Advance care planning in older patients with acute myeloid leukemia and myelodysplastic syndromes

Abstract

Introduction:

Older patients with acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS) have worse survival rates compared to younger patients, and experience more intense inpatient healthcare at the end of life (EOL) compared to patients with solid tumors. Advance care planning (ACP) has been shown to limit aggressive and burdensome care at EOL for patients with AML and MDS. The purpose of this study was to better understand ACP from the perspective of clinicians, older patients with AML and MDS, and their caregivers.

Methods:

We conducted semi-structured interviews with 45 study participants. Interviews were audio-recorded and transcribed. Open coding and focused content analysis were used to organize data, develop, and contextualize categories and subcategories.

Results:

Guided by our specific aims, we developed 4 themes, 1) The language of ACP and medical order for life-sustaining treatment (MOLST) does not resonate with patients, 2) There is no uniform consensus on when ACP is currently happening, 3) Oncology clinician perceived barriers to ACP (e.g., patient discomfort, patient lack of knowledge, and lack of time), and 4) Patients felt that they are balancing fear and hope when navigating their AML or MDS diagnosis.

Discussion:

The results of this study can be used to develop interventions to promote serious illness conversations for patients with AML and MDS, and their caregivers, to ensure that patient care aligns with patient values.

Introduction:

Acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS) are common hematologic malignancies in older adults (1, 2). Previous population-based studies found that survival rates for older adults (≥60 years old) with AML and high-risk MDS are poor, compared to their younger counterparts (3, 4). For example, three-year survivals for patients with AML aged 0–18, 19–59, and ≥60 years old are 36%, 28%, and 8%, respectively (4). Compared to those with solid tumors, older adults with AML and MDS experience greater healthcare utilization at the end-of-life (EOL), including hospitalizations, intensive care unit (ICU) admissions, and use of life-sustaining treatments (5–7). They are also more likely to die in the hospital and less likely to use palliative care services or enroll in hospice, some of which can be attributed to transfusion dependency and the acuity of a leukemia diagnosis (5–10). Utilization of aggressive in-hospital services at the EOL is considered sub-optimal, based on well-accepted oncology quality metrics and previous studies (11–13). Improving care at the EOL is particularly important for this patient population.

Advance care planning (ACP) is a process that supports adults in understanding and sharing their personal values and preferences regarding medical treatment. Depending on the patient’s health status, ACP may include completion of advance directive forms [e.g., living will, Physician Orders for Life-Sustaining Treatment (POLST), and healthcare proxy (HCP)], in addition to exploration of the patient’s personal values (14). While ACP has the potential to improve the EOL experience for patients with AML and MDS, the frequency of its use is low for this patient population (15).

POLST forms assist patients who are seriously ill in recording their preferences for medical treatment in emergencies or if they can no longer make decisions for themselves (POLST, polst.org). MOLST is the New York State endorsed national POLST Paradigm Program (MOLST, molst.org). We previously found that early completion of a MOLST form was associated with decreased utilization of aggressive in-hospital services at EOL for patients with AML and MDS and increased hospice enrollment (16) and similar outcomes have been reported for patients with solid tumors (17, 18). Therefore, completion of advance directive forms serves as a means to elicit conversations surrounding patient preferences. Among patients with various hematologic malignancies, those who had a documented goals-of-care discussion more than 30 days before their death were less likely to be admitted to the ICU or die in the hospital and more likely to enroll in hospice (19).

Challenges to ACP for patients with cancer have previously been described (20–22). Challenges included patient understanding and expectations around their diagnosis, identifying the right time for ACP discussions, and barriers to patient-clinician communication (20–22). Communication barriers included time constraints, poor health literacy, and lack of formal training for clinicians on conducting ACP conversations (20, 23). Patients with hematologic malignancies, such as AML and MDS, face additional unique challenges due to the often acute, rapid, and uncertain nature of their diagnosis or prognosis (24–26). Furthermore, older adults with AML and MDS are also at risk of age-related functional and cognitive decline (25–27). Therefore, it is imperative identify possible solutions to improve ACP so that it can be implemented in clinical practice for older patients with AML and MDS.

