Clinicians’ views on the use of triggers for specialist palliative care in the ICU: A qualitative secondary analysis

Abstract

Purpose:

To understand clinicians’ views regarding use of clinical criteria, or triggers, for specialist palliative care consultation in the ICU.

Materials and methods:

Secondary analysis of a qualitative study that explored factors associated with adoption of specialist palliative care in the ICU. Semi-structured interviews with 36 ICU and palliative care clinicians included questions related to triggers for specialist palliative care. We performed a thematic analysis to identify participants’ views on use of triggers, including appropriateness of cases for specialists and issues surrounding trigger implementation.

Results:

We identified five major themes: 1) Appropriate triggers for specialist palliative care, 2) Issues leading to clinician ambivalence for triggers, 3) Prospective buy-in of stakeholders, 4) Workflow considerations in deploying a trigger system, and 5) Role of ICU clinicians in approving specialist palliative care consults. Appropriate triggers included end-of-life care, chronic critical illness, frequent ICU admissions, and patient/family support. Most clinicians had concerns about “trigger overload” and ICU clinicians wanted to be broadly involved in implementation efforts.

Conclusions:

ICU and palliative care clinicians identified important issues to consider when implementing triggers for specialist palliative care consultation. Future research is needed to longitudinally examine the most appropriate triggers and best practices for trigger implementation.

1. Introduction

The benefits of specialist palliative care have been widely reported, including improved symptom burden and quality of life, improved serious illness communication and decision-making at the end of life, and increased caregiver support [1]. In the intensive care unit (ICU), specialist palliative care has been shown to better patient and family experiences, increase provision of goal-concordant care, and impact treatment intensity during terminal hospitalizations [2–4]. However, integration of specialist palliative care in the ICU may be limited by heterogeneity of services, barriers to consultation, adoption of services, and lack of understanding of indications for palliative care consultation [2], resulting in its underutilization [3].

One way to improve integration of specialist palliative care in the ICU is to use an active case-finding approach with screening criteria, or triggers, to prompt consultation when appropriate [5–7]. Existing triggers for specialist palliative care in the ICU have been based on healthcare utilization (e.g. length of ICU stay), disease state (e.g. sudden and severe brain injury), and clinician estimates of prognosis [2,3]. Despite their enumeration in the literature for over a decade, triggers are variably used in practice and effective implementation strategies remain unclear [2,3,6]. Moreover, few data exist to understand clinician views about the use of triggers for specialist palliative care consultations [2,5].

This study is a secondary analysis of semi-structured interviews with ICU and palliative care clinicians that aimed to understand the process by which specialist palliative care is adopted, or integrated, in the ICU setting. While the parent study developed a theoretical model of specialist palliative care adoption [6], the current study analyzed clinicians’ responses to specific questions related to the use of triggers, with the objective of providing an in-depth exploration of ICU and palliative care clinicians’ views on the use of triggers for specialist palliative care consultation in the ICU.

2. Materials and methods

The parent qualitative study was conducted at two large medical centers in New York City between 2016 and 2020 [6]. This secondary qualitative data analysis utilized the rich data derived from semi-structured interviews with ICU and palliative care clinicians from multiple ICUs with reported variable specialist palliative care usage including medical, surgical, transplant, neurological, and cardiothoracic ICUs. Institutional review board approval was obtained at both institutions and written informed consent was obtained before subject enrollment in the study.

Participants were identified through purposive sampling of ICU clinicians, ensuring an adequate mix of clinicians (e.g., attending physicians, nurses, social workers) from each ICU type, as well as palliative care clinicians with regular interaction with critically ill patients (e.g., physicians, nurse practitioners, social workers, and chaplains) [6]. This sampling approach ensured that a multitude of clinician perspectives and experiences were captured. We aimed to enroll 2–4 clinicians identified from each ICU type in both centers with a planned total enrollment of 40 clinicians or until data saturation was achieved.

