An interprofessional team-based intervention to address barriers to initiating palliative care in pediatric oncology: a multiple-method evaluation of feasibility, acceptability, and impact

Abstract

Context:

Many children with advanced cancer are not referred to palliative care despite both professional recommendations to do so and bereaved parental preference for earlier support from sub-specialty palliative care.

Objectives:

To assess the feasibility, acceptability, and impact of an adaptive intervention to address individual and team-level barriers to specialty palliative care referrals.

Methods:

A multiple-method approach assessed feasibility and acceptability among clinicians from pediatric oncology teams at a single institution. Quantitative measures of comfort with palliative care consultations, team cohesion, and team collaboration were conducted before and after the intervention. Number of palliative care consults were examined before, during, and after sessions. Intervention satisfaction surveys and qualitative interviews were conducted after the intervention.

Results:

Twenty-six team members (90% of consented) attended at least one intervention session with 20 (69%) participants completing 75% or more sessions. The intervention was modified in response to participant feedback. After the intervention, participants reported greater team cohesion, comfort discussing palliative care consultation, team collaboration, process satisfaction, and decision satisfaction. Participants agreed that the training was useful, effective, helpful, and worthwhile, that they would use the skills, and that they would recommend the training to other providers. The numbers of palliative care consults increased before intervention sessions were conducted, but did not significantly change during or after the sessions. In the interviews, participants reported overall favorably regarding the intervention with some participants reporting changes in practice.

Conclusion:

An adaptive intervention to reduce barriers to initiating palliative care for pediatric oncology teams is feasible and acceptable.

The American Academy of Pediatrics advocates for all children with life-threatening conditions to have access to palliative care.(1) While interprofessional pediatric palliative care programs are becoming more common(2, 3), many children with advanced cancer are not referred to palliative care before death(4) or are referred only close to death, limiting the potential benefits of palliative services.(5) Many parents of children who die of cancer wish they had received extra support and palliative care during their child’s illness.(6, 7)

Palliative care referrals are delayed due to both individual clinician and team-level factors.(8, 9) Individual clinicians may be unfamiliar with palliative care, unsure of when a referral is appropriate, or uncomfortable with the uncertainty inherent in caring for children with serious illness.(10–13) Clinicians may conflate subspecialty palliative care with hospice services or worry about upsetting families by mentioning palliative care.(14, 15) Some clinicians may experience negative emotions when considering palliative care, or view referrals to palliative care as a professional failure.(16–21) At the team level, group norms may encourage clinicians to continue curative treatments as long as possible. Institutions may not have guidelines for initiating palliative care, clinicians may be unclear about which team members should begin palliative care discussions, and hierarchical barriers may prevent clinicians from suggesting palliative care to other team members.(22–25)

This study evaluated an interprofessional oncology team-based intervention, Confronting Adversity through Regoaling Engagement (CARE) Discussions, intended to facilitate oncology referrals to specialty palliative care. The intervention was designed to address individual and team-level inhibitions to referring families to a specialty palliative care team. The goals of the intervention were to: 1) increase understanding among pediatric oncology clinicians regarding palliative care services and explore variation in perceived ideal timing for initial consultation; 2) introduce techniques for handling prognostic uncertainty and negative emotions; 3) practice techniques for intra-team communication and team collaboration; and 4) practice communication with families of children with serious illness.

Intervention topics (see Table 1) and details of the intervention development (the Co-design phase) are reported elsewhere.(26) We used an approach based on adult learning theory, adapting the intervention within and across sessions to meet the interests and concerns of learners.(27–29) For adult learners, the theory posits the existence of an internal motivation to learn that is tied to their social role and is oriented to a perceived problem. An adult learning approach must identify what the learner perceives to be the problem to be solved and to build upon their previously acquired knowledge. To achieve an adult-based approach, revisiting what the participants perceived to be the impediments to sub-specialty consultation required adapting intervention components to address concerns as they were identified by participants. Previous studies with medical professionals found that flexible interventions may be more engaging, feasible, and effective in addressing concerns of participants.(30) Manualized interventions often require adaptation when implemented to tailor them to community and professional needs.(31–35) The current intervention also incorporated the idea of formative evaluation(36): the intervention would help teams figure out what skills needed improvement and serve as the beginning of a process of working together to decide when to introduce palliative care. A multiple method approach was chosen to evaluate this novel intervention because existing quantitative measures did not capture all the relevant aspects of acceptability and impact of the intervention.