The purpose of this study was to understand clinician, patient, and caregiver perspectives on ACP (including advance directive completion). Specifically, we sought to understand current ACP practices, identify challenges to ACP, and elicit strategies to improve ACP. The feedback from this study will be used to adapt an ACP intervention for older adults with AML and MDS to improve frequency and quality of ACP discussions for this patient population.

Methods:

Study Design, Population, and Setting

We conducted a single-center qualitative study and recruited oncology clinicians, palliative care clinicians, patients with AML or MDS, and their caregivers, if available, from the Wilmot Cancer Institute (WCI) and its affiliated community practices. Inclusion criteria were 1) Oncology and palliative care clinicians –physicians, nurses, or advanced practitioners (APPs)–who cared for at least one patient aged ≥60 years with AML/MDS in the past year; 2) Patients aged ≥60 years with a diagnosis of AML or MDS and able to provide informed consent [conventional definition of “older” in this population (28, 29)] and 3) Caregivers selected by the patient when asked if there is a “family member, partner, friend, or caregiver [age 21 or older] with whom you discuss or who can be helpful in health-related matters” and able to provide informed consent. Patients could enroll in this study with or without a caregiver. Caregivers were identified by enrolled patients, and therefore required enrollment of their patient in order to participate. The study was approved by the University of Rochester Medical Center Institutional Review Board.

Study Procedure and Data Collection

Data were reported according to the Consolidated Criteria for Reporting Qualitative Research (COREQ) guidelines (Supplemental Table 1). Consenting study participants completed baseline measures and participated in a 60-minute semi-structured interview (Supplemental Table 2).

Data Analyses

Descriptive statistics were used to summarize demographics (all participants) and clinical characteristics (patients). Interview transcripts were imported into MAXQDA 2020 (VERBI Software GmbH, Berlin, Germany), a qualitative research coding software. Two independent coders utilized open coding and focused content analysis (30, 31). An initial coding schema was developed using the first two transcripts listing/specifying codes identified during the open coding as well as a priori codes central to our aims; this schema was refined during subsequent analysis. Thematic saturation (defined as no emergence of new themes in transcript analysis) was achieved for patients and clinicians. Consensus was reached through discussion of coder’s independent ideas, until agreement on best representation of data was achieved. We included caregivers (if available) to provide additional perspectives. We did not aim to achieve thematic saturation for caregivers.

Results:

We approached 24 patients and 20 consented (consent rate: 83%); five patients withdrew between consent and baseline assessment/interview. Final sample size was 15 patients. Six caregivers were available for consent and 1 withdrew between consent and baseline assessment, resulting in a final sample size of 5 caregivers. Consent rate among clinicians was 89% (25/28). Consented clinicians completed all study procedures.

Demographics

Mean ages of oncology clinicians (n=16), palliative care clinicians (n=9), patients (n=15), and caregivers (n=5) were 46, 48, 70, and 69 years, respectively. Eleven of the patients were interviewed within a year of diagnosis; four patients were interviewed more than one year post-diagnosis. Table 1 describes participant demographics. Table 2 describes the clinical characteristics of patients.

Table 1:

