Pediatricians therefore deemed it necessary to develop the clinical practice guideline (CPG) Palliative care for children to improve the quality of care for these children. The objective of this guideline is to provide pediatricians support and guidance when caring for children with a life-threatening disease. The guideline was published in the summer of 2013 [3].

The introduction of the CPG Palliative care for children stresses the importance of knowing and discussing the needs of the patient and his or her parents. In addition to palliative intervention recommendations, the guideline consequently also contains recommendations on communication and shared decision-making (SDM) (see Table 1) [3]. The underlying assumption of SDM models is that the doctor-patient relationship is based on mutual respect and on a joint interest to achieve the most beneficial outcomes for the patient, with partners sharing decisions. In this process it is important to understand the patients’ preferences and values, and to acknowledge the doctor’s expertise [4]. However, it is still unclear if and how SDM can be incorporated in CPGs [5,6,7,8,9]. Van der Weijden et al. describe how guidelines can be adapted to facilitate SDM, e.g. by incorporating patient-centered questions, flagging recommendations for which considering patient preferences is particularly important, or by embedding relevant patient support tools [5]. The Institute of Medicine recommends including patients in the guideline panels to address heterogeneity in patient preferences and include trade-offs in recommendations [10]. Prior to 2013, only a few guidelines tried to do this [11,12,13].

Table 1 Quick overview of the CPG Palliative care for children

Sharing decisions in pediatric care is triadic in nature as parents or other care givers may be as much or even more involved as the child [14]. Furthermore, in the Netherlands children have a right to be involved in deciding about their care from the age of 12 years [15]. A majority of both parents and children prefer to participate in treatment decision-making and consider it an essential component of quality of care, together with relationship building, demonstration of effort and information exchange [16, 17]. Despite the large number of publications in support of children’s (ethical) right – and desire – to participate in their own health care and decisions, pediatricians usually do not involve them in the decision-making. One reason being the clinicians’ underestimation of children’s capabilities; because of the children’s age, pediatricians think they cannot or do not want to be involved in the decision making [18, 19]. A national survey (2011) among Dutch pediatricians showed that 21% perceived a need to share the final responsibility for an end-of-life decision with the parents. Half of the respondents would inform parents and ask for their permission to discontinue treatment, and a quarter would merely inform parents [20]. Reasons for not sharing the decision were: complexity/uncertainty of the decision at hand, and/or protecting the parents. Another important aim of the guideline Palliative care for children was therefore to trigger an attitude shift towards SDM in palliative pediatric care.

The aim was to gain insight if integrating SDM in CPGs can potentially encourage pediatricians to practice SDM. Its objectives were to explore pediatricians’ attitudes and thoughts regarding (1) recommendations on SDM in CPGs in general and the guideline Palliative care for children specifically; (2) other SDM enhancing strategies or tools linked to CPGs.

Methods

We applied the Standards for reporting qualitative research (SRQR) and Consolidated criteria for reporting qualitative studies (COREQ) for this article [21, 22].

Study design

We conducted semi-structured, face-to-face interviews to gather ‘deep’ information and perceptions of the participating pediatricians using a phenomenological approach [23, 24]. This qualitative research method emphasizes the importance of personal perspective and interpretation. It is a suitable method for gaining insight into participants’ motivations and actions and understanding subjective experience.

Researchers’ characteristics

The second author (LV) conducted the face-to-face interviews, with the assistance of JW and NW. LV, NW and JW were sixth year medicine students carrying out a science research internship at Maastricht University. There were no prior relationships between the participants and any of the authors. We informed the participants of the researchers’ background and their roles in the PhD-research of the first author (DD), (1) at the time of recruitment, and (2) at the start of the interview. The interviewer (LV) had no previous experience with interviewing. She was trained on the job by TVDW and DD and closely supervised by the research team through regular meetings and feedback on audio-recorded interviews.