Table 1:

Unrevised and Revised Intervention Curricula Outline

Session	Unrevised Topics	Revised Topics
1	Eliciting variations in understanding of primary palliative care and consultation appropriateness.	Eliciting variations in understanding of primary palliative care and consultation appropriateness. How to describe the palliative care team to families.
2	Medical uncertainty and challenges with navigating uncertainty. Activities about managing uncertainty.	Medical uncertainty and uncertainty about parent reactions. Activities about managing uncertainty for yourself and colleagues.
3	How to bring up topic of palliative care with different team members. Activities emphasizing building on existing team collaboration skills and diverse skill set of team members to handle challenge cases.	How to bring up topic of palliative care with team members and members of other teams. How does your team collaborate effectively? Activities emphasizing building on existing team collaboration skills and diverse skill set of team members to handle challenge cases.
4	Simulate group discussion of patient’s appropriateness for palliative care consultation when there is asymmetric information and different team perspectives.*	Review of challenges in introducing palliative care. Communication skills for discussing difficult topics with families. Role play activity for discussing the palliative care team with an upset parent.

^*Note this activity was dropped completely for the Neuro-Oncology and Bone Marrow Transplant teams. The Solid team went through a very brief version of this planned activity.

This report aims to evaluate the feasibility, acceptability, and impact of an intervention by: 1) describing the intervention implementation with three pediatric oncology teams in one institution and how we modified the intervention in response to the interests, needs, and skills of each group; 2) reporting quantitative data from pre and post surveys; 3) reporting quantitative data on the number of palliative care consults before, during, and after the sessions (including codesign and intervention sessions); and 4) reporting qualitative interview data on reactions to the intervention.

METHOD

This study was approved by the IRB at The Children’s Hospital of Philadelphia (CHOP)). All participants provided verbal informed consent except one who provided written consent.

Setting

Pediatric oncology teams and the specialty palliative care team (Pediatric Advanced Care Team (PACT)) at CHOP.

Participants

Interprofessional oncology staff (attending physicians, nurse practitioners, social workers, and fellows) from three pediatric oncology teams (Bone Marrow Transplant, Neuro-Oncology, and Solid Tumor) at CHOP were recruited for codesign sessions and intervention sessions from April 2016 to May 2017. Twenty-six staff members participated in the intervention sessions between September 2016 and May 2017.

Intervention Sessions

A customized curriculum was developed for each team in a co-design process.(26) Intervention development materials can be found at https://feudtnerlab.research.chop.edu/care.php. Each intervention included 4 one-hour sessions (Table 1), with materials and patient cases tailored for each team. The curriculum was modified over time independently for each team. Revisions to the curriculum resulted from post-session debriefings with the research team and real time alterations made by the facilitator (JW) to prioritize or skip elements of the curriculum in response to each team’s engagement with activities and feedback during each session.

Curriculum Revisions over Time

The suggested language for introducing the palliative care team to families was modified after the first session because team members in that session believed that the original wording was too reassuring that a palliative care referral did not mean the child was dying, which would be misleading in some cases. The session leader provided revised language in the following sessions that included ways that clinicians could acknowledge family concerns that their child would likely die from their cancer if appropriate.

