Hospital Culture and Intensity of End-of-Life Care at 3 Academic Medical Centers

structures may appear to be relatively distant to day-to-day clinical decision-making, their effect can have potential wide-ranging intended and unintended consequences. This study illustrates how hospital culture might under- minetheeffectofinterventionsthatnarrowlytargetindividu-als or groups of individuals (eg, team dynamics, communications,anddecision-making).Considerationofhospitalculture and its association with individual and clinical practice patterns should be incorporated into institutional policies, practices

][38][39] Culture is defined as the shared beliefs, values, and practices of a group of people (eg, an organization, institution, or profession) and influences the typical way of doing things. 40,41nstitutions vary in treatment intensity, which appears to be stable over time. 32,427][38] There is a gap in our understanding of how hospital cultures are associated with institution-specific structures (which we defined as the hospital's policies, practices, protocols, and resource allocation) as well as how it might be associated with the intensity of end-of-life care.Institutional structures and hospital culture influence one another cyclically and iteratively. 43In sociological scholarship, this feedback loop that is perpetuated over time is described as recursion, in which 1 entity cyclically reproduces and strengthens another.The objective of this study was to elucidate our understanding of the complex, recursive relationships between hospital culture, in-stitutional structures, and the provision of potentially nonbeneficial, high-intensity life-sustaining treatments.

Design
This comparative ethnographic study was conducted at 3 academic hospitals in California and Washington that were selected for differences in intensities of end-of-life care (eg, high, medium, and low) based on the Dartmouth Atlas (Table 1).The University of California, San Francisco institutional review board approved this study, and participants underwent a written or verbal consent process.The institutional review board protocol called for anonymization of respondents and hospitals.We interviewed clinicians and administrators with different clinical backgrounds and organizational responsibilities.This project was guided by a conceptual framework based on a literature review and prior work by the authors (Figure 1 44 ). 31,36,44,45This conceptual framework was continuously refined throughout the project as more data were collected and analyzed.

Data Collection
One interviewer (E.D.), a hospitalist and PhD-trained sociologist, conducted semistructured, in-depth interviews with inpatient-based clinicians, leaders, and administrators.Interviews were conducted in person until the COVID-19 pandemic,

Key Points
Question What is the role of hospital culture and institutional structures in the provision of potentially nonbeneficial, high-intensity life-sustaining treatments near the end of life?
Finding In this qualitative study, 113 semistructured, in-depth interviews were conducted with hospital-based clinicians and administrators at 3 academic hospitals in California and Washington.Hospital culture tended to be aligned with institutional structures (eg, policies, practices, protocols, and resources), which together shaped the provision of potentially nonbeneficial life-sustaining treatments for patients near the end of life at each site.

Meaning
The findings of this study suggest that institutional cultures should be considered when developing policies and interventions to mitigate nonbeneficial, high-intensity life-sustaining treatments for patients near the end of life.when interviews transitioned to virtual video conferencing.
Respondents were purposively sampled by profession, which contributed to understanding everyday dynamics within each hospital.Recruitment occurred through group email solicitations, individual requests, and snowball sampling. 46,47The participation rate could not be calculated because persons were contacted through unsolicited emails and list-serves, along with direct solicitations.Sampling occurred until theoretical saturation was reached, ie, the point at which interviews generated no new insights. 48,49We included further details of the study methods using the Consolidated Criteria for Reporting Qualitative Research in eTable 1 in Supplement 1.
An interview guide was used across all sites; minor adaptations were made for different roles (see eAppendix in Supplement 1 for the interview guide).The initial guide was derived from our conceptual framework and evolved during the pilot and subsequent interviews.The interviews were openended; participants were encouraged to explore topics they considered relevant.Interviews were audiotaped, transcribed, and anonymized.