2.1. Data collection

Demographic information including age, gender, race/ethnicity, clinician type, specialty, number of years in practice, and number of years working in the current hospital was collected from participants. Semi-structured interviews were conducted using face-to-face interactions and audio recording for transcripts. The interview guide was created to develop and inform a theoretical framework of the process of specialist palliative care adoption in the ICU setting and ensure that the objectives of the parent-study were met [6]. Specific to the aims of this secondary qualitative analysis, participants were asked about their views when they thought palliative care involvement was appropriate, whether their ICU had specific triggers for use, and how they might envision triggers (as well as other methods of integrating palliative care specialists) to be implemented successfully. Questions from the interview guide relevant to the study are included in Table 1.

Table 1.

Targeted questions from semi-structured interview guide.

ICU clinician questions

• Is there a process in place in your ICU for determining whether a patient needs palliative care consultation?

• How do you determine whether a patient needs palliative care consultation?

• Do you think there should be a formalized process for identifying patients who should receive palliative care consultation? (Why yes or no? What does that process look like?)

• Would you have been more likely to utilize the palliative care consultation service had it been rolled out in a certain way? (Formal implementation leaders, champions, triggers?)

• Who are the people who have the greatest role in affective how much palliative care consultation is used? (Stakeholders?)

Palliative care clinician questions

• Are there formal processes in place for determining whether an ICU patient needs palliative care consultation? (What does that process look like?)

• Do you think palliative care consultation should be mandatory for certain patients? (Types ofpatients, clinical conditions, triggers?)

• Do some ICUs consult more frequently than others? (Why? Are you consulted for different reasons?)

• What preparations do you think are necessary prior to introducing a palliative care consultation service in an ICU setting? (Formal implementation leaders, provider education, champions, triggers?)

2.2. Analysis

Qualitative interviews were transcribed verbatim. All transcripts were originally coded by 2 members of the research team (MH, LF) using an inductive and iterative process and constant comparison. In a separate round of coding, a thematic content analysis was conducted by 3 investigators (MH, LF, KPM) to identify the major themes that arose surrounding clinicians’ views on use of triggers for palliative care consultation in the ICU. Codes that were potentially related to discussion of trigger use were identified for thematic analysis, including codes related to triggers, the electronic health record (EHR), and why clinicians would involve palliative care (e.g. goals of care, family support, chronic critical illness). ATLAS.ti was used to manage the data. After identifying representative quotes from a heterogeneous sample of participants of different clinical backgrounds, quotes were grouped under each major theme heading for synthesis of findings. The decision-making process behind the selection of quotes and identification of themes were discussed in an iterative manner among all members of the research team until agreement was achieved.

3. Results

There were 36 participants enrolled in the study (16 from Center 1 and 20 from Center 2). The overall sample included 17 attending physicians, 11 critical care nurses, 1 social worker, and 7 palliative care clinicians. Table 2 displays baseline characteristics of study participants. The sample consisted of mostly White participants of a median age of 35 years and was 58.3% female. The average number of years in practice was 5.5. The average length of interview was 44 ± 14 min. Five main themes emerged: 1) Appropriate triggers for specialist palliative care, 2) Issues leading to clinician ambivalence for triggers, 3) Prospective buy-in of stakeholders, 4) Workflow considerations in deploying a trigger system, and 5) Role of ICU clinicians in approving specialist palliative care consults. Representative quotes for each theme are presented in Table 3.

Table 2.

Participant characteristics.

Age, median (IQR)	35 (33–42.5)
Sex, n (%)
Female	21 (58.3)
Male	15 (41.7)
Race/Ethnicity, n (%)
White	22 (61.1)
Asian	13 (36.1)
Other	1 (2.8)
Hispanic/Latino	0(0)
Clinician type, n (%)
ICU attending physician	17 (47.2)
ICU nurse	11 (30.6)
ICU social worker	1 (2.8)
Palliative care clinician	7 (19.4)
Years in practice, median (IQR)	5.5 (4–11.5)
Type of ICU, n (%)a
Surgical	17 (47.2)
Neurological	6 (20.7)
Medical	6 (20.7)
Center, n (%)
Center 1	16 (44.4)
Center 2	20 (55.6)

^aIncludes the 29 ICU clinicians. “Surgical” comprises surgical, cardiothoracic and transplant ICUs.