In session 2, designed to address clinician uncertainty about prognosis, teams expressed more concern with their uncertainty about potential family reactions to discussions of palliative care and the timing of introducing palliative care than about prognosis. Clinicians worried that if they suggested palliative care to the family too soon, the relationship with the family would be negatively impacted. In later intervention sessions with other teams, the session leader explored these issues in more detail during session 2 and discussed strategies for managing negative family reactions. Introducing the palliative care team to families caused many clinicians stress because doing so was perceived as another “serious news” conversation requiring specific communication skills. Consequently, we developed an alternative fourth session focused on speaking with parents about involving palliative care, using the Vital Talk methodology(37, 38) for giving serious news and responding to resultant emotions.

Based on the discussions in sessions 2 and 3 regarding uncertainty about discussing palliative care with families (versus with colleagues), the session leader gave each team a choice for session 4 topics: collaborating with colleagues or communicating about initiating palliative care with families. All teams chose to explore skills for communicating with families. We therefore created a role play to offer skills practice in discussing the recommendation for referral to the palliative care team with a resistant parent. The challenges of family discussions that arose in session 1 were addressed with and strategies and skills in session 4 for all teams.

Quantitative Outcome Measures

Assessment surveys were administered pre- and post-intervention. Participants reported: 1) their professional role; 2) five items from the Group Environmental Scale(39–41), modified to measure team cohesion in a professional medical setting, including “Our team is united in trying to reach its goals for performance”; 3) how comfortable they were discussing palliative care with their team, specifically “I am comfortable discussing with my interdisciplinary team the advantages and disadvantages of involving the PACT team in the care of one of our patients”; and 4) the Collaboration and Satisfaction about Care Decisions scale(42), modified to reflect a decision to refer a patient to palliative care (6 items for team collaboration, including “Team members planned together to make the decision about care for this patient”; an item regarding satisfaction with the decision process, specifically “How satisfied are you with the way the decision was made for this patient, that is with the decision-making process, not necessarily with the decision itself?”; and an item regarding satisfaction with the actual decision, specifically “How satisfied were you with the decision made for this patient?”). Participants also completed 10 items (post-intervention only) to assess intervention satisfaction. As this study was not designed for hypothesis testing, we only examined point-estimates and confidence intervals at baseline and post-intervention.

Qualitative Evaluation

We conducted interviews between October and December 2017 with 16 staff members (8 physicians, 5 nurse practitioners, and 3 social workers). Interviews were conducted in person using a semi-structured guide by a trained interviewer (TES) who was supervised by a medical sociologist (JES). The interview guide (Supplemental Table 1) included open-ended questions to elicit the respondent’s overall perception of the intervention, feedback about the intervention sessions, suggestions for improvement, and whether the respondent perceived having made changes to their practice based on the intervention. A more detailed description of our methodology, and the results from other sections of the interview including perception of the PACT team have been reported previously.(21) The creation of the interview guide was informed by implementation of the intervention, a review of the literature, and consensus amongst our research team. We monitored for thematic saturation during data collection via memo writing and discussions between TES and JES. Given the focused nature of our research question, we achieved saturation quickly, after 8 interviews.

Interviews were audio-recorded, transcribed, and uploaded to NVivo 11, a qualitative data analysis software for management and coding. We utilized a hybrid inductive-deductive approach in which we identified themes that emerged from the interview data and applied codes to the transcripts we felt would be analytically useful a priori.(43) We read through a subset of interview transcripts to identify inductive codes and added these to a list of predetermined codes that were derived from the interview guide. Codes were clearly defined and discussed within the research team for clarification and modification before application to the entire data set. One data analyst (TES) coded the majority of the interview data with periodic reliability checks from another analyst (JES).

Palliative Care Consults Among Oncology Patients

To identify initial PACT consults, the inpatient and outpatient electronic medical records of all oncology patients between 2015 and 2019 were searched for first-time orders for palliative care consultation or the first appearance of a palliative care note. For interpretability, the calculated variable used in analysis was the number of initial PACT consults per 1,000 oncology patients per month, accounting for monthly variation in the total monthly number of unique oncology patients.

