Abstract
Background:
Clinical practice guidelines recommend integrating palliative care (PC) into the care of patients with heart failure (HF) to address their many palliative needs. However, the incidence rates of PC use among HF subtypes are unknown.
Methods:
We conducted a retrospective cohort study of patients with the following HF subtypes in the Department of Veterans Affairs: reduced ejection fraction (HFrEF), mid-range ejection fraction (HFmEF), and preserved ejection fraction (HFpEF). Patients were included at the time of HF diagnosis from 2011 to 2015 and followed until a minimum of five years or death. Incidence rates of receipt of PC (primary outcome) were calculated using generalized estimating equations. We evaluated the time to incident PC by HF subtype with Kaplan–Meier analyses and with adjusted restricted mean survival time.
Results:
Of the 113,555 patients, 69% were ≥65 years, 98% were male, 73% White, and 18% Black; 58% had HFrEF, 7% HFmEF, and 34% HFpEF. Twenty percent received PC during follow-up, and 66% died. Adjusted PC incidence rates were higher among patients with HFrEF (47 per 1000 person-years, confidence interval [95% CI] 43–52) than for HFmEF and HFpEF (42 per 1000 person-years, CI 38–47 for both). Restricting follow-up to five years, patients with HFrEF received PC six weeks earlier than patients with HFpEF. There was no significant difference in time to PC between patients with HFmEF versus HFpEF.
Conclusion:
About 1 in 20 patients with HFrEF and 1 in 25 patients with HFmEF and HFpEF receive PC annually. Patients with HFrEF receive PC sooner than patients with HFmEF and HFpEF.
Keywords: heart failure, hospice, incidence, palliative care, patterns
Introduction
Over 6.2 million adults live with heart failure (HF) in the United States today.1 This number is expected to increase by almost 50% over the next decade.1,2 In parallel, advances in medications and implantable technology can prolong life among patients with HF but may complicate medical decision making and the estimation of patient prognosis. Consequently, clinical practice guidelines include the recommendation to consult with palliative care (PC) providers for patients with advanced HF to assist with symptom management, establish goals of care, and facilitate timely transitions to hospice.3–7
HF is categorized by quantitative assessment of left ventricular ejection fraction (LVEF), and care is guided by evidence-based clinical practice guidelines that differ by HF subtypes,4 including reduced ejection fraction (HFrEF), mid-range ejection fraction (HFmEF), and preserved ejection fraction (HFpEF).8,9 Patients with HFpEF were initially believed to have less severe HF and live longer relative to patients with HFrEF.10 However, investigators have found similar five-year mortality rates of 75.3%, 75.7%, and 75.7% among HFrEF, HFmEF, and HFpEF, respectively.11
Calls to integrate PC into HF management have increased over the past decade. Recently, authors of clinical guidelines emphasized the role of PC specifically for patients with HFrEF.3 However, guidelines for patients with HFmEF and HFpEF remain focused on risk mitigation and comorbidity management rather than care near the end of life.3,12 It is unknown whether rates of PC consultation are similar among HF subytpes, despite similar morbidity and mortality across these groups.11 To address this critical knowledge gap, we examined the differences and characteristics of incident and time to PC consultation and time from consultation to death by HF subtype. We conducted this analysis among patients with HF within the Department of Veterans Affairs (VA), the largest integrated health care system in the United States.
Materials and Methods
Data sources and study population
We conducted a retrospective observational study with data from the HEART-PAL cohort, a VA-based cohort designed to examine care near the end of life among patients with HF. We used data from two sources: (1) electronic health record data from the VA's Corporate Data Warehouse (CDW) and (2) VA Text Integration Utilities (TIU) documents. The Human Research Protection Program at the VA Connecticut Healthcare System approved this study. Our sample included all Veterans diagnosed with HF first identified within the VA system from July 1, 2011 to June 30, 2015 (n = 136,307).
Patients were followed for a minimum of five years or until death and were censored on July 1, 2020. We identified HF by either two outpatient visits with a primary ICD-9 diagnosis of HF (Supplementary Table S1) within 18 months or one hospital encounter record with a discharge diagnosis for HF in the primary billing position.13,14 We also allowed an 18-month lookback period before July 1, 2011 to account for outpatient HF visits preceding the start of the observation period.
We excluded 14.5% of the sample who did not have LVEF values up to six months before or six months after entry into the cohort (n = 19,820). Characteristics of patients without LVEF values are available in Supplementary Table S2. We also excluded patients with prior encounters (consultations or follow-up encounters) with PC (n = 2932, 2.4%). We included encounters with PC that occurred up to seven days before entry into the cohort to account for encounters with PC that occurred during hospitalization.