Table 3.

Major themes and representative quotes.

Theme 1: Participants’ views on appropriate triggers for specialist palliative care

• I usually think about palliative care when I feel like the medical interventions have reached their limit and...we need to talk about...goals of care especially when someone is reaching the end of their life, has drastically deteriorated, and we have nothing else to offer medically. (P12 - ICU Physician)

• I know I just said the definition doesn’t necessarily need a terminal diagnosis but it’s usually I would think of palliative care when...a patient has advanced directives that they may not want, that they don’t want any heroic measures, then I would say...palliative care could be involved. (P8 - ICU Nurse)

• There are certain patients that automatically, I know they will need it [palliative care]...people who have been in the ICU before, so our repeat visitors because I’m concerned about chronic critical illness, and I want to make sure they have a sense of the burden of critical illness associated with like just these repeated stays. (P26 - ICU Physician)

• They’re still calling us at the point where critical care is either not saving the patient’s life, and death is inevitable, and that is the point at which I think the intensivist clearly sees an indication to call palliative care, to either have the conversation with the family that the patient’s dying and what are the goals now that the patient is dying, and introduce hospice for example, maybe that is a way of bridging to be able to discharge the patient from the ICU to a hospice scatter bed. (P24 - Palliative Care Physician)

• One of the things that I would love to see happen is for us to develop a unified definition of what chronic critical illness is. Umm it’s a syndrome that has been acknowledged by all the major societies and associations for critical care clinicians, but I’ve noticed that over my years here every unit defines it differently. Umm and what I would love is that for that to be a mandatory consult if anyone reached that definition of chronic critical illness but in order to do that we have to all agree on that definition. (P3 - Palliative Care Social Worker)

Theme 2: Issues leading to clinician ambivalence for triggers

• …I think if we had formalized criteria, of we’re discussing goals of care or something then we would be calling them all the time…like our trach and peg rates are probably quite high because we have so many brain-injured patients so I think like if we didn’t do a lot of the primary work ourselves and if we just said okay this patient has poor prognosis call palliative care, I think we’d be calling them on like the whole ICU and I think that would just be overwhelming and not that helpful. (P28 - ICU Physician)

• I think that if you have a formalized process, patients and families and diseases aren’t cookie cutter. I don’t think it will be feasible – it would be disadvantageous as we would be calling them for patients that we may not actually, necessarily need it. And you may not call them for patients who actually may need it. (P11 - ICU Physician)

• I think it would greatly benefit us…palliative care consult is something that could be put in and then retracted if palliative care doesn’t think it’s appropriate to follow this patient, they can always retract their consultation. So it would be tough to uhh to put something in place that would trigger a palliative care consult…We have patients that wax and wane…I would have to sit down and actually think long and hard about how to implement something that. (P32 - ICU Nurse)

• Don’t know what the ideal way is. Cause I’ve actually seen it in other hospitals … I did my rotations at another hospital…they had something similar to this were there were…parts of the patient’s chart that…cause, like this alert to come up during admission and ask whether the provider wanted to consult palliative care and I think ideally in my head that sounds great but when I saw in practice, it started, there were just a lot of consults for palliative care that were not necessary, would be great probably, but they started to turn down a lot of patients because there were just too many and started to get, started to become frustrated for the palliative care providers because they weren’t necessarily being called for a reason or like being called to help in any way, it was just because this thing popped up and they clicked ‘yes’ and they don’t know why they’re consulting palliative care. It was just a prompt from the computer. So, I, I’m personally trying to figure out a better way to have palliative care be more involved earlier in a patient’s stay without the negative effects of over doing it I guess. (P19 - ICU Nurse)

• I am a big proponent that you call consult when you need them. When there is a particular question or there is a particular struggle, umm I am not a big believer in automatic consults for anything, including pall care. (P14 - ICU Physician)