Because a significant number of oncology team members participated in detailed discussions of palliative care during the codesign sessions, we predicted that PACT consults would increase during the co-design sessions and continue to increase during the intervention sessions. In order to assess the adjusted monthly proportion of first-time PACT consults before, during, and after the sessions, an Interrupted Time Series Analysis was conducted.(44) We considered three separate time periods of first-time PACT consults: the period prior to the sessions (January 2015–March 2016); the sessions period, which included both the co-design and intervention (April 2016–May 2017); and (3) the post-sessions period (June 2017–December 2019). Newey-West standard errors were used with a maximum lag of 1 month. All estimates included 95% confidence intervals.

RESULTS

Feasibility of Intervention

27 primary team members (and 2 fellows) consented to participate, with 44% of eligible attending physicians (n=14), 75% of nurse practitioners (n=9), and 80% of social workers (n=4) consenting (53% of eligible clinicians overall, excluding fellows). 26 team members (90%) attended at least one intervention session, with 20 (69%) participants completing 3 or more of the sessions. 22 (76%) participants completed both the baseline survey and the post-intervention survey. All of these metrics met pre-determined targets for feasibility. Table 2 shows how many participants from each team completed the pre- and post-intervention surveys.

Table 2:

Session participants who completed both surveys

Sub-Specialty	Attending Physician n=10 (% of enrolled)	Nurse Practitioner n=5 (% of enrolled)	Social Worker n=8 (% of enrolled)	Fellow n=2 (% of enrolled)	Total n=21 (% of enrolled)
Bone Marrow Transplant	5 (83%)	2 (50%)	1 (100%)	0 (0%)	8 (73%)
Neuro-Oncology	5 (83%)	1 (50%)	2 (100%)	0 (0%)	8 (80%)
Solid Tumor	0 (0%)	2 (100%)	1 (50%)	2 (100%)	5 (63%)
Total	10 (71%)	5 (63%)	4 (80%)	2 (100%)	21 (72%)

Acceptability of Intervention

On a 7 point scale (1=strongly disagree, 7=strongly agree), participants reported that the training was useful (mean 5.3, SD 1.5), effective (mean 5.1, SD 1.6), helpful (mean 5.3, SD 1.6), and worthwhile (mean 5.4, SD 1.6), and that they would utilize the skills (mean 5.2, SD 1.5), that they would recommend the training to other providers (mean 5.4, SD 1.5), and that they would be willing to attend future communication workshops (mean 5.1, SD 1.7).

Change in Number of Palliative Care Consults

Across the 84 months of observation, 777 total PACT consults were requested, ranging from 1 to 17 per month, and an average of 9.3 (SD 3.7) per month. Figure 1 illustrates trends of first-time PACT consults per 1,000 oncology patients per month during “before, during, and after” time periods of the study. In January 2015 (before the sessions), there were 5 total PACT consults which occurred in 4.2 (95% CI: 2.7, 5.7) patients per 1,000 oncology patients per month, and during each subsequent month first-time PACT consults increased by 0.2, or 0.1 per 1,000 patients (both P<0.001). During the first month of co-design, no significant change in the number of first-time PACT consults occurred (P=0.5), and no statistically significant trend occurred during the sessions period (P=0.7), although the number of first-time consults continued an upward trend. During the first month post-sessions (June 2017), no significant change in the number of first-time consults occurred (P=0.7,), and no statistically significant trend was observed (P=0.4), although the number of consults trended downwards. Similar trends and non-significance were observed when examining the total number of monthly PACT consults across the study periods.

Figure 1 –

Interrupted time series analysis of the number of oncology patients receiving a PACT consult each month per 1,000 oncology patients before, during, and after the intervention.

Participant Perceptions of the Intervention

In interviews, participants expressed a largely positive response to the intervention. While a few participants offered criticisms or suggestions for changes, every participant had primarily positive feedback about their experience of the sessions.