Measurement of ejection fraction
The LVEF values were obtained using a VA-developed natural language processing (NLP) tool. The tool has been validated in VA databases with positive predictive values of 0.968–1.000 and sensitivities of 0.801–0.899, respectively.15–17 The NLP system extracts numeric LVEF values from clinical/TIU documents, including progress notes, discharge summaries, echocardiography reports, nuclear medicine reports, cardiac catheterization reports, and other notes completed in either inpatient or outpatient settings. Among patients with multiple LVEF values, we selected the lowest value for analysis. We categorized patients into mutually exclusive groups by HF subtype as HFrEF: LVEF ≤40%; HFmEF: 40% < LVEF <50%; and HFpEF: LVEF ≥50%.9
Receipt of PC
Our primary outcome was an initial PC consultation provided by a VA-based Specialist PC Consultation Team. This measure was used to calculate the rate of first consultation by specialist PC per 1000 person-years. Identifying PC followed methods used in prior research and VA operations.18,19 We identified incident PC using a combination of Current Procedural Terminology codes, VA Stop Codes, and ICD-9-CM and ICD-10-CM codes. VA stop codes represent VA clinics and are attached to every encounter. Specialist PC Consultation Teams use codes 351 and 353 for both inpatient and outpatient encounters.
We identified a patient's first PC consultation and subsequent encounters thereafter. We identified whether the primary reason for the PC consultation was HF or not using ICD-9-CM and ICD-10-CM codes assigned to the encounter by the PC provider. We calculated the total number of PC encounters (i.e., initial and subsequent encounters). Finally, we calculated the time from cohort entry to PC consultation in days and from PC consultation to death in days among decedents during study follow-up.
Covariates and patient death
Covariates were selected to account for differences among HF subtypes at cohort entry. We included patient demographics, including age, sex, and race, obtained from the VA Vital Status File. We adjusted for global disease burden using the Charlson Comorbidity Index (CCI), a 12-item validated tool.20 The CCI was calculated from ICD-9 codes generated from inpatient and outpatient visits up to 18 months before entry.
We included the following comorbidities: chronic obstructive pulmonary disease (COPD), solid tumor malignancy, cerebral vascular disease, diabetes, myocardial infarction, and moderate to severe chronic kidney disease (CKD). We adjusted for how the patient entered the cohort (outpatient or during hospitalization) to account for HF disease severity at the time of cohort entry. We included cohort entry year and identified the VA medical center where patients had the plurality of their inpatient and outpatient visits during follow-up. We gathered patient death information from the CDW and the VA Vital Status File. We identified all-cause mortality during study follow-up and the proportion of patients who died while inpatient in a VA medical center.
Statistical analysis
We compared cohort characteristics by HF subtype and receipt of PC using chi-square and t-tests. Using Poisson regression, we first calculated the unadjusted incidence of PC for the entire cohort and by HF subtype and subsequently with adjustment for age, gender, race, the CCI, individual comorbidities, entry type, and entry year. These models used generalized estimating equations to account for the clustering of patients within different VA medical centers.
We compared unadjusted time to receipt of incident PC and from receipt of PC to death by HF subtype using Kaplan–Meier survival analysis. We further examined these times using restricted mean survival time (RMST)21 with adjustment for the aforementioned covariates. The RMST is an average measure of the time until the event of interest occurs within a restricted (i.e., pre-specified) time frame. Within each specified time frame, differences in RMST can be interpreted as either gains or losses in days until the event of interest.
Because median time-to-event does not lend itself to a simple calculation of confidence intervals (CIs), RMST is a more robust surrogate.22 As an absolute measure of the outcome (average time to the event of interest), RMST serves as a more intuitively interpretable alternative to the hazard ratio,23 a measure of relative risk. We calculated adjusted RMST for pairings of HF subtypes and reported RMST and differences in RMST for time to initial PC consultation at two and five years and time from consultation to death at one year of follow-up.24 In all cases, statistical significance was defined as a two-tailed p-value <0.05, and all analyses were completed in SAS version 9.4.
Results
Sample characteristics
The sample included 113,555 patients with HF, contributing 555,461 person-years of follow-up. Fifty-eight percent of the sample had HFrEF, 7.4% HFmEF, and 34.3% HFpEF (Table 1). Sixty-nine percent of the sample were 65 years of age or older, 97.8% were male, 73.4% were White, and 18.2% were Black. Multimorbidity was common, with a mean baseline CCI score of 4.18 (standard deviation = 2.5). Forty-six percent of the sample had COPD, 21% had diabetes, and 30% had moderate to severe CKD. More than three-quarters of the sample entered the cohort via two outpatient encounters for HF. Almost 66% of the sample died during study follow-up, and 19.8% died in a VA medical center.
