Low Hospice Utilization in New York State: Framework for Compiling and Ranking Barriers

Abstract

Background:

The hospice benefit can improve end-of-life outcomes, but is underutilized, particularly in low enrollment states such as New York. Little is known about this underutilization.

Objective:

The first part of a mixed-methods study aimed to compile and rank barriers to hospice utilization and identify differences between New York and the rest of the United States.

Setting/Subjects and Design:

Clinicians, administrators, and hospice employees participated in six sessions (6–12 per session) across New York State, USA. During each session, a methodology known as nominal group technique was used to elicit barriers to hospice, identify those specific to New York, and suggest interventions to improve access. The analysis involved first categorizing and ranking barriers, and then conducting a thematic analysis of session transcripts to examine barriers specific to New York and proposed interventions to improve utilization.

Results:

Fifty-seven participants ranked 54 barriers, which were grouped into nine categories. These reflected concerns about clinician knowledge and attitudes or beliefs; patient and family knowledge, attitudes or beliefs, and resources; and both structural elements and practices of hospices, nursing homes, palliative care services, and other entities in the health care system. Thirteen barriers from eight categories were ranked among the top five by ≥10% of participants; only 10 of the 54 were judged to be specific to New York. Thematic analysis highlighted 14 barriers important in New York and suggested 11 interventions to improve hospice access.

Conclusions:

A categorization and ranking of barriers may guide future interventions to improve low hospice utilization. Novel studies with heterogeneous stakeholders are needed.

Introduction

In the United States, hospice is a government-supported benefit for eligible patients with advanced illness and limited prognosis. Eligible individuals who elect the benefit receive services consistent with specialist palliative care, provided in the home or nursing home without cost to the patient or family. The benefit includes case management by an interdisciplinary team; access to a home health aide and volunteers; inpatient care when needed for acute conditions; bereavement services; and all tests, treatments, supplies, and equipment related to the terminal illness.

Although hospice enrollment has been associated with positive outcomes,^1,2 only half of Medicare decedents nationally receive hospice services before death, the median length of stay is less than three weeks, and about one in four patients enroll for less than a week.³ In some states, hospices serve fewer than one-third of Medicare decedents.³ These figures suggest the existence of barriers that delay or prevent hospice enrollment by eligible patients. Some of these barriers may be national, but have variable impact across states^4–7 and others may be specific to a state or region. To date, there have been no systematic studies of these barriers.

New York State is ranked 50th among all states and the District of Columbia in the proportion of Medicare decedents who receive hospice before death.³ The reasons for this low utilization are unknown and the lack of data has impeded initiatives to promote timely and appropriate hospice enrollment. We designed a mixed-methods study to examine the barriers to low utilization of hospice in New York State and identify those that may distinguish this state from others. The first part, described herein, obtained information from professional stakeholders, and the second part⁸ evaluated national data sets to quantitatively compare state-level barriers to the rest of the country.

Materials and Methods

The study was approved by the Ethical and Independent Institutional Review Board. Participants provided verbal consent for participation and audio recording.

The study included six stakeholder group sessions in different regions of New York State, each of which was conducted using the nominal group technique (NGT) approach.^9–13 NGT is a structured methodology involving the conduct of focus groups in which responses to specific questions are elicited and stakeholders rank these responses in terms of perceived importance.¹⁴ The method yields both the pooled rankings of responses and a secondary thematic analysis of the discussions that are interspersed with the ranking activities.

Sessions

We convened sessions in New York City (Manhattan), two locations in the New York City suburban area (one in Nassau County and the other in Rockland County), and three locations in Western and Central New York State (near Albany, Syracuse, and Buffalo, respectively). Sessions were held between August 2019 and October 2019. The leaders of six hospice organizations serving these respective areas (Visiting Nurse Service of New York, Hospice Care Network, United Hospice of Rockland. The Community Hospice at Albany, Hospice of Central New York, and Hospice of Buffalo) were asked to identify individuals with a known interest in the appropriate use of hospice from among local health care administrators, clinicians, and hospice employees. Each hospice agency provided at least one participant, helped confirm the participation of others, and provided space for the session. Each 90-minute session was attended by 6–12 participants and 2–3 investigators (C.B., R.P., and L.D.).