“I think it was really, it was nice to have the opportunity to get our team together to really talk about this stuff at length. Because, you know, like I said, we’re a small, tight team that talks all the time, like, all day long. But it was really nice to have that 45 minutes or whatever it was to have a really long, in depth conversation about the theoretical nature of it. It was really nice, it’s just something we don’t get the opportunity to do much.”

(social worker)

“But I thought it was well run and I liked the small group environment, and it was non-threatening, even in simulation scenarios that can sometimes be uncomfortable, even though we’ve kind of gotten used to doing them over the course of training.”

(physician)

Many participants expressed some frustration about those who did not participate in the intervention. Since participation was voluntary, team members who were more resistant to palliative care involvement and perceived by colleagues as needing the intervention (to help them learn new skills or gain comfort with palliative care consultation) were less likely to choose to attend.

“I think it was helpful, but the problem is, the people who attended the interventions are the ones most likely and most comfortable with actually already consulting [PACT]. So, for the actual intervention, I think it would have been more helpful to physicians who don’t refer [to PACT] as often.”

(physician)

“The people that really needed it didn’t come. You know what I mean? And so – because I think all the people that participated in it are those – ‘preaching to the choir’ kind of thing. And so I just wish some other team members would’ve participated.”

(social worker)

Impact of Intervention on Practice

After the intervention, participants reported increased team cohesion, comfort discussing palliative care consultation, team collaboration, satisfaction with the decision process, and satisfaction with the actual decision (Figure 2).

Figure 2 – Pre- and post-intervention response scores regarding five attributes of impact of intervention on oncology team practice.

The “rating scores” were based on a 10-point scale of agreement, ranging from 1 (Strongly disagree) to 10 (Strongly agree). Items are sorted from largest change at the top to smallest change at the bottom. “Comfort” was 1 item, “I am comfortable discussing with my interdisciplinary team the advantages and disadvantages of involving the PACT team in the care of one of our patients.” “Team Cohesion” was the mean of 5 items including “Our team is united in trying to reach its goals for performance.” “Collaboration” was the mean of 6 items including “Team members planned together to make the decision about care for this patient.” “Process satisfaction” was 1 item, “How satisfied are you with the way the decision was made for this patient, that is with the decision-making process, not necessarily with the decision itself?” “Decision satisfaction” was 1 item, “How satisfied were you with the decision made for this patient?”

In response to open ended questions, participants reported that the most useful aspects of the intervention were: “having the opportunity to talk with my team in a protected setting about these multidisciplinary issues”; “getting written tips or phrases to state”; “role playing”; and “hearing how my whole team perceives we work together without getting too annoyed by each other.” Participants reported that the least useful aspects of the intervention were: the amount of time required; that some of the information was too basic for their team; that their team did not have a problem with referring patients to palliative care; that some team members were not positively engaged in the process; and that the intervention did not address the specific skills of their team members.

In the interviews, participants reported several themes including developing “new skills”, some of which impacted their or their team’s clinical practice. Skills included increased confidence, active listening, mindfulness about the timing of conversations, and new ways of communicating with families (See Table 3 for examples). Several physician participants felt that they had learned “new language” or phrasing for communicating with families about involving palliative care (Table 3). Nurse practitioners and social workers highlighted the quality and value of the “role-playing exercises” especially watching the physicians navigate a parent who was very resistant to sub-specialty palliative care. Some participants reported modest “changes to their own and their group’s practices” surrounding specialty palliative care consultations including social workers feeling more comfortable advocating for palliative care, especially when communicating with other team members who had attended the intervention.