Table 1.
Sample Characteristics N = 113,555 of Patients with Heart Failure Receiving Healthcare at the VA
| |
Overall |
HFrEF |
HFmEF |
HFpEF |
|
||||
|---|---|---|---|---|---|---|---|---|---|
| |
113555 |
66255 |
8391 |
38909 |
|
||||
| N | % | N | % | N | % | N | % | p a | |
| Age in years | <0.001 | ||||||||
| Less than 65 | 34,933 | 30.8 | 22,282 | 33.6 | 2394 | 28.5 | 10,257 | 26.4 | |
| 65-75 | 35,964 | 31.7 | 20,867 | 31.5 | 2701 | 32.2 | 12,396 | 31.8 | |
| 76-85 | 25,304 | 22.3 | 13,978 | 21.1 | 1993 | 23.8 | 9333 | 24.0 | |
| 86 and older | 17,354 | 15.2 | 9128 | 13.8 | 1303 | 15.5 | 6923 | 17.8 | |
| Sex, Male | 111,024 | 97.8 | 65,125 | 98.3 | 8238 | 98.2 | 37,661 | 96.8 | <0.001 |
| Race | <0.001 | ||||||||
| White | 83,361 | 73.4 | 47,490 | 71.7 | 6356 | 75.7 | 29,515 | 75.8 | |
| Black | 20,713 | 18.2 | 13,009 | 19.6 | 1355 | 16.2 | 6349 | 16.3 | |
| Other | 2829 | 2.5 | 1690 | 2.6 | 216 | 2.6 | 923 | 2.4 | |
| Unknown | 6652 | 5.9 | 4066 | 6.1 | 464 | 5.5 | 2122 | 5.5 | |
| Cohort entry type | |||||||||
| Inpatient | 22,870 | 20.1 | 13,995 | 21.1 | 1426 | 17.0 | 7449 | 19.1 | <0.001 |
| Outpatient | 90,685 | 79.9 | 52,260 | 78.9 | 6965 | 83.0 | 31,460 | 80.9 | |
| Cohort entry year | 0.097 | ||||||||
| 2011 | 11,498 | 10.1 | 6745 | 10.2 | 862 | 10.3 | 3891 | 10.0 | |
| 2012 | 23,809 | 21.0 | 14,013 | 21.2 | 1729 | 20.6 | 8067 | 20.7 | |
| 2013 | 25,021 | 22.0 | 14,468 | 21.8 | 1896 | 22.6 | 8657 | 22.3 | |
| 2014 | 25,774 | 22.7 | 15,070 | 22.7 | 1950 | 23.2 | 8754 | 22.5 | |
| 2015 | 27,453 | 24.2 | 15,959 | 24.1 | 1954 | 23.3 | 9540 | 24.5 | |
| Charlson Comorbidity Score Mean, SD | 4.18 | 2.5 | 4.0 | 2.5 | 4.3 | 2.5 | 4.57 | 2.6 | <0.001 |
| Comorbidities at cohort entry | |||||||||
| Chronic Obstructive Pulmonary Disease | 51,759 | 45.6 | 27,623 | 41.7 | 3855 | 45.9 | 20,281 | 52.1 | <0.001 |
| Malignancy (solid tumor) | 17,977 | 15.8 | 9852 | 14.9 | 1383 | 16.5 | 6742 | 17.3 | <0.001 |
| Cerebrovascular Disease | 20,001 | 17.6 | 11,174 | 16.9 | 1573 | 18.7 | 7254 | 18.6 | <0.001 |
| Diabetes | 23,366 | 20.6 | 11,766 | 17.8 | 1910 | 22.8 | 9690 | 24.9 | <0.001 |
| Myocardial Infarction | 19,137 | 16.9 | 13,220 | 20.0 | 1438 | 17.1 | 4479 | 11.5 | <0.001 |
| Moderate to severe Chronic Kidney Disease | 34,440 | 30.3 | 18,427 | 27.8 | 2620 | 31.2 | 13,393 | 34.4 | <0.001 |
| Died | 74,865 | 65.9 | 43,155 | 65.1 | 5526 | 65.9 | 26,184 | 67.3 | <0.001 |
| Died inpatient | 14,852 | 19.8 | 8428 | 19.5 | 1065 | 19.3 | 5359 | 20.5 | 0.0061 |
From chi-square test of equality across subtypes.
HFmEF, heart failure with middle-range ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; SD, standard deviation.