Theme 3: Prospective buy-in of key stakeholders

• An organization can look at the literature and say these are reasonable triggers to get for palliative care consult. I think by doing that…it will be very hard to sustain unless you have very clear victories with that process. So I believe that the triggers should be developed by the local organizational culture. So that everyone is on the same page with what those triggers should be. And in the surgical ICU you have a lot of conflict between surgical attendings, critical care staff and palliative care. All three groups and-and include nursing in there - four groups need to have the same mental model for what the triggers should be. If everyone agrees on that this program will fly because we can say we’ve developed out triggers, we’re all on the same page this is our organizational culture. If you don’t have buy in, then you are going to force people down and they’re not going to want to come to work. So you need people to want this to happen and that’s something I think we make a mistake with whenever we want to change how we function in a hospital. (P1 - ICU Physician)

• I think if you have an ICU that has a highly integrative model of palliative care where several attendings are champions of palliative care probably the efficacy of implementing a formalized process would probably be small in terms of efficacy in delivering substantially better palliative care to patients. (P13 - ICU Physician)

• The trigger system needs to have buy in from those in the front lines, not the hospital hierarchy, not the, not the – the chairman, I mean it’s nice that they might bless it and provide the funding for it, but you can’t have a mandate that says this is what you do. …The nurses, the people on the front lines. And frankly my own belief is that the more, the lower on the totem pole that person is on the hierarchy the better it is for them to develop the triggers. I think if you can get that group to buy in, and then refine it from input with my senior input overtime people will feel excited about making it happen. (P1 - ICU Physician)

• I think hospital administration finally understood the importance of palliative care and they became more supportive, so we have expanded now to I think four attendings, two social workers, three or four – four nurse practitioners…the chaplain and other teams. (P9 – ICU Physician)

Theme 4: Workflow considerations in deploying a trigger system

• Now what we’re essentially doing, not very sophisticated but it is very effective, we have a weekly conversation on Mondays between the medical ICU attending and the palliative care and they just look through their list with us there and tell us about a few of the patients which strike them as maybe benefitting from a palliative care consult and we give our input and we say oh in this case we may be able to be helpful with pain management, this sounds like complicated family dynamic we could probably help with providing some emotional support, and giving some resources for anticipatory grief that the family is going through, or maybe it would make sense for us to plan a family meeting for some time early next week, whatever the situation is we sort of give our input. (P24 - Palliative Care Physician)

• In other hospitals I’ve seen, well actually I don’t know if they have that here, but in other hospitals I’ve seen umm where there is a – actually I do think they have it here – a like a best practice alert that comes up, recommends whether or not you want to consult palliative care on certain, certain I don’t know, whether its age, disease something that in the patient’s chart that prompts the provider to say ‘well this person actually is a good candidate for palliative care’ like something besides a personal recommendation from a random nurse but I think something more structured could definitely help because that forces them to understand that there is more to palliative care than goals of care discussions. (P19 - Palliative Care Social Worker)

• Yeah so we have that as a – that’s a policy or guideline that exists. So when patients in particularly poor prognosis just based upon EMR criteria. I know [clinician] put this together like a dialog box that comes up. I don’t know if it ever comes up on a surgical patient huh – it comes up plenty on medical patients, it came up so much I told, I told [clinician] that I’m getting like you know I’m getting dialog box palliative care prompt fatigue. (P13 - ICU Physician)

• Well – so I’m thinking similar to like the …how RRT gets called for like sepsis triggers…so like maybe it should just get called for everybody and then they could determine whether or not it’s appropriate. That might be better, because then people wouldn’t hesitate to call and that would get rid of the attending thinking maybe we don’t need them…but I don’t know how that goes because I know like with RRT its nurses calling and nurses normally – we don’t have as much autonomy as doctors, so I don’t know if doctors would really like being told ‘oh like based on this score you have to have palliative care involved.’ (P20 - ICU Nurse)

• I mean currently there is a note that pops up for patients on the EMR, and you have to acknowledge that the patient has 6 months or less to live, and that they should be considered for a palliative care consult. And I think that you know we sometimes use that EMR trigger as a help but we don’t – the triggers aren’t necessarily binding, they don’t force people to call palliative care consults, they are just reminders…if you think pall care will be useful in these patients. So the triggers are just reminders but like I said I think we would encourage people to call them, beyond the triggers for more resources. (P14 - ICU Physician)