Table 3:

Quotes from Qualitative Interviews

Themes	Quotes
New Skills	“I think understanding how to provide hope without unrealistic expectation is good, one of the people who was in my group – one of the attending physicians – is a long-standing [physician] here who is very straight forward with patients. So it was interesting to hear him discuss how he maintains rapport with his patients after he is sort of just flat out with them, where they’re at. That was helpful.”
	“So I think that’s something they also emphasize, is, being a good listener. Instead of kind of jumping in and making assumptions about what families want, just ask them. So I think that was a useful thing that they talked about in the session.”
New Language	“[L]ike you heard me say, ‘layering support.’ I think that’s something that they talked about using as a kind of word that we can be consistent in saying we’re going to layer in extra supports, so families get a visual.”
Benefit of Role Playing Exercises	“The thing I think was the most helpful, at least when we did one piece in here was with the doctors kind of acting out those conversations. Because I’m sure they did that in med school and grad school. It’s like the thing you hate the most, right? But I think it’s the most helpful. […] And I remember when we were acting it out, and I forget which doctors it was, but one kept—it was one scenario when the family just kept saying, ‘Nope. I don’t want it. Nope,’ and just kept pushing back, and how to get around that. [Laughter] So, it was really, I think, very helpful, because we’ve all been in those situations, and sometimes you need a script and you need those words to be able to really navigate it. So, I felt that that was really helpful, and I think it definitely improved—I mean, we were already, I think, in a good spot in our team, but helped us improve how we communicate with each other about getting PACT involved.
Changes to Own Practice	“I think it helped me as a social worker feel even more confident about advocating for Palliative Care, because when the physician is reminded of what Palliative Care can do and will, I think it helped me to say, ‘Remember?’ Especially in that, I feel like that time frame, like in the month or so after the intervention, I felt like we were talking a lot more, like, ‘Oh, we have PACT for this person and this person.’”
	“I think sometimes it helps the little voice inside me to speak up a little bit more, thinking back to the concepts that as a group, we came together and decided were appropriate times to consult palliative care. And so it gives a little bit of a signpost to go back to and think about, ‘Okay. We said as a group that any patient getting a second transplantation should think about having a palliative care consult.’ And it’s a common agreement that we came up with. And so it’s a little bit – makes it a little bit easier to bring it up and say, ‘Remember, we discussed this. Are we still thinking that for this particular patient situation?’”
Challenges	“It might have helped us individually in how we reflect about palliative care and it might actually – it may have influenced the ease with which our newest attendings go towards palliative care. I’m not sure if it made one iota of difference in our more senior attendings and how they approach palliative care and those conversations.”
How to Improve Intervention	“I still feel like, “What was the Intervention?” Like was it just having a meeting and talking? I’m not sure what we did. I don’t know. Just being totally honest.”
	“It was so much time. And I think that, you know, we’re all stretched so thinly that it could’ve been condensed into one session. Like, I would’ve had the intervention developers cut out as much as possible to make it even one slightly longer session, but it was a lot of time.”
	“I think it would’ve been nice to have had that earlier when I was sort of younger and having started out, because I think it was really helpful.”
	“The only thing I would improve … was … the participation, we had at least two providers who probably needed it more who weren’t there. So, I think that’s, just like anywhere, the people who—it’s like preaching to the choir, sometimes.”

Some “challenges” to changing group practices were described. Members of one group that had discussed creating automatic “triggers” for consultation reported that while the team was making and discussing referrals more frequently, the practice had not been standardized, though no participants identified the reason why (beyond a general perception the intervention did not impact the attendings who had the most authority in making these decisions about consultation). Some participants described that team members who had opted not to attend the sessions were unlikely to change their practices.

Some participants suggested “improvements in the intervention” to clarify the intent, structure, and expected impact of the intervention sessions. Other suggestions included streamlining the content, providing content online, and presenting the intervention to providers early in their careers.

DISCUSSION

Our project demonstrated four main findings. First, an intervention to discuss barriers to palliative care referrals with pediatric oncology teams that is co-designed with team members and based in adult learning theory is both feasible and acceptable to interprofessional groups of clinicians. Second, we demonstrated the ability to maintain an adaptive approach to curriculum design and implementation to improve the intervention impact on clinicians with varying needs and group dynamics. Third, the intervention did not change the number of sub-specialty palliative care consults that were requested. Finally, we found that teams were most engaged with activities related to individual skill building and team processes and less engaged in activities requiring insight about personal behaviors or conflict with their team members present in the session.