There were several differences in demographic and clinical characteristics by HF subtype. Patients with HFrEF were younger than HFmEF and HFpEF patients at entry into the cohort (<65 years; HFrEF = 33.6%, HFmEF = 28.4%, HFpEF = 26.4%; p < 0.0.001). Patients with HFrEF were more likely to have had a previous myocardial infarction relative to patients with HFmEF and HFpEF and were more likely to enter the cohort during hospitalization for HF. Mortality was slightly lower among patients with HFrEF, with 65.1% of patients with HFrEF dying during study follow-up versus HFmEF = 65.9% and HFpEF = 67.3% patients, respectively (p < 0.001 for all).
Characteristics and incidence of PC by HF subtype
Of the sample, 20% (n = 22,885) received a PC consultation during study follow-up (Table 2). Patients who received PC were older (≥86 years; PC = 22.9%, no PC = 13.4%). They also had more comorbidities, including COPD (PC = 52.7%, no PC = 43.8%), moderate to severe CKD (PC = 38.4%, no PC = 28.3%), and solid tumor malignancy (PC = 22.8%, no PC = 14.1%; p < 0.001 for all) compared with patients who did not receive PC. Patients who received PC were more likely to die and more likely to die in a VA medical center (p < 0.001 for both).
Table 2.
Sample Characteristics N = 113,555 of Patients with Heart Failure Receiving Health Care within the Department of Veterans Affairs by Palliative Care Status
| |
No Palliative Care |
Palliative Care |
|
||
|---|---|---|---|---|---|
| |
90670 |
22885 |
|
||
| N | % | N | % | p a | |
| Age in years | <0.001 | ||||
| Less than 65 | 30,042 | 33.1 | 4891 | 21.4 | |
| 65-75 | 29,485 | 32.5 | 6479 | 28.3 | |
| 76-85 | 19,039 | 21.0 | 6265 | 27.4 | |
| 86 and older | 12,104 | 13.4 | 5250 | 22.9 | |
| Sex, Male | 88,597 | 97.7 | 22,427 | 98.0 | 0.0091 |
| Race | <0.001 | ||||
| White | 66,117 | 72.9 | 17,244 | 75.4 | |
| Black | 16,806 | 18.5 | 3907 | 17.1 | |
| Other | 2297 | 2.5 | 532 | 2.3 | |
| Unknown | 5450 | 6.1 | 1202 | 5.2 | |
| Cohort entry type | <0.001 | ||||
| Inpatient | 17,277 | 19.0 | 5593 | 24.4 | |
| Outpatient | 73,393 | 80.9 | 17,292 | 75.6 | |
| Cohort entry year | <0.001 | ||||
| 2011 | 9036 | 10.0 | 2462 | 10.7 | |
| 2012 | 18,666 | 20.6 | 5143 | 22.5 | |
| 2013 | 19,788 | 21.8 | 5233 | 22.9 | |
| 2014 | 20,862 | 23.0 | 4912 | 21.5 | |
| 2015 | 22,318 | 24.6 | 5135 | 22.4 | |
| Charlson Comorbidity Score Mean, SD | 4.0 | 2.4 | 4.92 | 2.7 | <0.001 |
| Heart failure subtype | <0.001 | ||||
| HFrEF | 53,202 | 58.7 | 13,053 | 57.1 | |
| HFmEF | 6757 | 7.4 | 1634 | 7.1 | |
| HFpEF | 30,711 | 33.9 | 8198 | 35.8 | |
| Comorbidities at cohort entry | |||||
| Chronic Obstructive Pulmonary Disease | 39,724 | 43.8 | 12,035 | 52.7 | <0.001 |
| Malignancy (solid tumor) | 12,773 | 14.1 | 5204 | 22.8 | <0.001 |
| Cerebrovascular Disease | 15,029 | 16.6 | 4972 | 21.8 | <0.001 |
| Diabetes | 18,313 | 20.2 | 5053 | 22.1 | <0.001 |
| Myocardial Infarction | 14,399 | 15.9 | 4738 | 20.7 | <0.001 |
| Moderate to severe Chronic Kidney Disease | 25,666 | 28.3 | 8774 | 38.4 | <0.001 |
| Died | 54,075 | 59.6 | 20,790 | 90.9 | <0.001 |
| Died inpatient | 5471 | 10.2 | 9111 | 43.8 | <0.001 |
From chi-square test of equality across subtypes.
HFmEF, heart failure with middle-range ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; SD, standard deviation.