• I think it should be definitely in the computer, everybody documents electronically rather than paper and so for that caregiver is that going to be the nurse or resident or maybe it would be - ideally it would be whoever, whichever team member for the patient at some point of the electronic documentation when it comes to some triggers, in the electronic documentation. (P5 - ICU Nurse)

Theme 5: ICU clinicians’ involvement in approving specialist palliative care consults

• So now we use triggers, and we use other suggestions to help us remind who should get pall care consults, but we are mindful of the fact that we all don’t necessarily want to give up our practice of pall care and then to burden some other group to just take care of all the pall care needs, we want that to be shared. (P14 - ICU Physician)

• If there were a formalized process, I think having – if it were automated in terms at least in saying. If we knew what the EMR system was or – and it was reasonable with a transparent set of guidelines or a check list that you can say oh met all of these criteria and it popped up automatically and then, I would still leave it to the clinician I think to consult palliative care directly. Because I do think there are situations where families are absolutely and willing to address – or there are situations in which palliative care will not be receivable by the family and I don’t know if there is a way for EMR to pick that up. (P2 - ICU Physician)

• I mean if the family would have been receptive to palliative care and also if we had had some…protocols in case. Like when they are not a surgical candidate, maybe that is a referral, or maybe when they want to come to the ICU, maybe the ICU attending or the fellow needs to say “this is a candidate for palliative care based on these criteria, check, check, check, check.” (P8 - ICU Nurse)

• Maybe if providers can be more involved, involving nurses more, that would be useful for a lot of things but also for pall care if providers would just have more involved communication with the nurses and asking for their opinions something like that. (P35 – ICU Nurse)

3.1. Theme 1: Appropriate triggers for specialist palliative care

There were clinical situations in the ICU where many participants broadly endorsed the use of palliative care specialists. First, participants felt that involvement of specialists was appropriate when a patient was approaching care that would be inconsistent with meaningful recovery. These scenarios were often described for patients who were at the end of life or for patients whose expressed goals may not be achievable, necessitating a goals of care discussion. For example, an ICU physician stated: “If the person appears to be quite critical in terms of having multiple system organ failures and needing dialysis and being on a vent and perhaps having other lab abnormalities that suggest having other organ dysfunctions, when someone reaches that state, it’s helpful to have palliative care assist us in making sure we all are achieving the same goals.” Often in the ICU, goals of care discussions and approaching the end of life seemed to be interchangeable, and across different clinical scenarios, many clinicians expressed that palliative care specialists were helpful to guide conversations.

Chronic critical illness was commonly endorsed by participants as an appropriate trigger for consultation, which was often operationalized as “patients who have tracheostomies” or “patients who are there more than a week… progressing from intermediate duration ICU stay to long term.” Both ICU and palliative care clinicians viewed these patients as being at high risk for therapeutic failure and expressed that involvement of palliative care specialists would be helpful for introducing this possibility and setting expectations. In particular, specialists were viewed as being helpful for communicating prognosis and the medical team’s view of the clinical picture, and to help patients and families understand the likely need for continued and prolonged aggressive medical care. Also, because chronic critically ill patients have long hospitalizations and ICU staff often change on a weekly or biweekly basis, many clinicians expressed that palliative care specialists were valuable to provide much needed continuity and consistent messaging to families because they would be able to develop a longer-term relationship with patients and family members. As one ICU physician stated, “When I’ve seen them the most valuable is for relationships with family lasting more than a week. So the ICU, our attendings change every week but the palliative care team does not, so if they’re going to be having these relationships over time, I think that’s a really nice way for the patients and families to not get a new message about prognosis or goals of care every Monday with a new attending.”