Participant Demographics

	Oncology Clinicians n = 16	Palliative Care Clinicians n = 9	Patients n = 15	Caregivers n = 5 (1 caregiver missing data)
Mean age in years (SD, range) Median age in years (IQR)	46 (11, 30 – 67) 45 (14.8)	49 (11, 33 – 67) 44 (11.5)	70 (5, 61 – 80) 72 (6.0)	69 (5, 62 – 75) 70 (4.8)
Female (%)	9 (56.3)	2 (22.2)	6 (40.0)	3 (75.0)
Race  White (%)  African American (%)  Asian (%)  Not reported (%)	14 (87.5) 1 (6.3) 0 (0.0) 1 (6.3)	8 (88.9) 0 (0.0) 1 (11.1) 0 (0.0)	14 (93.3) 0 (0.0) 1 (6.7) 0 (0.0)	4 (100.0) 0 (0.0) 0 (0.0) 0 (0.0)
Ethnicity  Hispanic or Latino  Not Hispanic or Latino	1 (6.3) 15 (93.7)	0 (0.0) 15 (100.0)	1 (6.7) 14 (93.3)	1 (25.0) 3 (75.0)
Mean years in practice (SD, range) Median years in practice (IQR)	13 (12, 0 – 40) 10 (19.0)	15 (10, 1 – 34) 14 (13.0)	N/A	N/A
Physician (%)	7 (43.8)	7 (77.8)	N/A	N/A
Advanced practice practitioner (%)	4 (25.0)	2 (22.2)	N/A	N/A
Nurse (%)	5 (31.3)	0 (0.0)	N/A	N/A
Level of education (%)  High school  Some college  College  Postgraduate	N/A	N/A	4 (26.7) 3 (20.0) 2 (13.3) 6 (40.0)	0 (0) 1 (25.0) 2 (50.0) 1 (25.0)
Marital status (%)  Married  Committed significant other  Divorced  Widowed  Single	N/A	N/A	7 (46.6) 1 (6.7) 3 (20.0) 3 (20.0) 1 (6.7)	3 (75.0) 1 (25.0) 0 (0.0) 0 (0.0) 0 (0.0)
Employment status (%) (1 patient missing)  Retired	N/A	N/A	14 (100.0)	4 (100.0)
Living at home with patient (%) (1 patient missing)  Partner  Friend  None	N/A	N/A	7 (50.0) 1 (7.1) 6 (42.9)	N/A
Caregiver not living with patient (%) (1 patient missing)  Children  Partner  Other relative  None	N/A	N/A	2 (14.3) 1 (7.1) 1 (7.1) 10 (71.4)	N/A
Relationship to patient (%)  Partner	N/A	N/A	N/A	4 (100.0)

Table 2:

Clinical Characteristics of Patients

	Patients (n = 15)
Diagnosis (at time of interview) (%)  Acute myeloid leukemia  Myelodysplastic syndrome	9 (60.0) 6 (40.0)
Eastern Cooperative Oncology Group (ECOG) performance status (at time of interview) (%)  0  1  2  3 or 4  Not documented	8 (53.3) 3 (20.0) 3 (20.0) 0 (0.0) 1 (6.7)
Initial treatment at diagnosis (%)  Intensive treatment  Lower intensity treatment  Best supportive care	4 (26.7) 8 (53.3) 3 (20.0)
Comorbidities  Hypertension  Hyperlipidemia  Diabetes  Depression	8 (53.3) 4 (26.7) 4 (26.7) 5 (33.3)
Antecedent hematologic malignancies  Myelodysplastic syndrome  Chronic myelomonocytic leukemia  Essential thrombocytosis  Myeloproliferative disorder  None	2 (13.3) 1 (6.7) 1 (6.7) 1 (6.7) 11 (73.3)
Time from diagnosis  Less than 1-year  Greater than 1-year	11 (73.3) 4 (26.7)

Four themes emerged from the data: 1) The language of ACP and MOLST does not resonate with patients, 2) There is no uniform consensus on when ACP is currently happening, 3) Oncology clinician perceived barriers to ACP, and 4) Patients felt that they are balancing fear and hope when navigating their AML or MDS diagnosis. Most representative quotes are shown in Table 3.