The total sample's unadjusted incidence of PC consultation was 41.2 per 1000 person-years (95% CI 41.1–42.2; Table 3). Unadjusted incidence rates of PC consultation were highest among patients with HFpEF at 44.1 per 1000 person-years (95% CI 43.2–45.1) compared with HFrEF and HFmEF. When adjusted for covariates, patients with HFrEF had the highest incidence of PC consultation at 46.9 per 1000 person-years (95% CI 42.6–51.6; Fig. 1) compared with HFmEF at 42.3 per 1000 person-years (95% CI 38.5–46.2) and HFpEF at 42.1 per 1000 person-years (95% CI 37.8–46.9).
Table 3.
Characteristics of Palliative Care by Subtype of Heart Failure
| |
Overall |
HFrEF |
HFmEF |
HFpEF |
|
||||
|---|---|---|---|---|---|---|---|---|---|
| (N = 113555) | (N = 66255) | (N = 8391) | (N = 38909) | p a | |||||
| Unadjusted incidence of PC per 1000 person–years (95% CI) | 41.2 | 41.1–42.2 | 40.4 | 39.7–41.1 | 40.0 | 38.1–41.9 | 44.1 | 43.2–45.1 | <0.001 |
| N | % | N | % | N | % | N | % | ||
|---|---|---|---|---|---|---|---|---|---|
| Received PC |
22,885 |
20.2 |
13,053 |
19.7 |
1634 |
19.5 |
8198 |
21.1 |
<0.001 |
| Primary reason for PC was heart failure |
7495 |
32.7 |
5266 |
40.3 |
470 |
28.8 |
1759 |
21.5 |
<0.001 |
| Number of PC encounters, median (IQR) |
1 |
1–2 |
1 |
1–2 |
1 |
1–2 |
1 |
1–2 |
0.706 |
| Received PC and died | 20,790 | 90.9 | 11,864 | 90.9 | 1473 | 90.1 | 7453 | 90.9 | 0.596 |
From chi-square test of equality across subtypes.
CI, confidence interval; HFmEF, heart failure with middle-range ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; IQR, Interquartile range; PC, palliative care.
FIG. 1.
Adjusted incidence of palliative care consultation per 1000 person-years by heart failure subtype.
Patients with HFrEF were more likely to receive a PC consultation for HF compared with HFmEF and HFpEF (HFrEF = 40.3%, HFmEF = 28.8%, HFpEF = 21.5%; p < 0.001; Table 3). There was no difference in the number of PC encounters during study follow-up by subtype of HF (p = 0.706).
Time to initial PC consultation
In unadjusted analysis, patients with HFpEF had shorter time to PC consultation relative to the other HF subtypes (log-rank p < 0.001; Fig. 2, Panel A). With follow-up restricted to two years, adjusted RMST to PC consultation was the shortest for patients with HFrEF at 1.77 years (95% CI 1.75–1.78; Table 4) compared with HFmEF at 1.79 years (95% CI 1.78–1.82) and HFpEF at 1.80 years (95% CI 1.78–1.81). With follow-up restricted to five years, adjusted RMST to PC consultation among patients with HFrEF was 3.99 years (95% CI 3.94–4.04), and it was again shorter than for patients with either HFmEF or HFpEF.
FIG. 2.
Kaplan–Meier survival curves: from cohort entry to palliative care consultation and from consultation to death.
Table 4.
Restricted Mean Survival Time to Palliative Care Consultation and from Consultation to Death
| HFrEF | HFmEF | HFpEFa | |
|---|---|---|---|
| RMST to PC consultation at two years, years (95% CI) | 1.77 (1.75 to 1.78) | 1.79 (1.78 to 1.82) | 1.80 (1.78 to 1.81) |
| RMST difference to PC consultation at two years, days (95% CI) | −11.69 (−13.99 to −9.39)b | −0.98 (−3.06 to 5.02)c | — |
| RMST to PC consultation at five years, years (95% CI) | 3.99 (3.94 to 4.04) | 4.08 (4.02 to 4.14) | 4.10 (4.05 to 4.15) |
| RMST difference to PC consultation at five years, days (95% CI) | −41.92 (−49.07 to −34.77)b | −8.08 (−21.60 to 4.0)c | — |
| RMST from PC consultation to death at one year, years (95% CI) | 4.80 (4.48 to 5.11) | 4.95 (4.57 to 5.33) | 5.02 (4.69 to 5.34) |
| RMST difference from PC consultation to death at one year, days (95% CI) | −6.78 (−10.93 to −2.63)b | 2.12 (−9.86 to 5.61)c | — |
HFpEF = Referent.
RMST difference p-value <0.001.
RMST difference p-value >0.05.