Analysis
Data collection and analysis occurred concurrently.During data collection, emerging findings were incorporated into ongo-ing interviews for further exploration to determine their consistency, robustness, and salience.We continually selfreflected around our own lens (eg, perspectives, assumptions, positionality) and how that might influence the research process and interpretation.We paid attention to counterfactual data, which were used to inform ongoing data collection, refine emerging themes, and enhance rigor.Respondents' diverse perspectives provided a composite view of clinical care at each hospital.Although no one respondent was able to fully explain or recognize what was attributable to hospital culture, analysis of the corpus of interviews provided these broader insights. 502][53] We deductively and inductively generated codes through line-by-line analysis and discussion and came to consensus on coding definitions. 53,54Twenty percent of the interviews were subsequently double or group coded by T.M., L.P., J.N.B., J.B., and C.B., which contributed to further refinement of the codebook.Analysis was conducted using ATLAS.tisoftware.Disagreements were resolved through discussion and clarification of code definitions until consensus was achieved.Member checking occurred through presentations, discussions, and review of manuscript drafts with cli-   2), as well as defaults toward high-intensity care (Q 4-6; Table 2).
Respondents noted that this default reflected cultural norms in the US (Q 7; Table 2).However, respondents at each of the 3 sites also described distinct hospital cultures around the intensity of end-of-life care that differed between sites and were broadly consistent with the intensity indicated by the Dartmouth Atlas data.High-intensity care was particularly notable at the high-intensity hospital (Q 3; Table 2), as was the receptiveness toward palliative care and a mindset toward deescalation at the lower-intensity hospitals (Q 8; Table 2).

Consensus and Coordination Required to Deescalate
In the absence of patient or surrogate preferences, respondents at all sites defaulted toward ICU admission if the patient met criteria, regardless of whether it was beneficial.At the high-intensity hospital, this was associated with persistent escalation, whereas at lower-intensity hospitals, there were multidisciplinary efforts to engage in shared decisionmaking within the ICU setting to deescalate (Q 9 and 10; Table 2).Institutional structures and hospital cultures at the lower-intensity hospitals facilitated deescalation; some respondents perceived it as relatively easy to deescalate lifesustaining treatments (Q 11; Table 2).Respondents at the lowerintensity hospitals, but not at the high-intensity hospital, felt comfortable not offering or strongly recommending against nonbeneficial treatments, including dialysis, pressors, and feeding tubes (Q 12; Table 2).At all hospitals, consensus and coordination was required among clinicians and families to deescalate lifesustaining treatments (Q 13; Table 2).At the high-intensity hospital, respondents described consensus around deescalation to be challenging to achieve (Q 14; Table 2).Respondents at the lower-intensity hospitals described alignment and teamwork

Medium intensity
"Sometimes I feel like it's just gotten to the point where it's like bordering on true cruelty…I would say almost everything almost always gets offered…and then we're put in a situation of having to carry out all of these things that we don't really agree with and that can be a really distressing situation."(ICU MD 6) 1 Low intensity "[Treatments are] very high intensity…we've had occasions where pretty much everybody agreed that the care that we were providing was inappropriate… the reality is that we have to work within a framework of the way our system operates."(ICU MD 4) 2 High Intensity "There is a culture from the top of the healthcare system down.We are going to go to eleven on every situation, appropriate or not.It's not explainable that each individual physician has independently arrived at this very aggressive practice pattern…I think it's the institution wants to be aggressive and so that just kind of steeps everyone in doing that kind of stuff."(ICU MD 2) 3 High Intensity "If…the son is freaking out and saying that you're killing and murdering her and that you need to code her, they're going to ask you to code her because that's the path of least resistance…you're probably going to end up coding her because [it's] the easier thing to do." (ICU nurse) "Generally it's if there is not consensus between the teams, the default is to then allow whatever decision the family makes to just ride…we'll just go ahead and allow that to evolve."(ICU MD) 13

High intensity
"I think it can be really distressing to see…[patients] suffering, and to see them getting treatments that you know that they wouldn't want…it's just really hard to see people getting forced into things that you know that they wouldn't want here.But you have essentially no control to stop it.Because if the doctors aren't going to say it, I can try, but then they may or may not listen to me, and then the family doesn't want to.I don't have a choice.I'm forced to do it."(ICU nurse 2) 14 Medium intensity "Most of the time when we say we're not going to offer a treatment such as dialysis, I can't remember a family pushing really hard to do it.Especially when we all come together as a team…There are usually multiple teams involved to reinforce that we don't think there's going to be any benefit from this." (ICU MD 6) 15 Low intensity "I would say more often than not, it feels like the various teams are on the same page.There are a few instances in which nephrology is the team to raise the alarm bells of, 'Hey, we're being asked to offer this intervention and it just doesn't make sense.And we don't feel comfortable doing it.'…wehave a discussion with our colleagues and we're all on the same page about what makes the most sense."(IM PGY-3) 16 Low intensity "I call it 'the system wins'…the system dictates where it goes…I can say, 'I don't want to do this.I don't think it's in best interest; I talk to the patient and the family.'The nocturnist who's cross covering, sees the labs and doesn't feel comfortable…and then next thing I know, they're like, 'I just decided that we should consult nephrology.'…Idon't think our system's created in a way that unless everybody's fundamentally feeling it's futile, It's almost like a majority gets what they want for the patient."(Hospitalist MD 2) 17