Feasibility and Acceptability of Intervention

Feasibility of the intervention was confirmed with 53% of eligible clinicians enrolling and 69% of enrolled clinicians retained for at least 75% of the sessions, which is consistent with other published pilot studies of physician interventions.(45, 46) The response rate of 76% for the completion of two online measures was also adequate to confirm acceptability.(47–49) Additionally, the quantitative measures of comfort discussing palliative care, group integration, team collaboration, and satisfaction with palliative care decisions all increased after the intervention. Instead of requiring clinician participation, to achieve even higher levels of meaningful participation leaders may need to find ways to enhance clinicians’ motivations for participation to ensure the intervention has the greatest potential impact on care.

Ability and Need for Flexibility in Curriculum and Implementation

Despite feedback from a subset of members of each subfield while co-designing the initial intervention materials, we still needed to adapt the materials to maintain engagement with participants. Post-intervention interviews showed that some of the added materials (e.g. how to talk to families and role play) had the most impact. Some participants reported using language and communication tools from the intervention and being more comfortable bringing up palliative care after the intervention.

One reason for the needed adaptations to improve impact may have been that physician participants were under-represented in our co-design group. Prior co-design studies have noted the challenges of designing materials and interventions based on groups with fewer physicians than other professions(50), and the fact that health care professionals in different roles may have different perspectives on the ideal ways to provide patient care.(51, 52) Our materials may have included concepts and activities (e.g. cognitive insight, coping with uncertainty) that social workers were more comfortable discussing but that physicians found less relevant. Particular attention should be paid to matching the professional disciplines of those who participate in the co-design process with those in the intervention.

Number of First-Time Palliative Care Consults Among Oncology Patients

Although we did not find a significant increase in new oncology palliative care consults after the study sessions began, the number of new consults were increasing in the time period before sessions started. This trend leveled off and declined slightly after the sessions ended. These findings may indicate that awareness of palliative care among the oncology teams was already increasing before the sessions were conducted, or that attempts to inform and recruit oncology team members for a study about palliative care increased awareness of the palliative care team.

Relative Success of Individual Skill Building and Team Processes

Many team members preferred skill building in communication or discussing triggers for consulting palliative care over discussions requiring reflection on individual cognitive errors (such as when to consider consultation) or team conflict. These latter activities may have been less successful in our intervention because team members from different levels in the medical hierarchy were included in each session and we did not have adequate time to build psychological safety to discuss topics requiring more participant vulnerability.(22, 23) As a result, participants may have been reluctant to raise certain concerns or disagree with higher status participants from their team.(53, 54)

Study Limitations

Our study has several limitations. The participants came from three pediatric oncology specialty teams at one pediatric hospital, thus the results may not be generalizable to other oncology teams and institutions. In addition, social workers and nurse practitioners participated to a greater degree than did physicians, so physician preferences may not be accurately represented. Finally, due to scheduling difficulties, bedside nurses were not invited to participate in the intervention sessions, thereby reducing the comprehensiveness of the intervention to impact all clinicians who may identify families’ palliative care needs.

CONCLUSION

The current project demonstrated the feasibility and acceptability of conducting an adaptive intervention to reduce barriers to initiating palliative care for pediatric oncology teams. Intervention elements focusing on individual skill building and team processes may have the most acceptability and likelihood of influencing practice. Further research should investigate the disconnect between agreement with the need globally for palliative care but individual reluctance to involve sub-specialty palliative care for one’s own patients.

Key Message.

Many children with advanced cancer are not referred to palliative care. We assessed the feasibility, acceptability, and impact of an intervention to address individual and team-level barriers to specialty palliative care referrals. We found that the intervention to reduce barriers to specialty palliative care referrals was both feasible and acceptable but did not increase sub-specialty palliative care referrals post-intervention.