RMST, restricted mean survival time.
Patients with HFrEF received a PC consultation 12 days sooner over two years of follow-up (95% CI −13.99 to −9.38; p < 0.001) and 42 days sooner over five years of follow-up (95% CI −49.07 to −34.77; p < 0.001) than patients with HFpEF. There was no difference in days until PC consultation between patients with HFmEF and HFpEF during either two or five years of follow-up (p > 0.05).
Time from PC consultation to death
In unadjusted analysis, patients with HFrEF had a shorter time from PC consultation to death than the other subtypes (log-rank p < 0.041; Fig 2, Panel B). With follow-up restricted to one year, adjusted RMST from PC consultation to death during was 4.80 months (95% CI 4.48–5.11; Table 4) among patients with HFrEF, 4.95 months (4.57–5.33) among patients with HFmEF, and 5.02 months (95% CI 4.69–5.34) among patients with HFpEF. Patients with HFrEF died seven days sooner after the PC consultation than patients with HFpEF (95% CI −10.93 to −2.64; p < 0.001). There was no significant difference in adjusted RMST from PC to death between patients with HFmEF compared with HFpEF (p > 0.05).
Discussion
In this large cohort of patients with HF in the VA, we found that 1 in 20 patients with HFrEF and 1 in 25 patients with HFmEF and HFpEF received PC annually. Accounting for differences in demographic and clinical characteristics, we found small but statistically significant differences in time to initial PC consultation among patients with HFrEF, but not HFmEF, compared with patients with HFpEF. These findings have clinical significance, as investigators have found similar morbidity and mortality among these subtypes.11 Thus, patients with HFmEF and HFpEF who could benefit from PC may receive PC less frequently or may receive PC later in their disease trajectory than their HFrEF counterparts.
In our study, patients with HFrEF were more likely to receive a PC consultation for HF compared with the other HF subtypes. Clinicians may be more familiar with referring patients with HFrEF to PC, which could explain why patients with HFrEF more often received PC, received PC sooner, and received PC for HF compared with the other HF subtypes. Historically, clinical practice guidelines were more focused on the treatment of patients with HFrEF, including care near the end of life3 compared to patients with HFmEF and HFpEF. Thus, clinicians may recognize the palliative needs of patients with HFrEF more readily than other HF subtypes, particularly those HFrEF patients needing advanced implantable devices. Clinicians might also associate increasing symptom burden and other palliative needs of patients with HFmEF and HFpEF with chronic conditions other than HF. Given our findings, additional research is needed to determine if variation in the rates of PC consultation and timing by HF subtype is associated with differences in care outcomes.
More than 20% of the sample received PC during follow-up. Rates of PC consultation among patients with HF are variable worldwide. Investigators reported rates of PC consultation of 4% in the United Kingdom to 13% in the United States,25 and as high as 47% among community-dwelling decedents who died from HF in Ontario, Canada.26 Our findings likely represent rising trends over time to refer more patients with HF to PC. Investigators have documented these trends in past reports.27,28 Our findings also reflect evolving clinical practice guidelines, which increasingly recommend PC for HF patients.3,4,29
Our study is representative of PC delivery within the VA, a health care system that has prioritized increasing access to specialist PC for all patients with serious illness.30 These VA policies likely also contributed to the increased consultation rates found in this study. We also followed patients over time, with 66% of the total sample dying during study follow-up. As our cohort not only comprised decedents, this could explain why rates of PC consultation were lower than in other studies using mortality follow-back designs.
Most patients in our study received PC shortly before death, with >90% dying after their initial PC consultation. Many factors contribute to delayed referrals to PC among HF patients. These include difficulty determining prognosis10 and clinician and patient perceptions of PC as synonymous with hospice.31 The ideal timing to refer patients with HF to PC is unknown. However, consulting PC near the end of life may result in lost opportunities for PC providers to discuss care goals, connect patients with services, and implement effective symptom management. As PC is, by nature, a multidimensional intervention, there is a need for additional research to determine the ideal components, dose, and timing of PC for this population.
This study has several limitations. We excluded patients without available LVEF values from the analysis, which might result in selection bias. Potentially these patients received care for HF primarily outside of the VA and therefore did not have documented EF values for inclusion into the cohort. However, missing LVEF values were not missing at random, which precluded our ability to impute them.
There is also the potential for the crossover of HF subtype, as LVEF values of the cohort may improve with guideline-directed medical therapy. Misclassification of LVEF values could also occur due to errors in the NLP program. We also did not require patients to have HF for a minimum time period. These limitations could alter rates of PC consultation by HF subtype. In the past decade, there has been a rapid acceleration of medications and devices for patients with HF that improve cardiac function and may extend life. Although these therapies likely do not alter the eventual need for PC among patients with HF, they may alter the timing.