Frequent ICU admissions were also cited as reason for specialist palliative care consultation. An ICU nurse stated: “I think right when the patient hits the ICU...that was not expected or if they bounce back to the ICU…I think palliative care would be perfect…to…introduce themselves and say ‘hi we’re palliative care…I know you’re back in the ICU or you might be here for a long time…I think we have to set some goals and see how you feel about your illness.’” These patients were viewed as having an uncertain prognosis and care trajectory, and the potential implications of requiring frequent admissions were viewed as burdensome and indicative of need to introduce specialist palliative care. In such clinical situations, both ICU and palliative care clinicians described the ideal scenario as one where palliative care specialists would be introduced and provide psychosocial support for patients, families, and staff, develop trust, and then if needed, specialists would be wellpoised to guide and support the family through difficult circumstances and complex medical decision-making.

3.2. Theme 2: Issues leading to clinician ambivalence for triggers

While certain clinical scenarios were commonly endorsed as being appropriate for specialist palliative care involvement, participants also raised concerns that led them to feel that trigger systems may not be helpful. Some expressed difficulty in discerning which ICU cases that would benefit from specialist palliative care consultation. Also, because clinicians commonly viewed specialist involvement as being associated with the occurrence or potential for therapeutic failure, they often expressed uncertainty regarding the utility of specialists for patients with ambiguity in their clinical status or trajectory. An ICU physician expressed, “I mean I feel like everybody comes to me with a poor prognosis because they’re so critically ill so that’s really hard for me to gauge, like I don’t think that certain prognoses…or certain conditions or disease processes lead me more toward getting palliative care involved or needing palliative care more than others.”

Some participants were skeptical about the benefit of trigger systems. Participants described concerns stemming from the inability of triggers to appropriately capture true patient need, and how trigger use may result in inefficient use of a scarce resource. For instance, one ICU physician stated: “I don’t like the shot gun approach that everyone who’s been there for five days needs palliative care, everyone who has this surgery needs palliative care, anyone who’s on a ventilator five days needs palliative care. And I don’t think that’s true, and I think it results in a lot of over consults and spreads the palliative care team too thin on patients that will not benefit.” Participants also relayed concerns that an undesirable and unintended consequence of frequent use of specialists was that generalist palliative care practice for ICU clinicians would decrease. An ICU physician stated, “The downside was obviously…we don’t sortof wannalikerelytoo muchon the presence of palliative care because you know, the last thing I would wanna see is…for the MICU or whatever primary ICU team to say ‘ok just call pall care’ and then let’s move on because I do really think that there should be that responsibility from the primary team to at least try to address some of those issues.”

3.3. Theme 3: Prospective buy-in and input of key stakeholders

Organizational and clinician buy-in was viewed as a critical component of implementation efforts. Specifically, having adequate stakeholder (e.g., institution, unit-level, ICU and palliative care clinicians, administrators, etc.) engagement and buy-in for palliative care consultation was often mentioned as a pre-requisite for triggers to be successfully implemented. One ICU physician commented, “The most effective ways of doing it are to have a system that makes something happen, so like a trigger system where everyone buys in and have your leadership talk about why this is important. And to do that visibly. When that happens, change starts to happen.”

Participants described how implementation of triggers without buy-in or prior input from users would not be successful. For example, one participant stated: “I know that…there is a flag…in EMR, that says ‘this patient will benefit from palliative care services,’ what that’s based on I have no idea. Most of those patients are already getting palliative care or they’re not appropriate for palliative care. I guess in, at least in the way we are using currently, because they are say listed for transplantation and they have just been in the hospital for a long time.” Without input from end-users, participants felt that a trigger system would not be able to appropriately identify patients who would benefit from palliative care, resulting in low clinical utility, and that clinicians should be actively involved in troubleshooting or refining triggers. Additionally, participants felt that successful implementation would likely be influenced by existing practices for clinical decision-making within a respective unit, underscoring the importance of unit-level buy-in.

3.4. Theme 4: Workflow considerations in deploying a trigger system

Participants universally stressed that triggers needed to be implemented correctly to be effective and described their experiences with successful and unsuccessful implementation. On the topic of how trigger systems might be implemented for individual patients in practice, participants held a wide range of views. Most participants agreed that palliative care specialists were underutilized (either overall or for specific types of patients) and expressed that their existing method of deciding which patients received a consultation could be improved. Some felt that patient needs were so prevalent that they favored a system that prioritized access to palliative care with fewer barriers, acknowledging that this may lead to unnecessary consultation. For example, an ICU nurse stated, “I think it would greatly benefit us…palliative care consult is something that could be put in and then retracted if palliative care doesn’t think it’s appropriate to follow this patient.” However, participants were also worried that triggers may identify too many patients for palliative care specialists to see resulting in overutilization and fatigue.

