Trends in Palliative Care Utilization in Deceased Veterans With Heart Failure

Background: Specialist-level palliative care in the final days does not allow time to alleviate symptoms and suffering. This analysis examined the change in the time from initial specialty-level palliative care to death among Veterans with heart failure. Methods: This retrospective cohort study examined Veterans with a diagnosis of heart failure (HF) who died between 2011 and 2021. We examined the decedents from each year as a separate cohort. The primary outcome was time from specialty-level palliative care (SPC) encounter to death in the year death occurred. Results: Of the cohort (n = 232,079), 56.5% did not receive SPC. Specialist-level palliative care >90 days before death more than doubled from 10.1% (2011) to 26.2% (2021), and Specialist-level palliative care in the last day of life was cut from 2.5% to 0.9%. Conclusion: For Veterans with HF, specialist-level palliative care moved earlier in the disease course and has a substantial growth opportunity.


Introduction
Approximately 6 million American adults suffer from heart failure (HF), accounting for over 800,000 hospital discharges annually.As HF prevalence nears 8 million by 2030 (3.0% of the US population), annual costs are projected to approach $70B. 1 Atop an *50% 5-year mortality following diagnosis, HF's physical and psy-chosocial symptom burden is considerable.Patients reported a mean of nine discrete symptoms including edema, lethargy, dyspnea, syncope, depression, anxiety, spiritual distress, and a reduced health-related quality of life (HRQL). 2,3pecialty-level palliative care (SPC) (where a dedicated team of trained palliative care providers are consulted to assist in the care of a patient) involvement in the care of patients with HF is associated with improved HRQL, reduced symptom burden, and significant reductions in health care utilization, including a 42% reduction in the risk of rehospitalization with home-based SPC. 4 Patients with HF can often benefit from palliative care for prolonged periods.SPC prioritizes quality of life by managing end-stage HF symptoms, advanced care planning, and a multidisciplinary care approach. 5Some of these tasks can be provided by primary providers, shifting as needed to more specialized care.By approaching high-disease burden this way, we can better set expectations and provide the care patients and their families want.
Although there is a lack of consensus regarding referral criteria for the involvement of palliative care in patients with heart failure, various organizations in the United States and worldwide, such as the American College of Cardiology, the International Society for Heart and Lung Transplantation, and the European Society of Cardiology, have made recommendations supporting the early integration of palliative care into HF management. 6Because of the lack of consensus and unfamiliarity with SPC, most SPC occurs late in the disease course.For this reason, we undertook a retrospective analysis of the timing of SPC in HF decedents within the US Department of Veterans Affairs (VA) and analyzed the results for relevant care trends over a 10-year period to help determine if a baseline penetrance level of SPC exists which can be used as a metric for systemwide improvement to increase SPC access and allow earlier SPC in the HF disease course.

Ethics
This study was approved by the Providence Veterans Administration Medical Center Institutional Review Board.Informed consent was waived.

Cohort
This study examined the national population of patients with HF who received their care at the VA from January 1, 2011 to June 30, 2021.Using VA electronic medical records (EMR) we identified deceased Veterans with HF who experienced at least one inpatient admission to the VA from January 1, 2011 to June 30, 2021.

Measurement of palliative care
We define a SPC encounter by the presence of VA stop code (351 or 353).Unique to the VHA, stop codes are mechanisms for workload credit for encounters but do not describe the makeup of an encounter.Prior studies have shown validity in using VHA stop codes to identify SPC encounters. 7Those with no identified palliative care stop code were considered not to have VA-delivered SPC.

Time from palliative care to death
From the cohort, we identified veterans' dates of death from the VA Death Registry, and retrospectively examined the VA EMR for PC stop codes.Treating the earliest SPC stop code as the date of first SPC involvement, we categorized veterans into three cohorts: SPC beginning more than 90 days preceding death; SPC beginning <90 days before death; and no SPC before death.The 90-day cutoff was based on prior literature used for oncologic patients receiving SPC. 8,9riables VA data sets were used to obtain demographic information (age, race, sex).Comorbidities were collected using International Classification of Diseases (ICD)-9 and ICD-10 codes from the year before admission.Utilization (cost, days admitted) was collected from the VA accounting system.

Analyses
Characteristics of the cohorts were compared using standard mean difference (SMD), as prior studies have shown a preference for SMD in large populations.For reference, an SMD >0.2 is considered unbalanced.We examined all decedents within a calendar year as an individual cohort.We did not analyze or determine individual differences between the yearly cohorts as we believed the trends were self-evident.

Outcomes
The primary outcome was time from palliative care to death in the year of death.Secondary outcomes were the number and proportion of Veterans receiving palliative care in the year of death, along with SPC encounters on the last day of life for each cohort.

Results
Among the 232,079 Veteran decedents with HF and one or more admissions during the study period (Table 1), the mean age (standard deviation [SD]) was 76.5 (10.7) and 227,528 (98%) were men.A significant majority-172,469 (74%)-of these Veterans were White.The mean Elixhauser Comorbidity Index (SD) was 7.9 (3.4).Hypertension, diabetes mellitus, chronic lung disease, arrhythmias, electrolyte imbalances, and renal failure each had greater than 50% prevalence in this population.The mean number of SPC consultations (SD) for this population was 2.1 (6.7), and the mean total number of days hospitalized in the final year of life (SD) was 14 (20).
The time from the initial SPC encounter to death is shown in Table 2. Notably, the number of Veterans with HF and at least one admission who died each year increased annually from 10,608 in 2011 to 30,019 in 2020 (Fig. 1).Over the entire decade observed, 56.6% (n = 131,122) of these Veterans did not undergo a single SPC encounter at VA medical centers.The proportion of Veterans receiving any SPC increased from 41.6% (n = 4415) in 2011 to 42.5% (n = 6536) in 2021 (Fig. 2).Among Veterans that did receive SPC, the proportion with SPC involvement more than 90 days before death increased annually, from 10.31% (n = 1094) in 2011 to 26.2% (n = 4028) in 2021 (Fig. 3).The proportion with SPC initiation one or fewer days before death declined from 2.5% (n = 271) in 2011 to 0.9% (n = 140) in 2021.