We did not examine hospice use by HF subtype, as most patients died outside of the VA. This analysis would require the inclusion of Medicare claims data and is the focus of our future research. We did not have clinical data, including blood pressure or pulse values, brain natriuretic peptide values, or New York Heart Functional Class, functional status, symptoms, or symptom severity. These characteristics may also account for differences in PC consultation and timing. Finally, as this is a primarily male cohort, findings may not be generalizable to female patients with HF. However, investigators have found similar rates of PC consultation among people with HF in Veteran and Nonveteran mixed-sex populations, suggesting the generalizability of populations across health care systems.27,28
Conclusions
About 1 in 20 patients with HFrEF and 1 in 25 patients with HFmEF and HFpEF receive PC annually. Palliative care consultation in HF remains infrequent and is influenced by HF subtype. When it does occur, patients with HF generally receive their initial PC consultation within the last five months of life. Future research is needed to identify the patients with HF who would benefit the most from PC and determine the most efficacious timing of PC referral.
Supplementary Material
Authors' Contributions
All authors contributed significantly to this work. Study concept and design: All; Data analysis: S.L.F., T.E.M., E.A.A. Data interpretation: All; Preparation of article: All Sponsor's Role: The analysis described here is based on work supported by the National Heart, Lung, and Blood Institute, the Yale Center for Implementation Science, and the Department of Veterans Affairs, Veterans Health Administration, which had no role in the design, methods, participant recruitment, data collection, analysis, or preparation of this article or in the decision to submit this article for publication.
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the United States Department of Veterans Affairs or the United States Government.
Funding Information
S.L.F. received funding from K12HL138037 from the National Heart Lung and Blood Institute and the Yale Center for Implementation Science, and T.F./T.E.M. received funding from the Yale Claude D. Pepper Older Americans Independence Center (P30AG021342).
Author Disclosure Statement
No competing financial interests exist.
Supplementary Material
Correction added on September 21, 2022 after first online publication of June 28, 2022: In the third paragraph of the Materials and Methods section, the first sentence of the paragraph was inadvertently cut off. The sentence has been corrected and it is now complete.
References
- 1. Virani SS, Alonso A, Benjamin EJ, et al. : Heart disease and stroke statistics—2020 update: A report from the American Heart Association. Circulation 2020;141:e139–e596. [DOI] [PubMed] [Google Scholar]
- 2. Heidenreich PA, Albert NM, Allen LA, et al. : Forecasting the impact of HF in the United States: A policy statement from the American Heart Association. Circ Heart Fail 2013;6:606–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Writing Committee, Maddox TM, Januzzi JL Jr., et al.: 2021 Update to the 2017 ACC Expert Consensus Decision Pathway for Optimization of HF Treatment: Answers to 10 pivotal issues about HF with reduced ejection fraction: A report of the American College of Cardiology Solution Set Oversight Committee. J Am Coll Cardiol 2021;77:772–810. [DOI] [PubMed] [Google Scholar]
- 4. Yancy CW, Jessup M, Bozkurt B, et al. : 2013 ACCF/AHA guideline for the management of HF: A report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. Circulation 2013;128:e240–e319. [DOI] [PubMed] [Google Scholar]
- 5. Fang JC, Ewald GA, Allen LA, et al. : Advanced (stage D) HF: A statement from the HF Society of America Guidelines Committee. J Cardiac Fail 2015;21:519–534. [DOI] [PubMed] [Google Scholar]
- 6. Allen LA, Stevenson LW, Grady KL, et al. : Decision making in advanced HF: A scientific statement from the American Heart Association. Circulation 2012;125:1928–1952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Feldman D, Pamboukian SV, Teuteberg JJ, et al. : The 2013 International Society for Heart and Lung Transplantation Guidelines for mechanical circulatory support: Executive summary. J Heart Lung Transplant 2013;32:157–187. [DOI] [PubMed] [Google Scholar]
- 8. Ponikowski P, Voors AA, Anker SD, et al. : ESC Scientific Document Group (2016). 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) Developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J, 37(27), 2129–2200. 10.1093/eurheartj/ehw128 [DOI] [PubMed] [Google Scholar]
- 9. Bozkurt B, Hershberger RE, Butler J, et al. : 2021 ACC/AHA key data elements and definitions for HF: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards (Writing Committee to Develop Clinical Data Standards for HF). Circ Cardiovasc Qual Outcomes 2021;14:e000102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Warraich HJ, Allen LA, Mukamal KJ, et al. : Accuracy of physician prognosis in HF and lung cancer: Comparison between physician estimates and model predicted survival. Palliat Med 2016;30:684–689. [DOI] [PubMed] [Google Scholar]