Cultural norms around decision points
Medium intensity "There's global input coming in from the [various]  teams…looking at the whole global picture and saying, 'This is not right.We need to really start putting the brakes on this thing.'…It'sinteresting because you'll start to bring this up, and then you'll start to see the gearbox move, and then in a day or two you can start to see it swing.The idea is implanted.I think part of that clinical momentum is that if you don't have somebody else to look at that, then that idea never gets planted."(ICU MD 5) 18 Low intensity "I feel like there are a lot of checks in place to prevent untethered clinical momentum in the ICU…It just seems like things don't naturally just continue to accumulate and progress unchecked over time.We talk with our team members and with families pretty frequently about how things are going and then what makes the most sense to do next based on what has happened so far…Everyone's very thoughtful and thorough about making medical decisions and reassessing those decisions over time."(IM PGY-3) 19 High intensity "[Doctors feel] it's not your job to decide whether or not someone's eligible and appropriate for a catheter, it's the cardiologist's…We keep thinking that the sub-specialist is going to decide for us.And overwhelmingly [the] specialists want the primary [team], to decide before they call…all the specialists I talk to have a lot of grief about feelings about forced to do these procedures."(Palliative care MD) 20 Undermining of efforts to deescalate nonbeneficial, high-intensity life-sustaining treatments High intensity "Sometimes the ethics team will make a recommendation to make the patient a unilateral DNR if the patient has a poor prognosis…so we make [that recommendation].The family then goes to the hospital leadership, and there are times when leadership has actually stopped us or stopped that unilateral DNR and said, 'Let's give it another week', which to be honest with you, it makes it challenging."(ICU nurse) 21 Committee, [the] committee agreed with us, and then the next doctor was like, "Let's just trach and PEG and get them out of the ICU.This is not worth the fight."(ICU MD 6) 24 High intensity "Bringing [cases] to the ethics committee is not a small deal.It's made up of clinicians and community members, so you present the case and the family presents the case, but it almost seems like you're sitting in court.The members are asking you questions and questioning every single judgment and stuff like that…[the ethics chair] tells me that the ethics mechanism is supposed to be hard.It's a big deal to override a family, so you want to make sure that all your i's are dotted and all your t's are crossed."(ICU MD 4) 25 Low intensity "There's still this discomfort around pushing against the aggressive care motto that we tend to default into.The ethics committee…can be this extra layer of support to clinicians to validate and verify the decisions that are being made…They liaise very closely with risk management because sometimes in these scenarios, a decision to not escalate or a decision to withdraw interventions can raise these flags for our institution…the ethics committee manages up and down and makes everyone feel like the decision is okay on multiple levels…they have a very low threshold to sort of get the institutional leadership involved…they have a good relationship and partnership with the administrators who also happen to be phenomenal clinicians themselves."(PC 5) 26 Medium intensity "The most surprising thing about the role that we play in risk management is the fact that unlike many risk managers we're often telling physicians 'Please don't give care that's medically ineffective if it goes against your conscience or isn't in the best interest of the patient.'It would be easier sometimes to just let it go on and on, but we advocate for people to please do the right thing, and that is not very common for a risk manager.The profession is reputationed as being more risk adverse so I think my approach and my office's approach is kind of different than a lot of risk managers."( among and within care teams to achieve consensus to deescalate and present a unified message to families.Potentially nonbeneficial high-intensity treatments at lower-intensity hospitals were discussed and sometimes resisted (Q 15 and 16; Table 2).In contrast, consensus was not always required to escalate life-sustaining treatments (Q 17; Table 2).