The investigators developed an initial list of 24 potential barriers to hospice based on clinical experience and expert opinions.^4–7 It included both national (e.g., “Clinicians have concerns about the quality of hospice care”) and New York-specific (e.g., “Clinicians are reluctant to refer to hospice because there is greater access to academic medical centers in New York and greater access to specialist care”) factors. This initial list was expanded and modified through the NGT approach (Table 1).

Table 1.

Nominal Group Technique Approach and Analysis

1. Before each session, participants planning to attend were sent a list of barriers and asked to rank them in terms of perceived importance as determinants of low hospice utilization in NYS and consider other factors that could be added to the list at the time of the session.

2. The list of barriers available at the start of each session was written on a whiteboard visible to those participating in the session. At the start, each participant was asked to report the three most important factors from the list and these rankings were then recorded on the whiteboard.

3. In a second round, each participant was asked to describe additional barriers to be added to the list and explain their rationale. These new barriers were recorded on the whiteboard after the moderator clarified the wording and eliminated redundancies with other factors.

4. In a third round, each participant was asked to review the expanded list of factors and rerank the three most important in terms of its potential as a determinant of low hospice utilization in NYS. These rankings were recorded.

5. After these rounds, the group was asked to discuss the specific reasons, if any, that some or all the factors listed on the board would be specific to NYS and whether any could be targeted for interventions that were likely to improve hospice access in the state. Approximately 15 minutes were provided for this discussion.

6. After all the sessions were completed, the list of barriers had increased from 24 at the start of the first session (developed by the investigators) to 54. This list of 54 barriers was sent to all participants with a request to select the five factors that were the most important determinants of low hospice utilization in NYS.

7. To create a framework describing the barriers to hospice utilization in New York, the barriers listed and ranked at the sessions were edited for clarity, grouped into categories, and designated by the rankings provided by all participants.

8. Thematic analysis was applied to the audio recordings of the sessions to obtain additional information about the barriers that may be specific to NYS and about potential interventions that could improve hospice access.

NYS, New York State.

After the last session was completed, the expanded list of 54 barriers was emailed to all 57 participants with instructions to select the five that were the most important determinants of low utilization of hospice in New York State. These final rankings, representing a review of all proposed factors by all participants, were used in the analysis.

Analysis

Consistent with the standard use of NGT,^9–14 information from the sessions was analyzed in two ways. First, the barriers to hospice utilization were edited for clarity, categorized, and organized by the rankings of the session participants. Second, thematic analysis was applied to the transcripts of the audio recordings.

A meeting of three investigators (L.D., R.P., and C.B.) was held to discuss the NGT analysis. Each investigator then independently reviewed the list of barriers, editing items for clarity and placing each barrier into a defined category. Three meetings were subsequently held to develop consensus through discussion. After these meetings, the investigators independently designated each barrier as likely or unlikely to be specific to New York State based on clinical experience, and disagreements were again resolved through group discussion. An uninvolved investigator (K.R.) then separately reviewed the wording and categorization of the barriers¹⁰; this produced two wording changes and a shift of four barriers from one category to another. As a last step in this part of the NGT analysis, each barrier in this framework was labeled by the percentage of the 57 participants who ranked the barrier as one of top five in importance.^9–14

The second part of the NGT analysis involved thematic analysis of the session transcripts.¹⁴ Data trustworthiness was promoted through adherence to consolidated criteria for reporting qualitative research guidelines for the conduct of qualitative research¹⁵ and the use of independent textual coding by two investigators (S.D. and K.R.), successive rounds for consensus, and a separate review by an additional investigator (L.D.).

The six session transcripts were inputted into Dedoose, a web-based, qualitative analysis program, and a framework was created for the systematic coding of participant statements based on the questions posed to the participants. Four parent codes were selected, one for content focused on differences between New York and other states, and three for content focused on interventions to improve hospice access (educational, health care system, and regulatory, respectively). Each coder independently read each session transcript, identified text pertinent to one of the four codes, and proposed child codes (themes) that reflected varied aspects of the parent code.

The coders met multiple times to reach consensus and to iteratively modify the codebook as new child codes were generated. Differences between the investigators were resolved through discussion. This process yielded 25 child codes distributed among the four parent codes. Reports of codes and associated transcript excerpts were generated through Dedoose and exported for review.