Participants

We purposively sampled pediatricians with experience in providing palliative pediatric care. The respondents were also purposeful sampled regarding tenure (professor, specialist, fellow); age; and gender. We recruited pediatricians from five subspecialty divisions: oncology, metabolic diseases, neurology, intensive care and neonatology at three university based pediatric centers in the Netherlands. These centers were UMCG, MUMC+ and Amsterdam UMC location AMC. We tailored the approach to each pediatric center to ensure timely recruitment of participants. Participants were approached via personal mailings sent by the author linked to the specific pediatric center or via personal invitations by one of the authors (LK, EV). The interviews took place at an office in the hospital in the second quarter of 2014. During the interview, only the participant and interviewer were present.

Data collection

Based on the relevant literature and the guideline Palliative care for children (PCFC), an interview guide was developed [25]. Table 1 contains an overview of the (recommendations of the) guideline. The research team discussed the interview guide within the project team and the guide was piloted in two interviews. We followed an iterative approach; meaning that the interview guide was adjusted – if needed – in response to the preceding interviews (see Additional file 2).

During the pilot interviews, it became clear that not all participants might have an in-depth knowledge of the guideline PCFC. Consequently, it was decided that if the pediatrician was not familiar in detail with the guideline that the researcher would walk the pediatrician through the guideline and the SDM-recommendations at the start of the interview.

In addition to the PCFC guideline, all participants were presented with two other SDM-approaches [5]. First, participants were shown two patient decision aids (PDA) for children: one on anticoagulants and another on bone infection [26, 27]. Second, we showed them the PCFC guideline recommendation on pain relief (see Additional file 1) that we had rewritten so that it offered several options for pain relief and could stimulate discussion with the patient and/or parents. Approval for the study protocol was granted by the Medical Ethical Committee of the University Medical Center Groningen, reference number M14.150.681/METc 2014.057 (10 February 2014).

Verbal informed consent was gained from all participants, which included consent for use of anonymized quotes in publications. No follow-up interviews took place. The interviews were audio-recorded and transcribed verbatim by LV, JW and NW, with at least one author carrying out a reliability check on each transcript correcting any transcribing errors. The participants received the transcripts for comments, correction and approval (member check). Furthermore, LV made field notes before and after the interviews, and these were discussed with DD.

Data analysis

Data analysis was performed by using the constant comparative method [23]. Three authors (LV, NW and DD) read and reread all transcripts independently. Text fragments related to the research questions were selected and coded. The research team did not identify themes or codes in advance.

The codes were developed inductively, while reading the transcripts and making notes about selected fragments in the transcripts. After the first five transcripts, the researchers (LV, DD) compared their observations and developed preliminary coding categories. After analyzing individual interviews, we added and altered codes and categories in an iterative process. LV and DD discussed disagreement about the coding until consensus was reached. When all the interviews were analyzed and coded, we collected the codes in a preliminary coding scheme. Through axial analysis, we constructed (sub-) categories. LV and DD discussed and compared all the codes and categories resulting in the final coding scheme (see Additional file 3). Data analysis was supported by NVivo-software. The research team met to discuss and complete the study report and manuscript. We did not invite participants to provide feedback on the findings.

Results

Sample

In total 22 pediatricians were approached to be interviewed and 15 agreed to participate (Table 2). Three were moderately interested and preferred to decline, three did not respond and one pediatrician was not interested in the subject and declined. The participants consisted of two professors, 12 medical specialists, and one fellow. Nine of them had 15 years of experience or more.

Table 2 Characteristics of the interviewed participants

Data collection

The 15 interviews lasted between 20 and 75 min (average 48 min). Data saturation occurred after eight interviews; no new codes emerged from the data.

Main findings

Three main findings became apparent when analyzing the interviews.

• (1). Possibility (or desirability) of CPGs enhancing the practice of SDM;

• (2). Added value of integrating SDM in CPGs in general and the PCFC guideline specifically;

• (3). Usefulness of a PDA accompanying a guideline.

