Examining Variation in Hospice Visits by Professional Staff in the Last 2 Days of Life

Importance  In the final days of life, symptoms are exacerbated and often families need both education and emotional support from hospice professional staff. The Medicare Hospice Benefit provides a per-diem payment with no requirements on the frequency of patient visits.

Objective  To examine individual characteristics, hospice program, and geographic variation in hospice visits by professional staff during the last 2 days of life.

Design, Setting, and Participants  This retrospective cohort study of 661 557 Medicare hospice beneficiaries included all patients who died in federal fiscal year 2014 (October 1, 2013–September 30, 2014) in hospice services while receiving routine home care in the last 2 days of life. Data analysis was conducted from March 9 to August 25, 2015.

Main Outcomes and Measures  Medicare regulations require that hospice care centers report daily the frequency and duration of patient visits by professional staff, as well as the type of hospice staff who conduct these visits. We examined visit variation by hospice professional staff (physician, nursing staff, and social worker) during the last 2 days of life using a multivariable logistic model adjusting for decedent and hospice program characteristics.

Results  Of the 661 557 patients in the study, 81 478 (12.3%) received no professional staff visits in the last 2 days of life. State variation ranged from 590 of 15 399 patients (3.8%) in Wisconsin to 97 of 492 patients (19.7%) in Alaska, and hospice program interquartile variation ranged from 3.2% (25th percentile) to 19.5% (75th percentile). Black patients were less likely to have any visits than were white patients (7352 of 48 221 [15.2%] vs 70 365 of 585 377 [12.0%]; adjusted odds ratio, 1.30; 95% CI, 1.24-1.37). Those dying in a nursing home were 1.74 times less likely to have any visits than those dying at home (30 880 of 186 649 [16.5%] vs 40 973 of 386 654 [10.6%]; 95% CI, 1.62-1.88). Those dying on a Sunday were 3.35 times less likely to have a visit compared with persons dying on a Tuesday (19 519 of 95 979 [20.3%] vs 6784 of 92 127 [7.4%]; 95% CI, 3.17-3.55).

Conclusions and Relevance  Hospice visits in the last 2 days of life by professional staff vary by race, hospice program, and geographic region of the country. Visits were less likely when the Medicare beneficiary was black, dying on a Sunday, and receiving care in a nursing home.

A key component of the Medicare hospice benefit that is intended to ensure patient quality of life and safety is that hospice programs should provide visits from professional staff to monitor and treat the patient’s symptoms, provide emotional support to the dying patient and family members, and educate caregivers on what to monitor, as well as which treatments to provide for worsening symptoms and when to call the hospice program for additional support. These in-person visits are particularly important in the days immediately preceding death, when symptoms and needs for emotional support are extremely high.^1,2 Medicare-certified hospice programs are paid a per-diem rate to administer services that include staff visits, medications, and durable medical equipment to care for a patient’s terminal illness and related conditions. Similar to a managed care professional, hospice programs do not have any mandated minimum number of required visits for the most common level of hospice care referred to routine home care (RHC). However, a hospice program must deliver the highest possible quality of care for the dying person and support the family members in their role as caregivers with the payments they receive from Medicare.

In the development of the Family Evaluation of Hospice Care survey,^3,4 which asks families about their perceptions of the care the patient received from hospice, focus group participants noted the importance of professional hospice staff in providing education and training to family caregivers. This finding is illustrated by the following quotation from a focus group member who was family to a person dying at home: “I’m glad I (kept her at home) but I think towards the end…giving the medications and upping the medications seemed so fast. All of the sudden [sic] now she is on all this morphine and all this whatever it is. And that kind of bothered me too. It really did. Because it was like, my God, I’m giving her this stuff. Am I giving her too much? I’m not a trained medical person.^5(p743)

Another family member clearly expressed her distress in her caregiving role, stating, “I, being the caregiver, I was scared to death. I mean I’m not a nurse. I don’t have that experience.”^5(p743)

Visits by professional staff who can assess symptoms and make changes to the plans of care as well as work with the patient and the primary caregiver to provide the appropriate palliation and emotional support are important to the quality of care hospices deliver, as noted by the National Quality Forum’s preferred practices on the recognition and management of the actively dying patient.⁶ Previous research has shown that patients have worsening symptoms as death approaches and family caregivers need education about how to identify the signs and symptoms of impending death.^1,2 Since 2010, the Centers for Medicare & Medicaid Services has required hospices to report the frequency, duration, and type of visits they provide to their patients. Given the previous findings from the development of the Family Evaluation of Hospice Care survey, in which families voiced the importance of visits by professional staff in the last days of life, we examined Medicare administrative claims data corresponding to federal fiscal year 2014 to examine patterns in visits by hospice professional staff to the dying patient and family in the final 2 days of life.