Results

Participants

Demographic items assessed participants' age, sex, educational level, occupation, employer, and number of years working in health care and hospice. The average age of the 57 participants was 58.6 years (standard deviation [SD] 9.8, range 35–80) and 44 (77.2%) were women. Thirty (52.6%) worked in health care administration (employed by a hospital, a nursing home, or a home care agency), 11 (19.3%) were employed by hospice agencies in administration or clinical services, and 9 (15.8%) were clinicians (including physicians, nurses, and chaplains). Four (7.0%) were retired after careers in health care (a hospice administrator, two nurses, and a physician) and 2 (3.5%) were spiritual care advisors (a retired nun and a local pastor).

Among those employed, the average number of years working in health care was 29.2 (SD 10.9, range 4–50); hospice employees had worked for an average of 13.0 years (SD = 9.7, range 2–40). A total of 25 (43.9%) participants had doctoral-level degrees (24 physicians and 1 nonphysician), 21 (36.8%) had master's degrees, 7 (12.3%) had nursing degrees, and 4 (7.0%) had bachelor's degrees (Table 2).

Table 2.

Participant Demographics (n = 57)

Characteristic
Age, mean (SD)	58.6 (9.8)
Women, n (%)	44 (77.2)
Educational level, n (%)
Bachelor's degree/bachelor of science in nursing	11 (19.3)
Master's degree	21 (36.8)
Doctor of medicine/doctor of osteopathic medicine	24 (42.1)
Doctoral degree	1 (1.8)
Professional role, n (%)
Administrative^a	41 (71.9)
Clinical^b	10 (17.5)
Other^c	6 (10.5)
Years working in the health care system, mean (SD)^d	29.2 (10.9)
Years working for a NYS hospice agency, mean (SD)^e	13 (9.7)

Includes director, senior level officer, vice president/senior vice president, manager, administrator, and owner.

Includes palliative care coordinator, physician, hospitalist, social worker, and pastor.

Includes professor and retired administrators and clinicians.

n = 54 participants.

n = 29 participants.

SD, standard deviation.

Categories, factors, and rankings

The 54 barriers to hospice utilization were grouped into nine defined categories (Table 3): #1: clinician knowledge-based reluctance factors; #2: patient/family knowledge-based reluctance factors; #3: clinician attitude/belief reluctance factors; #4: patient/family attitude/belief reluctance factors; #5: nursing home factors; #6: hospice-associated factors; #7: patient/family resource-based reluctance factors; #8: palliative care service factors; and #9: health care system factors.

Table 3.

Categories of Barriers to Hospice Utilization

Category	Definition
#1 Clinician knowledge-based reluctance factors	Clinician reluctance to refer due to inadequate knowledge or skills in palliative care and the utilization of hospice.
#2 Patient/family knowledge-based reluctance factors	Patient/family reluctance to enroll due to limited knowledge of hospice or the types of outcomes that can occur if disease-modifying therapies with questionable benefits are employed near the end of life.
#3 Clinician attitude/belief reluctance factors	Clinician reluctance to refer due to attitudes, beliefs, religious concerns, values, or biases.
#4 Patient/family attitude/belief reluctance factors	Patient/family reluctance to enroll due to attitudes, beliefs, religious concerns, or values.
#5 Nursing home factors	Nursing home impediments to hospice referral.
#6 Hospice-associated factors	Hospice performance or operational factors that increase the reluctance of clinicians to refer, the acceptance of referral by patients/families, or effective/timely admission after referral.
#7 Patient/family resource-based reluctance factors	Patient/family reluctance to enroll due to insufficient resources, the expectation of lost income, or confusion about decision making.
#8 Palliative care service factors	Factors associated with palliative care services that negatively affect hospice referral.
#9 Health care system factors	Factors in the structure or performance of the health care delivery system that impede hospice referral or enrollment.

Forty-seven participants provided the final rankings (82.5% response rate). Thirteen barriers distributed across eight categories were ranked among the top five by ≥10% (six or more) of the participants (Table 4). The highest ranked—“Clinicians lack skills in ‘serious illness conversations,’ are uncomfortable about having these conversations, or are uncertain about who ‘owns’ responsibility for this” (category #1)—was selected by 31 participants (54.4%) and was not specific to New York State. The remaining 12 barriers ranked among the top five by ≥10% included five that were selected by 12 (21.1%) to 16 (28.1%) of the participants. One was specific to New York State—“Patients/families believe that New York health care is exceptional and insist on disease-modifying therapies” (#4)—and four related to all states.