Table 3:

Major Themes and Subthemes

Themes	Exemplar Quote
1. The language of ACP and MOLST does not resonate with patients	Patient 1: “The only time I’ve ever talked about those things was before something – when I was just in the hospital, so it must’ve been one of those surgeries, I suppose. Someone asked me about chest compressions and then did care for my answer.” Patient 8: “I don’t think people understand advance care planning and I had to be walked through it.” Patient 12: “I’ve done some of it with people, I’ve done a little bit myself. Every time you have a surgery, they want a healthcare proxy and sometimes and DNR or other kinds of things.” Caregiver 3: “Well, we haven’t had those conversations…the only conversation that we’ve had is I asked questions…which is, what are you looking at as far as treatment goes and how far along are we with treatment?”
2. There is no uniform consensus on when ACP is currently happening	Patient 8: “Not at the end, when a person is so sick that they can’t speak. When they’re more healthy than down the road when they might not be able to answer those questions that should be done early on.” Patient 10: “I would’ve liked more transparency in the beginning. When I first came here, materials were not available or given. There are so many things available through the cancer society and MDS in particular, support groups and information pamphlets.” Oncology clinician 2: “There are some patients, based on what I am seeing when I first meet them…we’ll discuss that this is not going to be compatible with long term survival. There are other patients where…I don’t have this discussion until something acute happens or they’re in the hospital”. Oncology Clinician 4: “For patients with treatment options, I usually lay out the treatment options and say what I hope to accomplish with all of these treatments…but you know, I try and have a caveat that all of them are going to stop working and then that at some point, we’ll need to make decision about what to do at that time.” Oncology Clinician 6: “If it was truly an acute thing of course, I want to get that addressed. You know, I’m very worried I don’t want you to be suffering and I feel like you’re suffering. We have to establish this now. I’m happy to do that. In an ideal setting, where you have that time, I would be okay foin it on telehealth. In fact, I’ve done it during COVID, with the intent that the appointment was solely for that discussion.” Oncology Clinician 8: “If I start seeing the lab work not getting better or signs that things aren’t going the way I think that should be, if they’re not responding to therapy or that they’re starting to decline, then I start preparing the patient.” Oncology clinician 9: “I think the best way to improve it is to do what you’re trying to accomplish here, to discuss it and plan for it as early as possible because so many times we’ve seen in the 11th hour, trying to scramble to get hospice evaluations or scramble to get home hospice or home care needs in place.” Oncology Clinician 16: “I guess I’m just speaking for my teams, but we’re not regularly looking a patient code status until they’re admitted to the hospital, or we think that things are going the wrong direction.” Palliative care clinician 7: “Putting full code on a MOLST form doesn’t do anything except makes it difficult for surrogate decision makers to make them DNR/DNI when they lose capacity.”
3. Oncology clinician perceived barriers to ACP
Clinician perceived patient discomfort	Oncology Clinician 3: “It’s something that everybody’s slightly more uncomfortable discussing than treatment, or at least focus on the positive that patients and families look for.” Oncology clinician 4: “Preexisting conceptions among patients in terms of success rates of resuscitative stuff…and there’s a general perception that if we have this conversation, it’s failure or giving up.” Oncology Clinician 9: “I think the biggest challenge is it’s an uncomfortable conversation to have, not just for the provider, but for the patient too. I mean, they’re oftentimes just told that they have probably a. terminal disease and they’re coping with that, and then to address goals of care, it’s got to be emotionally stressful for them.”
Clinician perceived patient lack of knowledge	Oncology clinician 11: “I think that it’s a knowledge gap that’s the biggest challenge. They don’t understand what these kinds of cancers even are, like lung cancer, colon cancer, it’s more tangible I think for that kind of concrete thinking, and they don’t understand what this means, like blood cancer.” Oncology Clinician 15: “I think some of them is just the fact that they don’t expect it because it is not out there in the community. People don’t talk about these things all the time so being the first time, it’s kind of difficult to grasp the concept of that – like, what are you talking about? Basically, it’s like a foreign language to them.”
Time	Oncology clinician 6: “But our system, our outpatient and inpatient system, is not designed for time, it’s designed for efficiency. And that those two things don’t mix well and so, you know, we do these things either on an off-clinic day or at the very end of the day so that you don’t have these external pressures on you to move quickly, when you that this patient and their family are in front of you and need this kind of discussion.” Oncology Clinician 12: “There a lot of consideration and time and it should be, when discussing it, but there’s no time to have that much of an in-depth conversation just in clinic.” Oncology Clinician 13: “Do we have the time in the beginning when they’re trying to understand their diagnosis and understand all these things and are they even open to any of this in the beginning? No, definitely not. But then you start going to treatment, you start doing all these things, and yet this is something that gets, if were talking about addressing it sooner, I don’t know how because from our end, I don’t even see how there’s time.”
4. Patients felt that they are balancing fear and hope when navigating their AML or MDS diagnosis	Patient 1: “I don’t want my kids to ever feel any guilt over decisions they’ve made, or my husband.” Patient 2: “I was too scared to know everything.” Patient 3: My generation was a ‘hush-hush’ generation. They really have to be pretty good at pulling and feeling that out. They would have to have some kind of question they could slide to them to gain their trust.” Patient 4: “That’s the secret right there. I think positive attitude is the greatest thing in the world because mentally you can destroy yourself. Mentally you get sicker. The more you look at a disease as yes, I know I’m not going to survive, but I don’t know how long. Don’t tell me I could live another 20 years. Nobody knows. The thing is, the more positive a person is treated, the more positive results you’re going to get in their healthcare.” Patient 6: “I think the patient has to be proactive in their own healthcare.” Patient 6: “I need the encouragement. I need the prayers. I need this. I don’t need negativity.” Patient 9: “You’ve got to give the patient hope.” Patient 7: “I’m thinking that if I don’t think about it, maybe the positive stuff will stay because that’s talking negative.” Patient 8: “To know that the person has support and that everybody on that team is rooting for you.” Patient 12: My hope would be that my medical team would not abandon me when there’s no longer anything they can do to preserve my life.” Patient 14: “My biggest fear is not having enough coherence to make a decision for myself.”