Cultural Norms Around Decision Points
At the lower-intensity hospitals, respondents noted that because multiple people were involved in decision-making, there were multiple checks to ensure that the treatments provided were beneficial.Respondents described a cultural norm that encouraged a shared desire toward appropriate deescalation (Q 18 and 19; Table 2).They recognized that defaults existed toward high-intensity life-sustaining treatments and described intentional mindsets and actions to resist, which required time and effort (Q 9; Table 2).At the high-intensity hospital, respondents noted that the involvement of multiple people in decision-making was associated with a diffusion of decision-making responsibility that allowed for unchecked momentum toward a high-intensity care (Q 20; Table 2).

Undermining of Efforts to Deescalate Nonbeneficial, High-Intensity Life-Sustaining Treatments
Despite efforts to deescalate, respondents described undermined attempts at deescalation of nonbeneficial, highintensity life-sustaining treatments at all hospitals by external entities, such as consultants, ethics committees, or hospital administrators (Q 21-23; Table 2).However, in the respondent sample, the theme of undermined deescalation was particularly notable at high-intensity hospitals (Q 21-22; Table 2).Transitions of care, such as attending physician turnover, were described as occasions in which reescalation of high-intensity treatments occurred (Q 24; Table 2).The ways that ethics committees made decisions and supported clinical teams appeared to be aligned with measured end-of-life care intensity.At the high-intensity hospital, respondents described the ethics committee's approach as favoring patient autonomy and placing additional burdens on clinical teams that were seeking to deescalate (Q 25; Table 2).At the lower-intensity hospitals, respondents reported that the ethics committees worked with clinicians and institutional leadership to support clinicians' clinical judgment (Q 26; Table 2).At 1 of the lower-intensity hospitals, respondents reported that risk management encouraged clinicians to make decisions based on the patient's best interests and supported them institutionally and legally to do so (Q 27; Table 2).

Shaping of Hospital Culture Through Institutional Policies, Protocols, Practices, and Resources
At the lower-intensity hospitals, protocolized approaches and hospital policies counteracted the momentum toward potentially nonbeneficial, high-intensity life-sustaining treatments.Respondents felt that institutional structures (eg, policies, protocols, practices, and resources) successfully created and sustained hospital cultures that supported goal-concordant end-of-life care at these hospitals (Q 28 and 29; Table 3).Respondents felt that palliative care services demonstrated their value as important and trusted sources of support and education not only for patients and families, but also for clinicians (Q 30; Table 3).At the lower-intensity hospitals, respondents described consultants (Q 16; Table 3), social services (Q 9; Table 3), hospital leadership (Q 31 and 32; Table 3),

Shaping of hospital culture through institutional policies, protocols, practices, and resources
Medium intensity "It's just part of our culture to consider palliative care…there are changes in the ICU like the comfort care huddle to have a more systematic approach…before meeting with the family to make sure that [everyone is] on board…have we explored all the options?Do we all feel that there's consensus on a particular recommendation?"(ICU MD 4) 28 Low intensity "There are now little checkpoints for teams who have previously been on the train that keeps going and doesn't stop…we have this [automated] algorithm that says make sure that if patients meet these criteria that you consult palliative care…within the first 24 hours."(PC MD 5) 29 Low intensity "Death rounds [is where the]…palliative team provides for our residents to talk about really difficult cases or deaths…it's really important that they provide that space for the resident team…it was just so traumatizing.They would also just be present within our COVID ICU…[to] check in with our team and offer debrief sessions…that kind of support…was really important for our team as a whole."(ICU APC 2) 30 Medium intensity "I have never experienced a time when medical administration has in any way got involved in any of these decisions.I wouldn't say that I've ever felt financial pressure or press pressure…it's just absent at this institution…" (Ethics consultant)

31
Low intensity "And it's normally not a 'you shouldn't do this.' [from hospital leadership], it's usually, 'our suggestion right now would be to wait or like to give more time'…it's a supportive culture…[for] decision making…it's mostly on our team."(ICU APC 2)