Table 4.

Categories, Factors, and Rankings

Factors ranked in the top five by ≥10% of participants	n (%) with a top-five rating
Category #1: Clinician knowledge-based reluctance factors
Clinicians lack skills in “serious illness conversations,” are uncomfortable about having these conversations, or are uncertain about who “owns” responsibility for this	31 (54.4)
Clinicians lack knowledge about hospice, including eligibility, “related” and “nonrelated” diagnoses, benefits, and services	12 (21.1)
Clinicians believe that hospice is intended for only those dying in the next few weeks	9 (15.8)
Category #2: Patient/family knowledge-based reluctance factors
Patients/families are not realistic/informed about the burdens associated with advanced illness or disease-modifying therapies	16 (28.1)
Patients/families reject hospice because it is equated with dying, or they believe that they will die immediately after referral	14 (24.6)
Category #3: Clinician attitude/belief reluctance factors
Clinicians perceive that patients/families are unwilling/not ready to discuss end-of-life care or discuss/accept hospice	13 (22.8)
Clinicians who are specialists, including those in the ICU/critical care setting, are particularly reluctant to refer because of their focus on disease management or their strong bias toward aggressive treatments	10 (17.5)
Category #4: Patient/family attitude/belief reluctance factors
Patients/families believe that New York health care is exceptional and insist on disease-modifying therapies^a	13 (22.8)
Category #5: Nursing home factors
Patients/families accept rehabilitation services in nursing homes because they believe these will be effective or are free-of-charge, or they believe the nursing home can provide palliative or “comfort” care^a	8 (14.0)
Category #6: Hospice-associated factors
Clinicians lack the time to discuss hospice with patients/families, perceive that hospice discussions and referrals are too complicated/time-consuming	7 (12.3)
Category #7: Patient/family resource-based reluctance factors
Patients/families are reluctant to enroll in hospice if it means losing access to other trusted clinicians	6 (10.5)
Families cannot provide the amount of caregiving required when a family member is enrolled in hospice	6 (10.5)
Category #8: Palliative care service factors
Clinicians do not have access to palliative care specialists in the ambulatory setting, where most care occurs	6 (10.5)