The language of ACP and MOLST does not resonate with patients.

While majority of patients had identified a health care proxy, they were not familiar with the term “advanced care planning.” Patients felt that the components of ACP were unclear, making ACP a difficult concept to fully comprehend. Many patients were unaware that ACP extends beyond completion of advance directives and includes exploration of personal values. Many patients thought that having a medical background would have made ACP more comprehendible. Some patients recalled interactions with clinicians and social workers who took the time to explain ACP in detail. They found that these interactions clarified that ACP is designed to help clinicians become aware of patient medical preferences, life goals, and personal values.

Most patients had not heard of a “MOLST form,” and required further explanation through which they realized that they had been asked about their wishes for resuscitation by a healthcare professional. Of the 15 patients in this study, four of them did complete a MOLST form.” Patients recalled being briefly asked if they would want resuscitation by emergency room clinicians, without formally completing a MOLST form. Patients did not identify previous more in-depth discussion with a clinician sharing their resuscitation and intubation wishes.

For patients that had completed a MOLST form, they felt that the MOLST form played an important role in their care. Patients found resuscitation to be a feared medical treatment, and therefore the MOLST protected them from that experience. Furthermore, patients felt that the MOLST form lifted burden off their loved ones. Patients feared that their loved ones would feel guilty making care decisions for them.

There is no uniform consensus on when ACP is currently happening.

Patients felt that ACP should take place early, ideally following in-depth treatment planning. From the clinician perspective, there is currently no clear consensus on when ACP happens for patients with AML and MDS. Oncology clinicians individualize the timing of ACP for each patient. ACP is started earlier for older patients with worse prognoses and later for patients with stable disease or those pursuing treatment with curative intent. Oncology clinicians frequently bring up ACP in the context of treatment. Specifically, they discuss ACP to prepare for potential treatment complications (earlier initiation) or to address treatment failure (later initiation).

Oncology clinicians consider their relationship with the patient when deciding whether to initiate an ACP discussion. Oncology clinicians felt that it was important to establish trust with the patient, prior to addressing their medical preferences and personal values. They also considered the patient’s emotions in each encounter. Oncology clinicians felt that they did not want to overwhelm patients with this emotional conversation, and therefore believe that it is more appropriate to initiate ACP at visits where the patient is less overwhelmed by other factors (i.e., treatment planning, worsening symptoms).