32
Relational and emotional aspects of end-of-life care High intensity "I think it takes away a piece of your soul.You feel horrible about it because you're not accomplishing anything good in the long run, which is why I think a lot of physicians don't call ethics.They don't even fight that battle.Why fight that battle?Just do your shift and move on."(ICU MD 5) 33 High intensity "It's like an exercise in futility almost…I think those cases are few.It's just that when they occur it just takes so much time, and energy, and work.Sometimes I wonder…is it worth it to even go through that whole process?" (ICU MD 5) 34 Low intensity "I really value ending the situation on the same team as the family or the patient… I've got to align with this family.That's my number one priority…if I am at odds with the patient or the family and they feel like they're not getting the care that they want, I feel like that's a much bigger loss in my mind than someone getting CPR when they're 85 years old…it's just far sadder to me when they say, 'Those doctors didn't care' or 'they didn't try to save my mom'."(PC MD 1) 35 Medium intensity "I think I'm aligned with the culture here…I don't know anybody who would disagree with the fundamental principles of building relationships…I really try to reserve unilateral decision-making as a last resort.As physicians we're making the final sacrifice to do what we think is best for the patient.But there's so much struggle and suffering that is going to continue after a patient dies with the family…if we can align with the families…I'm willing to wait and go the extra mile to try to achieve that."(Hospitalist MD 1) ethics committees (Q 9 and 26; Table 3), risk management (Q 27; Table 3), and other entities as facilitating consensus rather than undermining deescalation efforts.These institutional entities encouraged clinicians to deescalate when ethically appropriate and deferred to clinician judgment (Q 32; Table 3).

Relational and Emotional Aspects of End-of-Life Care
There was a notable emotional toll, especially at the highintensity hospital, that was associated with challenging cases that affected respondents' future willingness to attempt deescalation.Respondents described feeling powerlessness, particularly when institutional entities thwarted attempts at deescalation and asked them to provide potentially nonbeneficial treatments.Respondents believed that their efforts were not worth the fight (Q 33-34; Table 3).Although emotionally fraught and challenging cases that involved intense conflict were relatively infrequent, the potential presence of these conflicts was a concern for clinicians long afterwards.Respondents noted that the negative emotional valence surrounding prior efforts was followed by less willingness to expend effort to deescalate care in the future.
One theme that emerged only at the lower-intensity hospitals was the way end-of-life decision-making occurred.Respondents at these hospitals reported that they achieved deescalation by seeking nuance between the extremes, finding ethical middle grounds between providing all possible treatments and unilaterally not offering nonbeneficial, highintensity life-sustaining treatments.Respondents described focusing on relationship building and aligning with the family.They recognized the potentially traumatic and burdensome nature of these decisions.Although an informed assent approach, in which clinicians explicitly offer the choice to defer to a clinician's judgment about withdrawing or withholding lifesustaining treatments, was an accepted norm at the lowerintensity hospitals, it was used within a context of empa-thetic, values-oriented goals of care discussions.Respondents frequently noted that this approach sometimes took more time but was the right thing to do (Q 35 and 36; Table 3).