Before describing the main findings, it is important to note that interviewees interpreted the term SDM differently. When asked to describe SDM, half of the participants referred to it as reaching a decision together with other clinicians, e.g. during a (multidisciplinary) team meeting without involving the patient. After providing participants with a commonly used definition of SDM, and briefly discussing the concept, several participants were of the opinion that if there was a clear superior treatment option, the decision should lie with the clinician and there was no need for patient involvement. Almost half of the participants found that patients and/or parents needed to be included only when there was not an obvious best choice (“grey area”), or when decisions were not treatment related, e.g. decisions about where to die or how to say their goodbyes. Furthermore, there seemed to be confusion between sharing decisions and who has the final responsibility for the decisions made. Some participants did not want to burden the patient/parents with the responsibility of deciding on their child’s (end-of-life) care, stating: “The parents have to live with that decision.” (Participant 3).

Possibility (or desirability) of CPGs enhancing the practice of SDM

When asked if CPGs could potentially enhance SDM, the responses were mixed. About half of the participants mentioned that it could be helpful when there is disagreement; to make sure you cover all the relevant topics, including topics you yourself might consider less important, but are valuable to patients; and to ensure completeness of information provided.

Some pediatricians commented that CPGs and SDM do not go together as they regarded CPGs paternalistic in principle and are therefore irreconcilable with SDM. Others thought it could be a good combination (Table 3).

Table 3 Illustrative quotes on possibility (or desirability) of CPGs enhancing the practice of SDM

Another reason why some participants said that CPGs and SDM do not go together was because they regarded SDM an attitude or skill. A pediatrician is either willing or capable to share decisions with a patient or not, and adding SDM as a recommendation to a guideline will not change that attitude. “It is the difference between ‘the art of medicine’ and ‘medicine’ “, according to two participants (2 and 12). Several participants commented that to support SDM as a skill, the recommendations in the CPG needed to be more practical. For example: one of the guideline recommendations is to clarify and collaboratively set goals; but no practical guidance is offered on how to do this in clinical practice. Some participants, however, thought that incorporating recommendations on SDM in a CPG, might make it easier to practice it because e.g. it makes it less scary (Table 3).

SDM integrated in guidelines

Several options exist to address SDM in a guideline. The PCFC guideline opted to add recommendations on (shared) decision-making in the guideline as a separate chapter (see Table 1). The participants were asked how they felt about the innovative section on decision-making in the guideline. A participant noted that the recommendations on (shared) decision-making could help prepare the discussions with the patient/parents, and could help to put the patient in a social context. Several of the participants thought that most of the SDM-recommendations stated the obvious and that experienced pediatricians already practiced this. However, others said the recommendations might be useful for pediatricians who are less experienced or a bit hesitant. One participant indicated that the SDM-recommendations are valid for pediatrics in general and therefore need not be in this guideline. He suggested developing a separate guideline solely on SDM in pediatrics (Table 4).

Table 4 Illustrative quotes on added value of SDM integrated in CPGs

Modifying a recommendation into an SDM-recommendation

When showing the participants, the original single-option recommendation on pain relief and the modified multiple-option recommendation from the symptoms’ section of the PCFC guideline (see Additional file 1), most participants preferred the modified recommendation. The modified recommendation was viewed as beneficial to engage patients/parents in the decision-making by showing considerations that are important to the patient/parents (Table 4). According to one participant the usefulness of the short, single-option recommendation versus the longer, multiple-option recommendation depends on the context, e.g. in acute or neonatal intensive care, the original – shorter – recommendation would be more useful. Views expressed by the participants working in neonatal intensive care reflected this notion.

Usefulness of a PDA accompanying a guideline

Another option to address SDM in guidelines is paring a specific conditional recommendation with a PDA. A conditional recommendation is used when the underlying evidence is scarce or conflicting, or when more than one relevant treatment options is available. For conditional recommendations, it is known that an individual patient values the uncertainties and trade-offs differently compared to other patients [28]. Although most participants perceived the PDA to be beneficial for the patient/parents, and to a lesser degree for themselves. Most of the participants preferred to convey the information mentioned in the PDA in person, out of fear of losing rapport with the patient when looking at a piece of paper or tablet during a consultation. Others believed patients/parents would not want to use PDAs. Some interviewed pediatricians expressed that the PDAs contained useful topics that they would normally not address during consultation, and suggested it could be used as a checklist to ensure important topics were covered. Two participants mentioned that patients/parents often do not remember everything mentioned during a consultation and that PDAs are ideal for patients as a reference document to take home, re-read and reflect on what was discussed (Table 5).