Currently, there is not routine implementation of ACP for patients with AML and MDS in clinical practice. ACP currently varies across clinicians. Nonetheless, oncology clinicians had uniformly agreed that, ideally, ACP should happen early in the disease process. All study clinicians emphasized the importance of earlier ACP discussions. Oncology clinicians suggested dedicating the second or third outpatient visit to ACP.

Additionally, there was no clear consensus on when MOLST form completion should happen across medical disciplines (oncology vs. palliative care). Oncology clinicians had variable opinions on MOLST form completion for patients who are full code (as opposed to DNR/DNI). Some completed the form, and some did not. Contrary to this, there was a uniform consensus among palliative care clinicians that they do not complete MOLST forms for patients who are full code.

From the palliative care perspective, reasons for not completing a MOLST form for patients who wish to be resuscitated include the difficulty of changing code status later and fear of the “pink form” in emergency situations where other healthcare professionals in the emergency department may assume the order is DNR/DNI. This would, in turn, withhold care from patients who want to receive all medical interventions needed to live longer. Similarly, it would be more emotionally challenging for families to reverse a documented full code for their loved one, as opposed to making the decision to limit care from the beginning of a hospital admission. Palliative care clinicians explained that it is often easier for families to withhold resuscitation if there is no documented full code status, whereas it is harder to reverse a documented full code.

Oncology clinician perceived barriers to ACP.

Oncology clinicians in this study described challenges to ACP: 1) Clinician perceived patient discomfort, 2) Clinician perceived patient lack of knowledge, and 3) Lack of time. Oncology clinicians felt that ACP discussions added to patient distress. Based on their prior experience, oncology clinicians reported that patients often equate ACP with dying, giving up hope, and/or being denied treatment. Clinicians felt that these perceptions may threaten the clinician-patient relationship and rapport.

Some oncology clinicians attributed perceived patient discomfort to perceived lack of patient knowledge around the severity and nature of hematologic malignancies. Patients are generally more familiar with the terminology and staging of solid tumor cancers, requiring clinicians to spend more time educating patients on their diagnosis, as opposed to discussing their wishes and values. Palliative care physicians agreed and added to this challenge, reporting that patients often do not fully understand the purpose of ACP. Therefore, both oncology and palliative care clinicians felt that a major barrier to ACP is patient lack of knowledge around their disease, which is compounded by their lack of familiarity with ACP discussions.

Oncology clinicians identified lack of time as a major barrier to ACP discussions. Clinicians felt that outpatient clinics are the preferred setting for ACP discussions, but outpatient clinics schedules allot limit time with each patient in order to promote efficiency. Therefore, in order for clinicians to see more patients, they need to spend less time with each individual. Oncology clinicians believed that clinic visits do not provide adequate time for ACP discussions. In addition, ACP discussions should evolve over time and be revisited as patient’s values change with the inevitable progression of their diagnosis.

Patients felt that they are balancing fear and hope when navigating their AML or MDS diagnosis.

Patients felt that ACP discussions were emotional conversations. Many patients in this study felt that they were balancing fear and hope. Fear of knowing too much about their diagnosis and the severity of AML and MDS held patients back from exploring their medical preferences and values with their clinicians. Some patients attributed their fear to older age, describing that older generations of people are less likely to participate in conversations that explore vulnerabilities.

Patients were fearful about how their AML or MDS diagnosis impacted their loved ones. Patients prioritized their family in treatment decisions and ACP. Their hope was to minimize guilt and burden on their loved ones who may have to eventually make medical decisions on their behalf. Many patients suggested having early ACP discussions to ensure their loved ones would have a more peaceful experience if their disease progressed and/or declined.