Factors ranked in the top five by <10% of participants	n (%) with a top-five rating
Category #1: Clinician knowledge-based reluctance factors
Clinicians are unable to prognosticate accurately	4 (7.0)
Clinicians receive no education about hospice in medical schools, have had no exposure to hospice, or lack educational opportunities	4 (7.0)
Clinicians rely on colleagues who are not medically trained to identify hospice patients, but they do not recognize advanced illness	2 (3.5)
Clinicians believe that they will lose income by referring to hospice	1 (1.8)
Clinicians in hospitals are reluctant to refer to hospice because they perceive that hospice will not enroll the imminently dying inpatient	0 (0)
Category #2: Patient/family knowledge-based reluctance factors
Patients/families lack information or experience with hospice or do not accept hospice because of their own past negative experiences	3 (5.3)
Patients/families are confused about the differences between palliative care and hospice	2 (3.5)
Patients/families have concerns about hospice quality	2 (3.5)
Patients/families believe that a hospice referral means that their provider is abandoning them	1 (1.8)
Patients/families are reluctant to enroll in hospice because of low health literacy	1 (1.8)
Patients/families see no reason for hospice because end-of-life care is provided well enough by family and other services	0 (0)
Category #3: Clinician attitude/belief reluctance factors
Clinicians are reluctant to refer to hospice because there is greater access to academic medical centers in New York and greater access to specialist care^a	5 (8.8)
Clinicians believe that a hospice referral would take away a patient's hope	3 (5.3)
Clinicians' own personal values and beliefs are incompatible with hospice	3 (5.3)
Clinicians perceive pressure from their colleagues to offer patients aggressive treatments or believe that colleagues will be reluctant to refer if they offer patients hospice	0 (0)
Category #4: Patient/family attitude/belief reluctance factors
Patients/families view hospice care as incompatible with their values, culture, language, or religion, or hospice staff as lacking in diversity	4 (7.0)
Patients/families in NYS are generally distrustful of strangers (caregivers) in their homes^a	1 (1.8)
Patients/families discourage their physicians from discussing end-of-life care or hospice	0 (0)
Category #5: Nursing home factors
Clinicians are reluctant to offer hospice to patients who are transferring to skilled beds in nursing homes because hospice will not be provided in these settings and facilities are financially incentivized to keep patients in skilled care^a	5 (8.8)
Clinicians in nursing homes do not perceive hospice as offering collaborative practice	1 (1.8)
Category #6: Hospice-associated factors
Clinicians are reluctant to refer to hospice because hospices often cannot provide necessary skilled care	5 (8.8)
Hospices do not market themselves adequately because of the Certificate of Need requirements^a	4 (7.0)
Clinicians are reluctant to refer because they would be expected to be the hospice Attending-of-record	1 (1.8)
Hospice agencies, particularly nonprofit agencies, lack the financial resources to maintain staff	1 (1.8)
Clinicians have concerns about the quality of hospice	1 (1.8)
Hospices do not market themselves adequately because they lack the financial resources	0 (0)
Clinicians in primary care are reluctant to refer to hospice because of poor communication with hospice providers, or the belief that continuing involvement with patients will be prevented	0 (0)
Category #7: Patient/family resource-based reluctance factors
Patients/families often need a place to go for end-of-life care, and the lack of hospice residences in NYS is an impediment to hospice enrollment^a	5 (8.8)
Patients/families have concerns about losing services if they enroll in hospice	3 (5.3)
Patients/families perceive that hospice enrollment may lead to reduced access to a revenue stream that helps the family^a	1 (1.8)
Patients/families often have complex family structures, which cause confusion about who the primary decision maker is	0 (0)
Category #8: Palliative care service factors
Clinician shortages (physician or nurses with specialized training in hospice or palliative care) impede the referral of patients to hospice, or hospices' ability to accept patients	3 (5.3)
Clinicians have learned to refer to palliative care consultation services and assume that hospice referral will follow, but palliative care services are unwilling to refer	2 (3.5)
Category #9: Health care system factors
Patients in Assisted Living facilities are not referred to hospice because of specific regulations^a	5 (8.8)
Clinicians, managers, and leadership working in health care facilities, and payers, do not view hospice as an integrated or normal part of the health care system, which is focused on “cure versus care” and financially incentivizes disease-modifying therapies	4 (7.0)
Clinicians in hospitals are discouraged from offering hospice because they perceive that the hospital would be negatively impacted by hospice patients in hospital beds	1 (1.8)
Clinicians' ability to offer hospice is impeded by the lack of contracts between hospices and hospitals or nursing homes, or perceived conflicts	1 (1.8)
Patients enrolled in Medicare Advantage programs commonly receive case management that does not promote hospice referral	1 (1.8)
Patients/families, especially with low socioeconomic status, lack access to a medical home, and/or often use emergency departments for primary care, which limits hospice referral	1 (1.8)
Patients who are eligible for or are enrolled in Medicaid Long-term Care Programs, are not referred to hospice because of specific regulations^a	1 (1.8)
Clinicians have concerns about medicolegal risk if they do not emphasize aggressive disease-modifying therapy or lack knowledge about the regulations covering hospice care	0 (0)

Judged to be specific to NYS.

ICU, intensive care unit.

The seven other factors ranked in the top five were selected by 6 (10.5%) to 10 (17.5%) of the participants. Only one was specific to New York State—“Patients/families accept rehabilitation services in nursing homes because they believe these will be effective or are free-of-charge, or they believe the nursing home can provide palliative or ‘comfort’ care” (#5). The others included two clinician-related barriers, two related to family resources, one related to hospice, and one related to palliative care services.

There were 41 barriers that were ranked in the top five by fewer than 10% (five or fewer) of participants (Table 4). Only eight were specific to New York State, including two in category #7: patient/family resource-based reluctance factors (“Patients/families often need a place to go for end-of-life care, and the lack of hospice residences in NYS is an impediment to hospice enrollment” and “Patients/families perceive that hospice enrollment may lead to reduced access to a revenue stream that helps the family”) and two in #9: health care system factors (“Patients in Assisted Living facilities are not referred to hospice because of specific regulations” and “Patients who are eligible for or are enrolled in Medicaid Long-term Care Programs, are not referred to hospice because of specific regulations”).