We examined all Medicare beneficiaries who died while in hospice during federal fiscal year 2014 (October 1, 2013–September 30, 2014). In addition, the patient had to be receiving only RHC services (the most common level of care that is provided) during their final 2 days of life (ie, the last 2 calendar days, with a partial day counted as 1 day). Routine home care services are paid using a per-diem payment rate for hospice services, in contrast with care for the hospice patient during a crisis, which is provided under the continuous home care level of care or general inpatient level of care. Continuous home care has a mandatory requirement for visits by nursing staff. The general inpatient level of care requires 24-hour nursing services. Our focus was on whether the patient received a visit from the hospice’s nursing staff (RN and LPN), social workers, nurse practitioners, or physicians in the last 2 days of life.

The main focus of this analysis was to examine the variation in visits received by patients as well as characteristics of the professional staff. The Medicare enrollment database and Medicare claims data were used to determine a patient’s age, sex, race and ethnicity, primary diagnosis, residential setting at death (home, nursing home, assisted living facility, or other), the lifetime length of stay in hospice, and the day the patient died. We operationalized length of stay into quartiles, resulting in ranges of 7 days or less, 8 to 23 days, 24 to 91 days, and 92 days or more. The Provider of Services file was used to characterize the age of the hospice program (classified into 4 quartiles: ≤3, 4-8, 9-19, and ≥20 years), whether the hospice program was facility-based, the tax status of the hospice program (for profit vs nonprofit), the state in which the hospice program was located, and the size of the hospice program (classified into 4 quartiles based on the number of Medicare hospice deaths in federal fiscal year 2014: <33, 34-90, 91-201, and >201 deaths). This study was reviewed and approved by the Brown University Institutional Review Board, who waived the requirement for patient consent.

Simple descriptive statistics were used to show the variation in the proportion of hospices providing no visits by professional staff during the last 2 days of life for patients who died while receiving RHC services. For the purpose of describing hospice variation, we used only Centers for Medicare & Medicaid Services certification numbers for hospice programs that had at least 30 discharges during federal fiscal year 2014. For beneficiary-level analyses, we examined all hospices. In addition, we used descriptive statistics to characterize the association of sociodemographic characteristics, hospice primary diagnosis, and organizational characteristics of the hospice program with the lack of professional staff visits in the last 2 days of life. We used a multivariable logistic model to examine patient and hospice organizational characteristics associated with the lack of visits by professional staff. All models used robust SEs that adjusted for clustering of patients within a hospice program. All analyses were performed using Stata, version 13.1 (StataCorp). Data analysis was conducted from March 9 to August 25, 2015.

In federal fiscal year 2014, a total of 661 557 Medicare beneficiaries cared for by 3935 hospice programs died while receiving RHC level of hospice care during the last 2 days of life. Overall, 81 478 of these patients (12.3%) did not receive any visits from professional staff during the last 2 days of life. For those not receiving visits during their last 2 days of life, 32 743 (40.2%) received a visit 3 days before death, 17 257 (21.2%) received a visit 4 days before death, 9185 (11.3%) received a visit 5 days before death, and the remaining 15 867 (19.5%) received a visit 6 days or more before death.

Variation by state in visits by professional staff in the last days of life is shown in Figure 1. The state with the highest proportion of patients not receiving visits from professional staff in the last 2 days of life was Alaska (97 of 492 patients [19.7%]), followed by Washington (2825 of 14 394 [19.6%]), New Jersey (3417 of 17 753 [19.2%]), and Rhode Island (486 of 2528 [19.2%]). Only 590 of 15 399 patients (3.8%) in Wisconsin who died while receiving hospice care did not receive any visits from professional staff in the last 2 days of life. There were 3935 hospices in our sample; but, to examine the variation among the hospice programs, we restricted the analysis to only the 3448 hospice programs with at least 30 discharges. Variation among hospice programs ranged from 281 hospice programs (8.1%) that provided no visits during the last 2 days of life to 21 hospice programs (0.6%) that provided visits to 100% of their patients receiving RHC services during their last 2 days of life. Hospice program interquartile variation ranged from 3.2% (25th percentile) to 19.5% (75th percentile). The characteristics of the 42 128 patients in the 281 hospice programs that provided no visits from professional staff during the last 2 days of life compared with the 1 793 012 patients in the other 3167 hospice programs that did provide such visits differed in a number of ways. Patients from the 281 hospice programs that provided no visits from professional staff during the last 2 days of life were more likely to be in a rural area (14 444 [34.3%] vs 307 217 [17.1%]; P < .001), more likely to be treated by a hospice in existence for 3 years or less (7960 [18.9%] vs 137 898 [7.7%]; P < .001), less likely to be treated by a freestanding hospice (28 459 [67.5%] vs 1 390 669 [77.6%]; P < .001), and more likely to be treated by a for-profit hospice (20 951 [49.7%] vs 755 221 [42.1%]; P < .001).