Table 5 Illustrative quotes on usefulness of a PDA accompanying the CPG

Discussion

Several participants acknowledged the added value of SDM being included in the guideline PCFC, and more participants were of the opinion that guidelines in general potentially could enhance SDM. Regarding the specific SDM-recommendations in the PCFC guideline, several participants judged these as stating the obvious and lacking detail and practical guidance. Some even said they were offended by these recommendations, perceiving it as an attack on their professionalism because they already do this. Clinical observations, however, have shown that SDM during patient contacts is not standard practice yet. Clinicians think or say that they practice SDM, but when their interactions with patients are analyzed, it appears that the level of SDM leaves room for improvement [9, 29,30,31,32]. The felt attack could perhaps be unjust because of optimistic bias. Furthermore, research suggests that there seems to be a tendency with clinicians to share non-treatment related decisions with patients, and to share the treatment related decisions to a lesser degree [33, 34]. When it concerns children, this could be expected as parents and healthcare professionals might take a protective stance towards the child. However, the child might prefer to be protected in some situations and wants to share decision-making in other situations. In addition, some children prefer to leave the more ‘serious’ decisions to their parents and healthcare professionals, whereas other children prefer to share the decision [19, 35]. The views expressed by the participants seemed to confirm that mainly the non-treatment related decisions where shared, or the less ‘serious’ ones. Combined with the situation that the clinician ‘decides’ which information to share with a patient, the stage for SDM is not ideal [32, 36, 37].

The intention of the Dutch Pediatric Association was to potentially enhance the practice of SDM in pediatric palliative care and to do so by integrating SDM in the guideline PCFC. Therefore, recommendations on SDM were included in the guideline and the necessity of SDM was explained in a separate chapter. However, these SDM recommendations have not been integrated in the treatment related recommendations. The integration of SDM could be done e.g. by re-phrasing these recommendations to increase option awareness and/or to include patient preferences. Other possibilities are to structure the deliberation process and describe it more explicit in the guideline, and/or providing patient support tools [38]. These tools could be linked to a specific recommendation or to the guideline as a whole [5]. In the current format, the guideline might not encourage the interviewed pediatricians to involve patients/parents when talking about treatment choices which might inadvertently contribute to SDM only being used for non-treatment related decisions in children’s palliative care.

Participants were open to recommendations related to an SDM-approach to treatment decisions (i.e. structuring the options in a recommendation to increase option awareness), as was shown when we discussed the modified recommendation on pain relief. More than half preferred the ‘SDM’-recommendation because it can help open the discussion, it shows underlying arguments for different treatments, such as pros and cons, and it enables patients and/or parents to choose. Recommendations such as these are preferable according to the Institute of Medicine (IoM). In its report ‘Clinical practice guidelines we can trust’, IoM suggests refraining from so-called blanket recommendations: a recommendation for all patients to choose one particular treatment, irrespective of the patient’s characteristics, preferences and values. IoM recommends to describe the options and trade-offs in a recommendation encouraging SDM, as its respects the individual choice. In this way guidelines, according to IoM, become tools for patient engagement and activation [10]. The last couple of years more CPGs are being developed in which recommendations address trade-offs and mention more than one option [6, 39,40,41,42].