The other New York State-specific barriers included “Clinicians are reluctant to refer to hospice because there is greater access to academic medical centers in New York and greater access to specialist care” (#3), “Patients/families in NYS are generally distrustful of strangers (caregivers) in their homes” (#4), “Clinicians are reluctant to offer hospice to patients who are transferring to skilled beds in nursing homes because hospice will not be provided in these settings and facilities are financially incentivized to keep patients in skilled care” (#5), and “Hospices do not market themselves adequately because of the Certificate of Need requirements” (#6).

The remaining 33 factors were not specific to New York State (Table 4). Five participants ranked “Clinicians are reluctant to refer to hospice because hospices often cannot provide necessary skilled care” (#6) in the top five of importance and several participants identified lack of knowledge about hospice as a concern for both clinicians (e.g., “Clinicians receive no education about hospice in medical schools, have had no exposure to hospice, or lack educational opportunities” and “Clinicians are unable to prognosticate accurately”) and patients/families (e.g., “Patients/families lack information or experience with hospice or do not accept hospice because of their own past negative experiences” and “Patients/families are confused about the differences between palliative care and hospice”).

Similarly, several participants highlighted the importance of problematic attitudes/beliefs among clinicians (e.g., “Clinicians believe that a hospice referral would take away a patient's hope” and “Clinicians' own personal values and beliefs are incompatible with hospice”) and patients/families (e.g., “Patients/families view hospice care as incompatible with their values, culture, language, or religion, or hospice staff as lacking in diversity”).

Several other barriers were noteworthy. Palliative care services were noted to have both positive and negative influences on hospice (#8): “Clinician shortages (physicians or nurses with specialized training in hospice or palliative care) impede the referral of patients to hospice, or hospices' ability to accept patients” and “Clinicians have learned to refer to palliative care consultation services and assume that hospice referral will follow, but palliative care services are unwilling to refer.”

Problems in the health care system (#9) were highlighted: “Clinicians, managers, and leadership working in health care facilities, and payers, do not view hospice as an integrated or normal part of the health care system, which is focused on ‘cure versus care’ and financially incentivizes disease-modifying therapies”; “Clinicians in hospitals are discouraged from offering hospice because they perceive that the hospital would be negatively impacted by hospice patients in hospital beds,” and “Patients enrolled in Medicare Advantage programs commonly receive case management that does not promote hospice referral”).

Secondary thematic analysis

A total of 133 participant excerpts were coded. The parent code depicting the difference between New York State and other states yielded 14 child codes (themes) and the three parent codes describing interventions included, respectively, five educational themes, four focused on the health care system, and two describing regulatory change (Table 5).

Table 5.

Themes Identified in Session Transcripts

Parent code	Theme
New York-specific reasons for low utilization	Prevalence of academic medical centers Preference for disease-modifying therapies at the end of life Large elderly population Fragmentation of social networks Influence of structural and financial factors Patient and family distrust of strangers Fragmentation of health care and disruptions in provider relationships High utilization of emergency care and urgent care services Multiculturally diverse population Shortages in hospice workforce Medicolegal concerns Limited availability of hospice residences Attitudes about death and dying Beliefs about training and education
Education-based interventions that could favorably affect hospice utilization	Improve professional training and education on foundational elements of hospice and palliative care at the graduate and post-graduate levels Increase the availability of initiatives for training mid-career clinicians in hospice and palliative care Improve clinician education on the hospice benefit Improve patient, family, and public education on hospice services and benefits Improve marketing of hospice
System-level interventions that could favorably affect hospice utilization	Introduce the concept of hospice early in the patient illness trajectory to reduce associations with imminent death Integrate and reframe palliative and hospice care as part of the continuum of care rather than distinct care at the end of life Develop collaborative practice models and multidisciplinary team approaches to improve hospice access Increase the availability of hospice residences
Regulation-related interventions that could favorably affect hospice utilization	Modify regulations and existing payor policies for nursing homes and hospice Modify hospice eligibility criteria

Several differences between New York State and other states were described. Some participants perceived the New York health system as having inconsistent access to primary care providers and few medical homes, and over-reliance on specialist care. One noted that a lack of communication between primary care providers and specialists may diffuse responsibility for hospice-related conversations. Participants also described the relatively large elderly population in New York as relevant, suggesting that this demography resulted in a negative referral bias or reduced willingness to accept referral because older caregivers may struggle to provide support to a patient enrolled in hospice. This perception was reinforced by other participants, who described the salience of social fragmentation affecting families in the urban region of New York and noted that families of older adults may be unavailable for caregiving.