Of the full sample of 661 557 decedents, the proportion of patients who received no visits from professional staff in the last 2 days of life varied by patient sociodemographic characteristics (Table). Relative to white patients, black patients were less likely to have any visits from professional staff in the last 2 days of life (70 365 of 585 377 white patients [12.0%] vs 7352 of 48 221 black patients [15.2%]; adjusted odds ratio, 1.30; 95% CI, 1.24-1.37). Persons with a primary hospice diagnosis of unspecified debility, dementia, or failure to thrive were less likely to have professional staff visits in the last 2 days of life. In addition, hospice patients residing in nursing homes had a higher proportion of not having any visits from professional staff in the last 2 days of life compared with patients not in a nursing home (30 880 of 186 649 [16.5%] vs 40 973 of 386 654 [10.6%]; adjusted odds ratio, 1.74; 95% CI, 1.62-1.88). Figure 2 presents the proportion of visits by professional hospice staff in the last 2 days of life based on the day of death. More than 1 in 5 persons who died on a Sunday (19 519 of 95 979 [20.3%]) did not have a visit from professional staff in the last 2 days of life. The day of death was a strong predictor of whether a patient received a visit in the last 2 days of life, with those dying on Sunday being 3.35 times more likely to not receive a visit in the last 2 days of life compared with those dying on a Tuesday (95% CI, 3.17-3.55).

Several hospice program characteristics were associated with a lower likelihood of not receiving skilled visits in the last 2 days of life (Table). Smaller hospice programs (defined as those with <33 deaths and those with 34-90 deaths) were significantly less likely to provide visits by professional staff in the last 2 days of life compared with programs with more than 201 deaths. Hospice programs based in a nursing home were less likely to provide these visits compared with freestanding hospice programs (677 of 2240 [30.2%] vs 61 243 of 499 017 [12.3%; adjusted odds ratio, 2.78; 95% CI, 1.04-7.45). Visits did not differ by whether the hospice program was for profit or nonprofit.

Physical symptoms, such as pain, dyspnea, and excess respiratory secretions, worsen in the last days of life, with up to 80% of decedents experiencing 1 or more symptoms.^1,2 The National Quality Forum’s Framework and Preferred Practices for Palliative and Hospice Care Quality emphasized the importance of educating the family on what to expect and on symptom management during the last hours of life.⁶ Focus groups of family members attest to the importance of the hospice staff’s support in providing ongoing assessment, treatment of symptoms, monitoring the plan of care, and education about caring for the imminently dying patient.⁵ Professional staff, such as physicians, nurses, and social workers, may implement changes to the care plan that ensure expert symptom management. For these reasons, we studied visits by professional staff in the last 2 days of life and found that 12.3% of hospice decedents did not receive such visits when undergoing the most commonly provided level of care, RHC.

The Medicare Hospice Benefit does not mandate a set number of visits for RHC services and states only that hospices are responsible for providing high-quality care to dying patients and their families. Our findings that professional staff from 281 hospice programs, which had at least 30 discharges, did not visit any of their patients who received RHC services during their last 2 days of life raises concerns that deserve further research to understand whether a lack of visits by professional staff affects the quality of care for that dying patient and their family. In addition, black patients and frail, older patients residing in nursing homes often did not receive visits from hospice staff in the last 2 days of life, providing evidence of disparities of care found in other areas of health care.

Previous research reported concerns with black Medicare beneficiaries’ access to hospice services, including a higher proportion of patients discharged alive, and significant differences in bereaved family members’ reporting of unmet needs, concerns with quality of care, and the overall rating of the quality of hospice care.^7-9 Our finding that black patients are less likely to have visits by professional staff in the last 2 days of life is a potential explanation for observations that bereaved black family members rate their quality of care lower than do their white counterparts. In addition, bereaved black family members are more likely to have an unmet need for pain and symptom management in hospice and to want more information regarding management of the dying patient’s symptoms.⁹

Larger hospice programs (ie, those with >201 hospice deaths among patients receiving RHC services) are more likely to provide visits in the last 2 days of life. Previous research has found an association of the size of the hospice program with reduced live discharges,^8,10 implementation of National Quality Forum preferred practices,¹¹ and an increased likelihood of providing labor-intensive bereavement services.¹² Potential explanations include that larger hospice programs may have the financial resources to provide more staffing on weekends and/or create dedicated teams for off-hours care.