Discussion
At the 3 hospitals examined in this qualitative study, hospital culture appeared to recursively shape and reflect institutional structure, as manifested by its policies, practices, protocols, and resource allocation.Each hospital's culture and institutional structure either supported or undermined attempts to deescalate against the default of high-intensity lifesustaining treatments in US medicine.A clinicians' attempt to deescalate appeared to be more effective if they operated within a hospital culture that discouraged high-intensity lifesustaining treatments.The aggregate effects of a hospital's culture and institutional structures appeared to coalesce into an institutionally specific equilibrium that created and reproduced hospital culture.Overall, these observed patterns were consistent with our conceptual model (Figure 1) regarding how hospital culture and institutional structures recursively reproduce and strengthen one another.
In Figure 2, we illustrate a prototypical patient trajectory of how recursive reproduction and strengthening between institutional structures and hospital cultures might be associated with the ease of de-escalation.A default toward highintensity treatments occurred unless every element of a hospital's culture, as manifested by specific structural and procedural factors (ie, institutional structures), was aligned to resist this default.We described in the Results section instances in which hospital leaders, administrators, subspecialty consultants, and ethics committees undermined deescalation.As such, successful deescalation of nonbeneficial, high-intensity life-sustaining treatments required every individual involved to be aligned toward deescalation.If  1 individual undermined an attempted deescalation, respondents reported that the care trajectory regressed to the default of high-intensity treatments.
It is well known that the US is an outlier in its default toward high-intensity treatments. 55,56Respondents at all sites needed to use active and concerted efforts to deescalate nonbeneficial, high-intensity life-sustaining treatments.The tendency across hospitals was to escalate treatments unless there was explicit agreement with all relevant decision-makers to de-escalate.A relevant concept is that of clinical momentum, or clinical practice norms and patterns of usual care that promote the accumulation of multiple interventions over time. 57ven when clinicians, patients, or families resisted this default toward high-intensity care, multiple overlapping institutional structures (eg, clinical practices of the primary team or consultants, hospital policies) made it challenging to do so.
Dynamics that encouraged escalation at the high-intensity institution were similar to dynamics that facilitated deescalation at lower-intensity hospitals.As we described (eg, Q 20; Table 2), respondents at the high-intensity hospital described a diffusion of decision-making responsibility to multiple clinicians as encouraging escalation.Respondents at lower-intensity institutions described (eg, Q 16 and18; Table 2) similar involvement of multiple decision-makers and teams as positive checks and balances that facilitated multiple opportunities to raise concerns.Another example of similar dynamics that were associated with divergent clinical practice patterns occurred when respondents at the high-intensity hospital (eg, Q 22; Table 2) described consultants as undermining deescalation, whereas consultants helped mitigate unchecked clinical momentum at the lower-intensity hospitals (eg, Q 15; Table 2).Studies examining clinical practice patterns often focus on team dynamics and interteam structures. 58,59The results of the current study suggest that hospital culture should also be considered.Hospital culture manifests in many ways, such as the convergence of ethical perspectives around appropriate care among all members of the multidisciplinary team, or a preponderance of clinicians willing to attempt deescalation.
These findings potentially provide insights into the recursive relationship between institutional policies and practices that are designed to mitigate potentially nonbeneficial, high-intensity life-sustaining treatments and institutional culture by creating feedback loops.Deliberate hospital policies and protocols that encourage thoughtful pauses around treatment escalation decisions potentially help mitigate unchecked clinical momentum.The comfort care huddle described at one of the lower-intensity hospitals allowed for all members of the team to convene at regular intervals to discuss treatment deescalation opportunities.This is reminiscent of a protocolized process in France (limitation et arrêt des traitements) that we had previously described as beneficial in their ability to align team members toward a unified message with families. 45The ways these interventions, policies, and practices cyclically reinforce and are reinforced by hospital culture to affect care intensity is an example of the recursiveness of social processes.
Although extreme and rare, the results of this study suggest that challenging clinical cases had an outsized role in shaping clinicians' perceptions and willingness to engage in future difficult cases.This is another example of recursion, but in this case between the individual and the hospital's culture.Shared beliefs and experiences recursively reproduce cultural orientations to strengthen hospital culture.This feedback-feedforward reproductive process is a general feature of social life, as noted by Giddens' structuration theory 43 in which individuals' experiences and subsequent actions are influenced by institutional constraints in ways that reinforce hospital culture.

Limitations
This study was limited in its ability to ascribe causation from hospital cultures to care intensity.While this study potentially helps elucidate the dynamics between hospital culture, institutional structures, and care intensity, there may be other differences between hospitals, such as uptake of palliative care, economic incentives, and patient population demographic characteristics, that are associated with care intensity.While broader macrosociological, political, and economic forces affect individuals and institutions, we were not able to specifically examine those phenomena in this article beyond a general perception by respondents that these forces contributed to a default of high-intensity care.Our observations were based on 3 urban academic medical centers and may not generalize to community hospitals.

Conclusions
In this qualitative study, we described the significance of hospital culture and institutional structures in resisting the default toward high-intensity life-sustaining treatments.This study highlights the importance of the deliberate design of institutional structures (eg, policies, practices, protocols, and resource allocation) in mitigating the harmful effect of entrenched societal forces and defaults within the US health care system.These efforts should include purposeful consideration of how institutional values might reflect and be reinforced by specific policies and procedures, such as ethics committee decision-making processes and the structure of comanagement and consultation services.We also suggest attention to the relational aspects of care, including risk management values, thoughtful and consistent dialogue between institutional leaders and clinical teams, and attention to the nature of administrator oversight in high-stakes clinical decisions in extreme cases.While these entities' values and structures may appear to be relatively distant to day-to-day clinical decision-making, their effect can have potential wide-ranging intended and unintended consequences.
This study illustrates how hospital culture might undermine the effect of interventions that narrowly target individuals or groups of individuals (eg, team dynamics, communications, and decision-making).Consideration of hospital culture and its association with individual and clinical practice patterns should be incorporated into institutional policies, practices, and interventions.The design of institutional policies, protocols, practices, and resource allocation have the potential to shape hospital culture and, thereby, individual clinician behavior and patient and family experiences.