Participants noted that very ill patients in New York often are transferred from a hospital to a nursing home-based rehabilitation unit, rather than to hospice. Multiple reasons were offered: potentially easier for providers; perceived as a way to obtain a nursing home bed if a patient lacks resources for room and board expense; more acceptable to families; and preferred by both hospitals (rapid bed turnover) and nursing homes (higher reimbursement for skilled services) because of financial considerations. Finally, other participants highlighted a lack of hospice residence beds in New York State and shortages of hospice clinicians.

The participants proposed interventions (educational, health system changes, and regulatory changes) to improve hospice access and utilization. Several underscored the need for statewide efforts to enhance medical school, residency and fellowship training in serious illness conversations, as well as training opportunities for mid-career clinicians seeking to re-specialize in palliative care. The importance of funding for continuing education focused on hospice eligibility and referral, and education to address misconceptions about hospice, also was noted. The value of education was also noted in reference to informing the public about hospice services in a manner similar to other types of health care, with messages that address misconceptions, promote awareness, and normalize hospice care. Several participants noted the importance of introducing hospice to patients early after the diagnosis of a life-limiting disease, as part of a destigmatizing educational strategy.

Interventions focused on the health care system underscored the value of integrating the principles of palliative care and hospice into conventional medical practice. A participant noted that this integration would better communicate that palliative care improves quality of life throughout the course of a serious illness and that hospice provides this focus when the disease is advanced. Participants also suggested an increase in the number of hospice residence beds and support for innovations, such as collaborative practice models linking inpatient or primary care settings with multidisciplinary teams engaged in hospice.

Other suggested interventions were focused on regulatory reform. Participants recommended regulations that would encourage enrollment in nursing home hospice and suggested that existing regulations should be modified so that patients with longer life expectancies could receive hospice and concurrent care with disease-modifying therapy.

Discussion

Eligible patients who elect the hospice benefit receive numerous treatments and services implemented through a care plan that addresses multiple sources of distress and burden experienced by patients and families. Although hospice can improve end-of-life outcomes, routine timely referral of patients with advanced illness has not been achieved. Nationally, only half of Medicare decedents receive hospice before death and the median length of stay is ∼25 days.³ There is large state-to-state variation in these figures and some states, such as New York State, have very low utilization.³ As a large state with a very diverse population and a health care system that is both heterogeneous and highly regulated, New York provides an opportunity to examine multiple barriers that may operate to a greater or lesser extent in other states.

Although national barriers to hospice utilization have been suggested,^4–6 there have been no systematic studies of factors that may explain the low utilization of hospice and its variation across states. Studies that identify barriers are needed to guide efforts to improve access. In this first part of a mixed-methods study, we used the NGT approach^9–14 to explore the perceptions of 57 professional stakeholders from different regions of New York State. The aims were to identify and categorize barriers to hospice utilization, rank these barriers in terms of their perceived impact, and explore the perceptions of stakeholders that may illuminate the differences between New York and other states and the types of interventions that could improve hospice access. The second part of the study, reported elsewhere,⁸ analyzes federal claims data to further examine the differences between New York State and other states.

Nine categories of barriers to hospice utilization were identified. They reflect concerns about clinician knowledge and attitudes or beliefs; patient and family knowledge, attitudes or beliefs, and resources; and both structural elements and practices of hospices, nursing homes, palliative care services, and other entities in the health care system. Most of the 54 barriers identified apply to all states. The top ranked barrier underscores the need for improved clinical competency in the conduct of serious illness conversations. Barriers that may be specific to New York State, identified through either the ranking process or the thematic analysis of transcripts, were highly diverse.

Interventions to improve access to hospice are needed and the information gained from this study could be helpful in developing targeted approaches. As suggested by the participants, these approaches may be educational, targeting professionals and the public, or any of a group of changes that could modify the health system in a way that removes disincentives to hospice and integrates this care into routine practice.

The limitations of the information obtained through this approach should be noted. The barriers are observations and require confirmation. Participants were selected by hospice leaders and could be biased toward hospice with favorable opinions about this system of care. Our sample largely included hospice professionals and those familiar with hospice services and benefits. No community members, government workers, or elected leaders participated and the regional distribution of the sessions did not include several major population centers or very rural areas. We did not assess participants' race/ethnicity, cultural background, or religious affiliation and most participants were women.