Despite this, patients reported that they would like the opportunity to share what is meaningful to them with their clinicians. Patients were hopeful that the more connected they feel to their care team, the more positive they would feel during treatment, and therefore would have better outcomes. While patients were fearful to know about the severity of their illness, patients in this study reported that an ACP discussion would be helpful for them to better understand their disease. Patients felt that ACP discussions could serve as an opportunity for patient empowerment. These conversations should be patient-driven and could provide patients with the knowledge needed to navigate their AML or MDS diagnosis with more confidence.

Patients in this study were able to successfully balance fear and hope by upholding and endorsing a positive attitude regarding their care. The feeling of being supported by their care team directly impacted their experience during treatment. Specifically, optimism was essential to successfully navigate treatment and the side effects/unknowns that come with treatment.

Discussion

In this single-center qualitative study, we sought to better understand ACP from the perspective of clinicians, patients, and caregivers. We identified four major themes: 1) The language of ACP and MOLST does not resonate with patients, 2) There is no uniform consensus on when ACP is currently happening, 3) Oncology clinician perceived barriers to ACP, and 4) Patients felt that they are balancing fear and hope when navigating their AML or MDS diagnosis.

There is no uniform consensus on when ACP is happening for patients with AML and MDS. Oncology clinicians reported that the timing of ACP is individualized and variable for each patient, with earlier discussion for those with worse prognoses. Previous research examining ACP in patients with metastatic breast cancer found that the timing of ACP is not uniform, and ACP is generally addressed late in the disease course when few anticancer therapy options remain (32). Our previous work in patients with AML and MDS revealed that 51.7% of patients with AML or MDS completed a MOLST form ≤30 days prior to death, and 14.2% never completed a MOLST form (16). Patients who completed MOLST forms early were less likely to receive life-sustaining treatments (AOR 0.39, p <0.01) and be admitted to the ICU (AOR 0.24, p<0.01) (16). Similarly, a retrospective study of 383 patients with hematologic malignancies treated at an academic cancer center found that 66.8% of goals of care discussions occurred ≤30 days prior to death (19). Patients who had a goals of care discussion early (>30 days before death) were less likely to be admitting to the ICU (OR 0.37; 95% CI, 0.17–0.81) in the last month of their life (19). ACP discussions, therefore, often happen late in the disease course, despite previous research showing that earlier ACP can improve EOL experience (16, 19). This suggests the need for earlier ACP discussions for most older patients with AML and MDS, regardless of their prognosis, given the high risk of decompensation and death (27). Furthermore, participants in this study stressed the need for early ACP conversations, emphasizing that care preferences are ideally elicited from patients, themselves, rather than loved ones who may experience guilt on their behalf. Therefore, while it is important that ACP is individualized to each patient, systematic ways to incorporate ACP early in the disease course are needed (16, 19).

Challenges to ACP discussions were raised by oncology clinicians. Oncology clinicians perceived that ACP conversations make patients feel uncomfortable and take away hope. Previous research exploring physician perspectives on ACP found that the most common physician-perceived barrier to ACP is fear of the reaction from the patient or their family (33). Despite these reported findings, a study of 672 patients with advanced cancer found that hope is not decreased after patients have ACP discussions with their providers. Rather, hope was increased for patients after ACP discussions and advance directive completion (34). Patient satisfaction with ACP, however, was not analyzed in this study (34). A prior study including 200 patients with a cancer diagnosis found that only 24% of patients reported having an ACP conversation with their physician, while 82.5% of patients reported wishing they had an ACP discussion with their physician (35). Of the 24% of participants who had ACP discussions, the majority (89.4%) reported that the discussion was physician-initiated, seemed brief and rushed, and were dissatisfied (35). Therefore, while oncology clinicians in this study identified patient discomfort and hope as a challenge to ACP, previous studies have shown that ACP is associated with increased hope for patients with cancer. These findings also highlight the desirability and positive impact of ACP discussions for patients with cancer.