Some participants felt that triggers should be incorporated into patient care through the electronic health record, while others felt they should be used during rounds. Some participants expressed support for electronic health record (EHR) imbedded triggers as they may be an effective way of increasing consultation and in turn increasing access to palliative care for patients who might benefit. For example, a palliative care nurse practitioner expressed, “The other is establishing some kind of trigger set…I know there have been some tools developed already for ICU patients…also for emergency room patients, like if they meet X criteria like that results in automatic palliative care consult and that could just mean that they are on our radar, maybe on paper it meets the criteria, but we’re not going to be of so much utility to that patient at that time. And other times it probably would catch people a lot earlier.” Others disagreed with this approach given the high likelihood of potentially ineffective and overly sensitive triggers that would lead to unnecessary consultation and trigger fatigue. As an alternative, participants suggested that high need patients who would benefit from specialist palliative care could be identified on clinical rounds, discussed as a team, and approved by the ICU physician.

3.5. Theme 5: Role of ICU clinicians in approving specialist palliative care consults

In actual and hypothetical practice, most participants supported the idea that ICU clinicians should be involved in “approving” the consultation. Automated triggers based on a standardized set of criteria were not viewed favorably. An ICU physician stated, “I think triggers are good I just don’t…I would not like to see them be binding. I wouldn’t want to see an automatic person show up because a trigger told them to come. I would like for us to actually talk to them and request them or allow a family to request pall care…I think they’re helpful for reminding us, but I’m not sure necessarily it should be limited to those triggers plus like I said I’m not a big believer in automatic consults.”

Participants stated that requiring some discussion between ICU and palliative care clinicians promoted collaboration between teams and were helpful for making implementation of triggers successful. In one example, triggers were used regularly to prompt further discussion between ICU and palliative care clinicians: “We have these triggers and then we do round with them every day, not bed to bed but we go through the list of patients with typically the attending, and I think that works a lot better… because certain people they could identify that there is not really a trigger for.” Both ICU and palliative care clinicians involved in this model endorsed that this system allowed for triggers to be fine-tuned by giving ICU clinicians an opportunity to request consultation for patients not identified by triggers, as well as decline consultation for patients that they felt were not appropriate. Furthermore, some ICU clinicians felt that they should be involved in approving the consultation to ensure that care was ultimately being driven by the primary ICU team with support from palliative care specialists.

4. Discussion

In this qualitative secondary analysis, we used a robust dataset containing clinically diverse viewpoints to examine ICU and palliative care clinicians’ perspectives on the use of triggers for palliative care consultation in the ICU. We identified five major themes that illuminated participants’ views on which clinical situations should serve as triggers for palliative care consultation, as well important considerations for trigger implementation. We believe our findings provide a conceptual overview of when clinicians’ felt that specialist palliative care was appropriate for ICU patients in a recent era. This builds on upon prior work, as acceptance of palliative care has grown over time [3,5–7]. Furthermore, our findings provide specific information about the ways in which trigger implementation can be successful or ineffective within the ICU setting. The diverse range of views presented by these participants serve as important foundational evidence for future implementation efforts seeking to better integrate specialist palliative care in the ICU.

Wysham et al. [8] surveyed ICU clinicians to examine attitudes about integrating palliative care specialists and found that the most acceptable triggers were for individuals with metastatic cancer, goals of care complications or family conflict, and multiple organ dysfunction; free-text data revealed that some participants felt EHR-based triggers would be burdensome and lead to “trigger fatigue.” This echoes our study findings, that even with broad agreement regarding appropriateness of specific triggers, clinicians held varied and differing views on how triggers should be used and implemented. Moreover, a critical review by Nelson and colleagues [3] highlighted existing screening criteria for palliative care consultation and made recommendations with regards to trigger selection and implementation processes that were informed primarily by expert opinion, as empirical evidence was largely lacking. Overall, the authors recommended that any process for developing and implementing triggers should involve key stakeholders and be tailored for a given ICU. In our study, participants put forth a plurality of opinions about triggers and their use, and often mentioned the need to be sensitive to cultural norms, suggesting that these recommendations remain salient.