Discussion
This retrospective cohort study found that over the 10year period from January 2011 to June 2021, SPC was initiated earlier in the disease course for Veterans with HF; however, there was only minimal change in the proportion of patients receiving SPC.As HF is a disease with a high symptom burden and a nonlinear functional decline, offering SPC earlier in the disease course is essential to enhance HRQL and achieve positive outcomes for patients with HF.It is concerning that the overall percentage of veterans receiving SPC increased only modestly despite the earlier involvement of SPC.Notably, all patients were treated as inpatients in the same health care system, with eligibility and coverage determined by the same national criteria, thus removing much of the variability commonly encountered in obtaining SPC.
The trend toward earlier SPC involvement in HF cases correlates with care mandates, heightened SPC staffing, and administrative ease in SPC referrals across the VA system.As recently as 2000, only 38% of VAMCs offered inpatient SPC programs, a number which approaches 100% today following the US Department of Veterans Affairs Medical Center Hospice and Palliative Care Initiative in 2001 and the Comprehensive End-of-Life Care Initiative in 2009. 10These initiatives were launched in the setting of a nationwide cultural and financial recognition of SPC's benefits, as well as increasing demand by a growing population of veterans with HF being treated at VAMCs.The trend toward earlier SPC involvement with HF patients may also suggest increased recognition of SPC's benefits and willingness to consult SPC colleagues among clinicians.The minimal change in the proportion of veterans with HF who receive  SPC, however, suggests that the well-documented stigma, fear, or knowledge deficits regarding SPC persist within this patient population and their families, even though the VA allows Veterans to receive disease-modifying treatments while on hospice, or concurrent care.Additionally, some patients in this study may have received treatment for and/or died of comorbid conditions, whose treating clinicians may be less familiar with SPC's availability and indications for those conditions.Finally, despite the potential for relative equalization of care within VA facilities, geographic and economic disparities remain that may make SPC less accessible to some veterans.Information on the precise staffing levels of SPC departments at individual VAMCs is also imprecise, and this data may reflect a persistent scarcity of SPC resources at some VAMCs.The VA's unique population and care model pose limitations to the generalizability of these findings.For instance, our cohort in this study was 98% male and 74.3% White, and the VA at large serves a population that is disproportionately rural. 11The numerous comorbidities of many patients in this cohort and the treatment thereof may have impacted when and whether SPC was implemented.Further, an in-depth review of documentation to assess why SPC was or was not consulted for patient and clinical course is not feasible in a sample of this size, complicating efforts to identify the discrete impacts of staffing, institutional initiatives, and changes in medical and societal attitudes upon the frequency and timing of SPC in HF patients.Furthermore, as specific outcome measures are beyond the scope of this retrospective report, it should be noted that we used SPC as a surrogate marker for quality of care as SPC has shown to be beneficial for HF patients overall. 4In addition, our data are from VA databases and did not look at Medicare databases which most likely results in some Veterans who received SPC or hospice in the community being miscategorized.Future studies should combine both VA and CMS databases for clearer analysis.Lastly, even though stop codes are used to signify clinical encounters and get workload credit at VA medical centers, the process can be tedious and unreliable for tracking at times.A 2019 internal VA review found that SPC was provided without the corresponding stop code in 21.5% of consults systemwide, most likely leading to an undercount of Veterans who received SPC for HF and slightly skewing data to no SPC/Hospice. 12

Conclusion
Our study demonstrates sustained growth in early palliative care intervention in HF patients in tandem with the rapid expansion of availability within the VA system.As an increasing number of Veterans are undergoing earlier SPC, it is hoped that they are reaping the established benefits of SPC for a longer period of time thus improving the quality of care of these Veterans, which should strengthen the case for continued expansion of SPC staffing and involvement in the clinical care of patients with HF.Further inquiry is needed to determine the barriers to increasing the persistently static proportion of Veterans with HF who receive palliative care and how they might be surmounted.Such knowledge could ultimately benefit the quality of life of all patients who suffer from this disease and lead to high-quality collaboration between cardiology and palliative care.and editing.J.L.R.: conceptualization, methodology, software, validation, formal analysis, resources, data curation, writing-original draft, writing-reviewing and editing, visualization, supervision, project administration, funding acquisition.L.J.: software, validation, formal analysis.H.M.E.: formal analysis, visualization, writing-original draft, writing-reviewing and editing.J.S.P.: visualization, writing-reviewing and editing.W.C.W.: formal analysis, visualization, writingreviewing and editing.J.D.-D.: conceptualization, methodology, formal analysis, writing-original draft, writing-reviewing and editing, visualization, supervision, project administration, funding acquisition, supervision, project administration.

FIG. 1 .
FIG. 1. Number of patients who received or never received palliative care by year of death from January 1, 2011 to June 30, 2021.

Table 1 .
Patient Characteristics