Patient knowledge of their diagnosis and of the concept of ACP was identified as a challenge by clinicians. A previous study similarly found that a major barrier to EOL care from the oncologist perspective was unrealistic patient expectations (36). Current literature suggests that limited patient knowledge is a common experience for patients with AML (37). For example, a qualitative study of 32 patients with AML found that patients were uncertain about their prognosis, what treatments were available to them, and what to expect from their treatment (38). Older patients also suggest that ACP is something their family members will take care of (39). Therefore, interventions to promote patient understanding of their diagnosis may help facilitate ACP for older adults with AML and MDS (40). Specifically, patient education may improve diagnostic and prognostic understanding, allowing them to better understand and participate in ACP conversations (40).

Telehealth is a possible method to address limited resources such as time and palliative care services, while simultaneously promoting earlier integration of ACP into care for patients with AML and MDS. A previous study found that telehealth increased access to palliative care for homebound patients with Parkinson’s disease (43) and can address healthcare disparities by expanding access in rural communities (44). Telehealth can also help mitigate barriers such as long travel distances and promote inclusivity of distant family members.

Clinician usage and patient understanding of MOLST forms were explored in depth in this study. Oncology physicians reported variable MOLST form use and completion, especially for patients who desire full code. Palliative care physicians consistently reported that they do not complete MOLST forms for patients who desire full code status. While patients were not familiar with the term “MOLST form,” many of them recognized the concepts and therefore felt that their code status was an important part of their care to lift burden from their loved ones. Our study found that the role of the MOLST form is inconsistent and unclear at the physician and patient level. Nevertheless, previous work has shown that patients with hematologic malignancies who complete advance directives late are more likely to die during hospitalization (45), but those who complete them early are less likely to be admitted to the ICU (16). MOLST form can serve as a gateway to engage in ACP discussions with patients, and its routine implementation may promote ongoing goals of care discussions between patients and clinician.

In recent years, controversy has surrounded the impact of ACP and the distinction between ACP and serious illness communication (46, 47). The ability of ACP to lead goal-concordant care has been questioned, and thus there is a drive to shift care toward serious illness communication since the latter focuses on the patient’s current illness and occurs and can provide immediate benefit to patients and families (46, 47). ACP is increasingly used in the context of stable chronic illness to understand preferences for future treatment, while serious illness communication typically takes place in the context of acute or decompensated illness to understand preferences for in-the-moment treatment (48). In our study, we used the terminology “advance care planning,” since that is more familiar to clinicians, but in retrospect, “serious illness communication” might be more appropriate in the context of an AML or MDS diagnosis. The latter may also be more understandable by patients. Nonetheless, both ACP and serious illness communication address patient values, may use advance directive forms, and may include surrogate decision makers (48).

The major strength of our study is the inclusion of perspectives from several stakeholders including older patients, and their caregivers. We also included both community based and academic hematologic oncology clinicians (physicians, APPs, nurses), and palliative care clinicians (physicians and APPs). This study delved deeply into the understanding of the MOLST form and its role in patient care, which was not done in previous work. Our study also has several limitations. Participants were recruited from a single center (WCI) and its affiliated sites, although some patients were co-managed with community oncologists not associated with Wilmot Cancer Institute. Second, we did not quantify the number of patients with AML or MDS seen by clinicians, but the majority of clinicians who participated in this study treated >1 patients with AML or MDS in the last year. Third, our study sample was primarily white and non-Hispanic. Previous studies have shown that patients from minority backgrounds have different perspectives on ACP (49). Forth, caregiver enrollment was based on availability and therefore we did not reach thematic saturation for this group. However, we included caregivers in this study because often participate directly in the care and ACP for older patients with AML and MDS.

In conclusion, our study explored current perceptions of ACP from patient and clinician perspectives. Themes that emerged from this study underline the need to improve ACP for older patients with AML and MDS. Patients with AML and MDS are more likely to receive aggressive care at the EOL than patients with solid tumors, and older patients, compared to younger patients, have worse survival rates and higher risks of decompensation. Therefore, early integration of ACP will allow these patients to better prepare for EOL and ensure that their care reflects personal goals and values.