One commonly mentioned method of deploying triggers was the EHR. While some participants felt that triggers would be best integrated into the EHR to maximize reach, others felt that an approach using automatic prompts would only result in over-consulting and undue burden on the palliative care team. A prior systematic review by Bush and colleagues [9] described how the EHR can be operationalized to integrate palliative care and found that despite there being an increase in the number of research studies seeking to utilize EHRs for palliative care, there has been a slow uptake of this method. Our study elicited many reasons that may explain this underutilization, such as frustration with automatic triggers, unwillingness to give up generalist palliative care responsibilities, inappropriate automatic consultations generated by triggers and an overall lack of trust in such a system. Despite its appeal and advantages, more research is needed to address how best to enhance the EHR to facilitate triggers for palliative care with methods that are acceptable and informed by ICU clinicians’ viewpoints and everyday practices, and to ensure that the clinical indications for palliative care generated by the EHR are commensurate with ICU clinical decision-making [8–10].

From a practice standpoint, ICU and palliative care clinicians need to be involved in planning and decision-making processes for trigger implementation. Several participants in our study highlighted the importance of stakeholder engagement to ensure that palliative care triggers are helpful, sensitive enough to catch the most appropriate cases for consultation and accepted by the ICU clinicians who will use them. Within the ICU, care teams must reach consensus on how to proceed clinically when a patient meets a trigger for palliative care consultation, as these decisions require input from all clinical team members including palliative care specialists, especially in complex scenarios of chronic critical illness [11]. Given that patients with chronic critical illness often receive complex care across different units in the hospital [12], specialist palliative care consultation may provide continuity of care during and after transitions in and out of the ICU. Future research should examine how implementation of specific triggers can be built around the needs of this population.

Workflow considerations were notable concerns of clinicians regarding trigger use. Establishing workflow procedures may aid successful implementation of triggers by reducing confusion or frustration generated by such triggers. If triggers are established and found to be ineffective, overly sensitive, or result in increased documentation burden, there should be flexibility for improvement and evaluation processes in place to address these key issues that often arise to allow for regular input from both ICU and palliative care clinicians. Moreover, future research should explore and measure ICU clinicians’ primary palliative care expertise to understand where specialist palliative care consultation might be most impactful. Similarly, understanding whether specialist palliative improves or detracts from ICU clinicians’ primary palliative care skills is an important phenomenon needing more exploration.

This study has limitations. Despite recruiting participants from different institutions, both were urban academic medical centers. It is possible that the findings are not generalizable to rural or lessresourced medical centers. While both institutions had established palliative care programs in place, it is unclear how specific differences in the programs like number of available specialty-trained palliative care clinicians [13], clinical rotating schedules, practice-related characteristics, ICU-related structures and processes, and existing implementation initiatives may impact our findings.

5. Conclusions

We identified several themes associated with the use of triggers for specialist palliative care consultation in the ICU that suggest broad support among both palliative and ICU clinicians for certain triggers and highlight important factors to consider for any trigger implementation efforts in the ICU. Our findings represent a diversity of views from a variety of interdisciplinary clinicians who endorsed that specialist palliative care in the ICU was largely indicated for patients approaching end-of-life care or having continued need for intensive care, and for supporting patients and families through those experiences. Concerns were raised related to the prospect of over-consultation, trigger fatigue, and optimal implementation within the EHR. Future research is needed to longitudinally examine the most appropriate and effective triggers and best practices for trigger implementation, and patient- and family-centered outcomes associated with trigger implementation should be examined to establish effectiveness and clinical utility [14]. These research efforts should involve all relevant stakeholders to ensure that ICU and palliative care (and other) clinicians collaboratively elucidate ways to improve palliative care delivery to critically ill patients.