Figure
Figure 1.Conceptual Framework for the Association and Interactions Between Hospital and National Cultures and Policies ] we don't have a family member…or some other form of documentation…the assumption [is] of full court press and do everything that we can." (ED MD 1) 5 Low intensity "Until the family members or the patient [decides], then [care] will be typically intensive."(ICU MD 2) 6 Low intensity "In the United States…everyone has the opportunity to receive full care to the maximum amount of our abilities until they say no, or until, essentially, their body tells us that they can no longer tolerate it…that is 'the right thing to do'….it is [an] American right…There's a lot of conversations about 'If you don't do this, I will report you.I'll bring you to the media'…'sue you,' kind of thing.That conversation comes up with disgruntled family members here often."(ICU APC 2) 7 Low intensity "I think it's almost like passive diffusion of the knowledge and style and vision of palliative care.I see that there's a lot more palliative care consults happening…I think you can't help but learn their approach…some of that knowledge or management rubs off on the primary teams."(Hospital MD 2) 8 Consensus and coordination required to deescalate Low intensity "The default [for unbefriended patients] is the aggressive care pathway.But I think what ends up happening is we get to know them and their character…We try to reach out to case managers, to SNF providers… people in the community…but that takes time… [if] we are unable to track anyone down…we ultimately get our ethics committee involved and do a deeper dive into how we should proceed."(PC MD 5) 9 Low intensity "We are not interested in blocking or making the upfront decision-making hard to get into the ICU.We'd much prefer to be open to bringing people into the ICU and then do a good job with care decision-making there."(ICU MD 3) 10 Low intensity "I think in our institution, it's pretty easy to move to CMO (Comfort Measures Only)…and deescalate.That tends to be well-established."(ED MD 2) 11 Low intensity "We don't have to offer certain things…we may in the course of discussions say, 'This is perhaps not the most appropriate patient to offer renal replacement therapy to.'" (ICU MD 2) 12 Low intensity

Figure 2 .
Figure 2. Prototypical Patient Trajectory for How Institutional Structures Can Be Associated With the Ease of Deescalation of Nonbeneficial, High-Intensity Life-Sustaining Treatments

Table 1 .
42spital Selection Based on Dartmouth Atlas Indices Around Intensities of End-of-Life Care42 a Hospital Care Intensity Index score was based on the number of days a patient spent in the hospital and the number of physician encounters they experienced as inpatients and constitutes the ratio of a given hospital's utilization rate compared with the national average.
1. Conceptual Framework for the Association and Interactions Between Hospital and National Cultures and Policies

Table 2 .
Table of illustrative Quotations

Table 2 .
Table of illustrative Quotations (continued)

Table 2 .
Table of illustrative Quotations (continued) The family would not let go.She was dying for months and had no other treatment options left and [she] suffered horrendously…even though she had a DNR, I knew that it wouldn't be followed…our attending said, 'She is no longer a candidate for [treatment] so if she decompensates… do not offer [treatment]…' So I stood in front of the room for hours, and every doctor that came in, I said, 'We're not offering [treatment].Don't bring it up.'…Finally[the family started] to accept it…Later in the day, I came back and the [specialist] attending was in the room.The family said, 'Well she's not a candidate for [treatment].'[The [specialist] attending] then said, 'Well, yes, we can do that [give treatment].'I was devastated.I'm not sure I've ever been so devastated in my whole life…[The attendings] got into a screaming match in the hallway…She eventually went back on [treatment], she coded and died.They did CPR on her." (ICU nurse) The ethics committee was like, '[De-escalation is]…appropriate', but when they went to the hospital administration, they said, 'We can't.What if some family member emerges out of the background and then we made a decision to withdraw care.Let's just not escalate.'…evenif all the clinicians agree, the institution puts a block on certain decisions because of their concerns around litigation."It made the whole system look like a fool.The whole ethics [committee] mechanism kind of fell apart before our eyes [in] that we brought this to [Ethics]