- 11. Shah KS, Xu H, Matsouaka RA, et al. : HF with preserved, borderline, and reduced ejection fraction: 5-Year outcomes. J Am Coll Cardiol 2017;70:2476–2486. [DOI] [PubMed] [Google Scholar]
- 12. Hsu JJ, Ziaeian B, and Fonarow GC: HF with mid-range (borderline) ejection fraction: Clinical implications and future directions. JACC Heart Fail 2017;5:763–771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Justice AC, Dombrowski E, Conigliaro J, et al. : Veterans Aging Cohort Study (VACS): Overview and description. Med Care 2006;44:1–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Haskell SG, Mattocks K, Goulet JL, et al. : The burden of illness in the first year home: Do male and female VA users differ in health conditions and healthcare utilization. Women's Health Issues 2011;21:92–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Patterson OV, Freiberg MS, Skanderson M, et al. : Unlocking echocardiogram measurements for heart disease research through natural language processing. BMC Cardiovasc Disord 2017;17:151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Freiberg MS, McGinnis Ka, Kraemer K, et al. : The association between alcohol consumption and prevalent cardiovascular diseases among HIV-infected and HIV-uninfected men. JAIDS 2010;53:247–253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Patel YR, Robbins JM, Kurgansky KE, et al. : Development and validation of a HF with preserved ejection fraction cohort using electronic medical records. BMC Cardio Disord 2018;18:128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Feder SL, Redeker NS, Jeon S, et al. : Validation of the ICD-9 diagnostic code for PC in patients hospitalized with HF within the Veterans Health Administration. Am J Hosp Palliat Care 2018;35:959–965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Feder SL, Tate J, Ersek M, et al. : The association between hospital end-of-life care quality and the care received among patients with HF. J Pain Symptom Manage 2021;61:713–722.e711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Quan H, Li B, Couris CM, et al. : Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol 2011;173:676–682. [DOI] [PubMed] [Google Scholar]
- 21. Stensrud MJ and Hernán MA: Why Test for proportional hazards? JAMA 2020;323:1401–1402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Kloecker DE, Davies MJ, Khunti K, and Zaccardi F: Uses and limitations of the restricted mean survival time: Illustrative examples from cardiovascular outcomes and mortality trials in type 2 diabetes. Ann Intern Med 2020;172:541–552. [DOI] [PubMed] [Google Scholar]
- 23. Royston P and Parmar MKB: Restricted mean survival time: An alternative to the hazard ratio for the design and analysis of randomized trials with a time-to-event outcome. BMC Med Res Methodol 2013;13:152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Kim DH, Uno H, and Wei L-J: Restricted mean survival time as a measure to interpret clinical trial results. JAMA Cardiol 2017;2:1179–1180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Alqahtani F, Balla S, Almustafa A, et al. : Utilization of PC in patients hospitalized with HF: A contemporary national perspective. Clin Cardiol 2019;42:136–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Quinn KL, Hsu AT, Smith G, et al. : Association between palliative care and death at home in adults with heart failure. J Am Heart Assoc 2020;9:e013844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Robinson MR, Al-Kindi SG, and Oliveira GH: Trends in palliative care use in elderly men and women with severe heart failure in the United States. JAMA Cardiol 2017;2:344. [DOI] [PubMed] [Google Scholar]
- 28. Mandawat A, Heidenreich PA, Mandawat A, and Bhatt DL: Trends in palliative care use in veterans with severe heart failure using a Large National Cohort. JAMA Cardiol 2016;1:617–619. [DOI] [PubMed] [Google Scholar]
- 29. Yancy CW, Jessup M, Bozkurt B, et al. : 2017 ACC/AHA/HFSA Focused Update of the 2013 ACCF/AHA guideline for the management of heart failure: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America. J Card Fail 2017;23:628–651. [DOI] [PubMed] [Google Scholar]
- 30. Miller SC, Intrator O, Scott W, et al. : Increasing Veterans' hospice use: The Veterans Health Administration's focus on improving end-of-life care. Health Aff (Millwood) 2017;36:1274–1282. [DOI] [PubMed] [Google Scholar]
- 31. Kavalieratos D, Mitchell EM, Carey TS, et al. : “Not the ‘grim reaper service’”: An assessment of provider knowledge, attitudes, and perceptions regarding palliative care referral barriers in heart failure. JAHA 2014;3:e000544. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.