Future investigations should strive for greater heterogeneity in terms of participant sociodemographics, occupation, professional role, and experiences and attitudes toward hospice and its potential importance in end-of-life care outcomes. Such approaches could involve community-based participatory research methodologies that engage diverse stakeholders to share their personal experiences and elucidate additional barriers, as well as facilitators who could improve utilization and promote more favorable end-of-life outcomes.

Conclusions

Notwithstanding these limitations, this study is the first systematic effort to identify and rank barriers to hospice utilization and examine state-level variation. Combined with the analysis of national data in the second part of this mixed-methods study, the information may assist in future research and the design of initiatives to improve access to hospice.

Footnotes

Acknowledgments

We wish to thank Jack Chen for his assistance with study implementation, the hospice agencies and staff for coordinating and hosting the sessions, and the individuals who participated in the sessions. We gratefully acknowledge the MJHS Foundation for supporting this work.

Data Sharing

The dataset for this study, including the study protocol, analysis plan, and individual deidentified participant data underlying the results reported in this study, will be made accessible four months from initial request to investigators who provide a methodologically sound proposal, as determined by the study's lead investigators. The dataset will be provided after its deidentification, in compliance with relevant regulatory and privacy laws, data safeguards, and requirements for participant consent and anonymization.

Authorship Confirmation Statement

My colleagues and I have agreed to the authorship line and to the submission of the article in this form. All authors listed on the article have read and approved the article.

Funding Information

This work was supported by The MJHS Foundation.

Author Disclosure Statement

No competing financial interests exist.

References

Kleinpell

, Vasilevskis

, Fogg

, and Ely

: Exploring the association of hospice care on patient experience and outcomes of care. BMJ Support Palliat Care, 2019; 9:e13.

Teno

, Clarridge

, Casey

, et al.: Family perspectives on end-of-life care at the last place of care. JAMA, 2004; 291:88–93.

NHPCO Facts and Figures 2020 Edition: National Hospice and Palliative Care Organization: 2020. https://www.nhpco.org/wp-content/uploads/NHPCO-Facts-Figures-2020-edition.pdf (Last accessed March 9, 2022).

Fine

: Hospice underutilization in the U.S.: The misalignment of regulatory policy and clinical reality. J Pain Symptom Manage, 2018; 56:808–815.

Wallace

: Hospice eligibility and election: Does policy prepare us to meet the need?. J Aging Soc Policy, 2015; 27:364–380.

O'Mahony

, McHenry

, Snow

, et al.: A review of barriers to utilization of the Medicare hospice benefits in urban populations and strategies for enhanced access. J Urban Health, 2008; 85:281–290.

Aldridge Carlson

, Barry

, Cherlin

, et al.: Hospices' enrollment policies may contribute to underuse of hospice care in the United States. Health Aff (Millwood), 2012; 31:2690–2698.

Dhingra

, Braveman

, Kassner

, et al.: Low hospice utilization in New York State: Comparisons using national data. J Pain Symptom Manage, 2022; 63:522–529.

Mullen

, Kydd

, Fleming

, McMillan

: A practical guide to the systematic application of nominal group technique. Nurse Res, 2021; 29:14–20.

10.

McMillan

, Kelly

, Sav

, et al.: Using the Nominal Group Technique: How to analyse across multiple groups. Health Serv Outcomes Res Method, 2014; 14:92–108.

11.

Stoffelen

JMT

, Schaafsma

, Kok

, Curfs

LMG

: Views on sex using the nominal group technique to explore sexuality and physical intimacy in individuals with intellectual disabilities. Sex Disabil, 2019; 37:227–244.

12.

de Wolf-Linder

, Dawkins

, Wicks

, et al.: Which outcome domains are important in palliative care and when? An international expert consensus workshop, using the nominal group technique. Palliat Med, 2019; 33:1058–1068.

13.

Dening

, Jones

, Sampson

: Preferences for end-of-life care: A nominal group study of people with dementia and their family carers. Palliat Med, 2013; 27:409–417.

14.

Ven

, Delbecq

: The nominal group as a research instrument for exploratory health studies. Am J Public Health, 1972; 62:337–342.

15.

Tong

, Sainsbury

, Craig

: Consolidated criteria for reporting qualitative research (COREQ): A 32-item checklist for interviews and focus groups. Int J Qual Health Care, 2007; 19:349–357.