The Centers for Medicare & Medicaid Services recently finalized initial hospice payment reform policies. The 2 different rates for RHC services will provide increased payment for the first 60 days of hospice care and a service intensity add-on policy will pay up to 4 hours per day in the last week of life for an in-person visit by a social worker and/or registered nurse. The changes may encourage additional services to ensure that the hospice patient and family receive needed services. Hospice professional staff expressed concerns about reimbursement regarding the need for providing essential care, but not meeting the requirements to bill for the continuous home care level of hospice services. Thus, the service intensity add-on policy may alleviate some of these concerns expressed by hospices.

There are important limitations to our study. Aside from the patient’s diagnosis, age, and other sociodemographic information, we did not have knowledge of the severity of the dying patient’s symptoms or the family’s preferences for a visit. It is possible that professional staff from the 281 hospice programs that did not have visits on days the patient was receiving RHC services during the last 2 days of life provided care that anticipated problems and concerns at an earlier date, such that a visit immediately prior to death was not needed. However, we believe that our finding that death on a Sunday is associated with fewer visits compared with death on any other day of the week suggests that this explanation is less plausible. Persons who died on a Sunday were more than 3 times less likely to receive a visit during the last 2 days of life. Biologically, it is implausible that any particular day of the week would be associated with different severity of symptoms or other family needs for education and support in the dying process. It is possible that some of these patients suddenly died, so we would not expect the proportion of those receiving no visits by professional staff to be zero. Finally, we have no information on visits by spiritual counselors, which Medicare does not require hospices to report. We examined visits only by professional staff who can directly assess the patient’s condition and implement changes in the plan of care. While hospice aides, spiritual counselors, and volunteers play an important role in the care of patients and family, state law does not allow them to assess the patient’s condition and implement changes to the hospice plans of care. Therefore, we do not believe they can serve as an adequate substitute for professional staff in terms of their ability to assess the patient’s condition and make changes to the plan of care.

Hospice care began as a social movement led by persons with a vision to improve inadequate treatment of pain and other symptoms at the end of life and to provide emotional support to the family. Medicare policy has focused on the provision of hospice care at “home” whether it is the patient’s home, an assisted living facility, or a nursing home. Regardless of the site of service, a key part of safely caring for patients is that the hospice professional staff provides ongoing assessment, monitoring, and education about the medical management of the imminently dying patient. Our findings raise concerns that black patients and nursing home residents receive less support in the last 2 days of life from hospice professional staff. In addition, persons dying on a Sunday were less likely to receive any visits in the last 2 days of life. However, recent changes to the Medicare hospice payment system may address these problems. Starting in 2016, Medicare will provide a service intensity add-on payment to hospices that provide visits during the last 7 days of life (on days the patient is receiving RHC services). The intent of the payment is to promote visits by professional staff at the end of life and to better align the payments hospices receive with the increased care and service requirements at the end of life for all dying patients and those who care for them.

Accepted for Publication: November 10, 2015.

Corresponding Author: Joan M. Teno, MD, MS, Division of Gerontology and Geriatric Medicine, Department of Medicine, Cambia Palliative Care Center of Excellence, University of Washington, Campus Box 359765, Pat Steele Bldg, 401 Broadway, Ste 5123.11, Seattle, WA 98122 (jteno@uw.edu).

Published Online: February 8, 2016. doi:10.1001/jamainternmed.2015.7479.

Author Contributions: Dr Plotzke had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Teno, Plotzke, Gozalo.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Teno, Plotzke, Christian.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: All authors.

Obtained funding: Teno, Plotzke, Gozalo.

Administrative, technical, or material support: All authors.

Study supervision: Teno, Plotzke.

Conflict of Interest Disclosures: None reported.

Funding/Support: The analyses on which this study is based were performed under contract HHSM-500-2005-00018I, entitled “Hospice Study and Report,” funded by the Centers for Medicare & Medicaid Services, Department of Health and Human Services. This study was also supported by the Robert Wood Johnson Foundation Health Policy Investigator Award (Dr Teno).

Role of the Funder/Sponsor: As required by the contract, the manuscript was reviewed by program staff at the Centers for Medicare & Medicaid Services for the accuracy and protection of privacy of individuals as well as the decision to submit the manuscript for publication. Neither the Centers for Medicare & Medicaid Services not the Robert Wood Johnson Foundation had a role in the design and conduct of the study, management, analysis, or interpretation of the data. The Robert Wood Johnson Foundation had no role in the review, approval, or decision to submit the manuscript for publication.

Disclaimer: The content of this presentation does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does the mention of trade names, commercial products, or organizations imply endorsement by the US Government.