Another strategy to adapt guidelines so that they could enhance SDM is imbedding patient support tools – such as PDAs – in CPGs. The majority of participants thought that a guideline (recommendation) accompanied by a PDA would be beneficial to engaging patients and sharing the decision-making. Systematic reviews have shown that the use of PDAs supports patients to engage in deciding about their care [43]. Our interviews suggest that the use of PDAs could also help pediatricians to check if they have covered all the topics (PDA as a checklist) and address issues they would normally forget or not consider important. Another benefit of PDAs mentioned by pediatricians was that patients can take the PDA home and reread it. Patients often do not remember everything when talking with the clinician, and by taking it home they can weigh the pros and cons in a more comfortable and less time-pressured setting. PDAs can also help patients to ask questions, as patients not always dare to ask their pediatrician everything [44, 45]. The pediatricians who were not convinced of the benefit of a PDA claimed that patients did not want to use them [8]. Research has shown that patients do want to be involved in (deciding about) their care, and need to be involved [31, 46, 47] including children and adolescents (and their parents) [48,49,50,51,52,53]. Another reason why some pediatricians were not inclined to use a PDA was that they felt they know best what to advise based on their expertise, experience with the treatments and the patient sitting in front of them. This attitude is risky for three reasons; firstly, well-informed patients who actively participate in their care are more satisfied with their decisions [54]. Secondly, without using a PDA, the clinician will likely not be complete when providing information to the patient [36, 37]. And thirdly, it has also been shown that clinicians are not always correct in predicting what the patient wants [55, 56]. Moreover, the preferences of a clinician, from the perspective of being a patient herself or himself, often do not match with what they would recommend the patient [57]. Instead of projecting their own opinion on their patients, clinicians should ask what the patient prefers; not only when it concerns treatment decisions but also when it comes to being involved and deciding about their care.

Another important finding is that participants were worried about burdening parents with the responsibility of deciding on their child’s end-of-life care if they engaged them in SDM. However, instead of assuming that parents do not want to be burdened with this responsibility, discussing this with the parents, and/or child, acknowledging their autonomy in the decision-making process, seems to be justified and fitting with SDM. Some research has shown that parents want to be the ultimate decision maker for their child [58], and children value autonomous decision-making, without excluding their parents [51, 59]. In the interviews the pediatricians used the words responsibility and sharing the decision-making interchangeably. It appears that the pediatricians are not aware that even when the decision-making is shared with the child and/or parents, they have the final responsibility for the decision. This unclarity might make pediatricians wary of SDM when it comes to sharing decisions.

Regarding the comments made by the pediatricians that SDM is a skill, and that the recommendations in the PCFC guideline mainly addressed the what and not the how of SDM, the guideline could be enriched with examples of SDM in the pediatric palliative care context. For example, Van der Weijden et al. suggest using a vignette describing how a patient and a healthcare provider discuss the options to reach a shared decision, or could provide scripts modelling SDM-language. However, enhancement of the practice of shared decision-making requires more than an isolated guideline which integrates SDM, especially when it comes to skills and attitude. It needs to be imbedded in an overall SDM contextualized effort for SDM to become common practice. Healthcare professionals could be trained, receive feedback on SDM performance, be facilitated in an SDM approach, work in organizations which advocate SDM and where there is senior-level buy-in of engaging patients in their care and sharing the decision-making with them [9, 32].

Limitations and strengths

Limitations of this study are that the findings are limited to the 15 participants and the university medical centers (3) they work at. Even though the participating centers were considered forerunners of SDM, it does not mean that other pediatricians in other medical centers have the same opinions. However, data saturation occurred after the eight interviews. Another limitation was the timing of the interviews; in hindsight, the interviews might have taken place too early as the guideline PCFC had not been fully implemented yet. Which meant that most of the participants had not used the guideline yet, and some had only heard of it. This was countered by showing the participants the guideline and by walking through the recommendations on SDM. Key strengths were that we conducted the research with the developers of the PCFC guideline themselves (EV, LK) and that participants were very forthcoming and open during the interviews as they were being interviewed by a doctor to be. We also carried out a member check on the interview transcripts.

Conclusions

The interviews showed that most of the participants thought that CPGs in general potentially could enhance SDM. However, integrating SDM into a guideline seems not to be an easy feat, and guideline developers have to walk a tightrope on how to formulate recommendations on SDM. They have to avoid stating the obvious because it might offend and alienate the pediatrician. At the same time, they have to provide more detail on how to practice SDM. Furthermore, developers could consider formulating more ‘open’ recommendations. Especially in case of preference sensitive choices, the recommendation should describe (treatment) trade-offs and (treatment) alternatives and provide more detailed guidance. Another consideration is to provide tools amalgamated with specific guideline recommendations that enhance SDM, such as PDAs.