Economic Impact of Early Inpatient Palliative Care Intervention in a Community Hospital Setting

Abstract

Background:

Inpatient palliative care programs have demonstrated financial benefit for the hospital and improved quality of care for patients with advanced disease. Previous studies on this subject have focused on comparisons between palliative and traditional care. The financial and clinical effects of early versus late palliative care intervention are less well documented.

Objective:

The aims of this study are to review the financial and quality outcomes that early palliative care intervention has on appropriate inpatients in the community hospital setting.

Materials and Methods:

This retrospective study analyzed 449 palliative care patients. The independent variable was days to palliative care consultation, characterized as early palliative care (≤3 days) and late palliative care (>3 days). Dependent variables included length of stay (LOS) and financial considerations. The two groups were further stratified according to case mix index, medical versus surgical, as well as certain disease groups, such as sepsis, congestive heart failure (CHF), and chronic obstructive pulmonary disease (COPD) exacerbation. The patient's functional status, measured by the Victoria Palliative Performance Scale (PPSv2) was calculated to determine if this variable independently influenced the timing of consultation.

Results:

Patients in the early intervention group realized a reduction in LOS and a significant cost reduction. In the analysis of the entire group, the average LOS with early intervention was 6.09 days versus 16.5 days with late intervention (p < 0.001). The early intervention group demonstrated an earlier transition to comfort care, earlier referral to outpatient hospice, and did not have a negative effect on mortality. The patient's PPSv2 score did not influence the timing of intervention (p 0.25).

Conclusion:

Early intervention with inpatient palliative care consultation correlated with financial benefit as well as earlier referral to more appropriate levels of care. These effects were achieved with minimal expense in a medium-sized community hospital.

Introduction

As an adjunct to the more traditional care of patients in the hospital setting, the concurrent use of palliative care services is now gaining widespread acceptance.¹ The need for use of inpatient hospitalization and its related services is increasingly directed toward more frail, elderly patients who often have multiple advanced medical problems.

While most third-party payers help reimburse part of the expenses of home palliative care by compensating licensed independent practitioners, some of the other services necessary for the care of these patients, such as chaplains, social workers, and registered nurses, may not be covered. For example, in western New York if the patient belongs to a non-fee-for-service-managed Medicare program or Medicaid, then there will be full coverage for palliative care at home. This is not true for those patients with a fee-for-service Medicare or other traditional insurances. Many of these patients have not been introduced to, or are not eligible for outpatient hospice. Hospitalization usually becomes the only option when their chronic disease decompensates. This creates a financial burden for the medical system^2,3 and often exposes these patients to testing and procedures that may cause more suffering, without any obvious improvement in quality or patient outcomes.^4–6

Inpatient palliative care services have been demonstrated to provide more appropriate care for these patients once they are admitted.^7,8 Additionally, with inpatient palliative care, there is a reduction in the use of intensive care services, expensive laboratory and radiology tests, and invasive procedures.^9–14 This results in a reduction in costs to the hospital and improves the overall quality of care for patients.^15,16

Many of the financial studies done thus far have shown financial benefit largely achieved through cost avoidance. It would seem intuitively obvious that if earlier palliative care intervention is implemented, greater cost savings should be realized. This comparison was demonstrated in one large multicenter oncologic study.¹⁷ However, it was unclear to those authors whether these results could be extrapolated to all palliative care patients on the medical–surgical service in the community hospital setting.

The effects of early palliative care intervention versus late intervention will be investigated in this analysis.

Materials and Methods

Study design and setting

Kenmore Mercy Hospital is a medium-sized community-based hospital that is part of a larger hospital system (Catholic Health System of Buffalo), with an average census of ∼130 patients. The need for a palliative care service was identified and implemented in 2013. Although there are two inpatient acute hospice beds, there is no designated palliative care unit; therefore, the care of these patients was rendered on all of the clinical units. The palliative care team includes one physician who performs all of the consults, palliative care nurse champions on each clinical floor, and participation from case management, spiritual care, and social services. The nurse champions are voluntary and have received specific palliative care training, which allows them to act as a resource for the nurses on each clinical unit. Although the physician consultant is not board-certified in palliative care, he is board certified in internal medicine, pulmonary medicine, and critical care.

All palliative care interventions reviewed in this study were implemented at the specific request of the attending physician.

For this study an analysis was conducted for all 449 palliative care patients encountered in 2015 and 2016. Early intervention was defined as the initiation of palliative care at three days or less and late intervention was defined as four days or greater. There were 228 patients with early intervention and 221 patients with late intervention. The palliative care service frequently used the Geisinger screening tool (Appendix 1) to assist nursing staff and case management to identify appropriate palliative care patients and alert attending physician to this need.

Source of data

Cerner Sorian financials was used to extract payment information and Lawson G/L was used to obtain cost data (not charges) for each patient encounter. Factors, such as employee salaries and benefits, laboratory costs, X-ray costs, and pharmacy could all be assigned to each encounter. This type of analysis has been used in similar studies.¹⁷

Measures

Outcomes

We measured the duration of total hospital stay as length of stay (LOS). Additionally, to measure financial outcomes, we obtained expected revenue, margin cost, and contributing margin for each encounter from the database noted above.

Predictors

The key independent variable was days to palliative care consultation after admission, which was categorized as early palliative care (≤3 days) and late palliative care (>3 days). Other predictors measured demographic characteristics of the subjects, including: age, gender, race, religion, insurance, and PPSv2 (Appendix 2). Age was categorized as less than 65, 65–79, and 80 plus. Gender included female and male. Religion included Christian, none, and other. Health insurance was grouped as Medicare, Medicaid, and others (commercial, private, etc.).

In 2016, the PPSv2 was calculated for all patients on the palliative care service. This was based on the time immediately preceding admission. The goal was to distinguish if the patient's admission functional status influenced the decision to initiate a palliative care consultation. A PPSv2 score of 70–100% was designated as stable, a score of 40–60% was transitional, and less than 30% classified as near end of life or possibly hospice appropriate.^18,19

Statistical analyses

All analyses were performed using Minitab for Windows, version 17.1. The significant level was set at P < 0.05. Descriptive statistics (% or mean + standard deviation) were used to describe the characteristics of the overall study population. Chi-square test for categorical variables was performed to compare patient demographics between the early and late palliative care group.

Standard regression analysis with covariate adjustments (ANCOVA) was performed on the association between days to palliative care consult and the outcomes of LOS, expected revenue, margin cost, and contribution margin. The adjusted regression model controlled for age, race, religion, insurance, and PPSv2.

The study population was further stratified according to case mix index (CMI), medical versus surgical admissions, as well as some common diagnosis related group (DRG) categories, such as sepsis. Two-sample t test (unadjusted for covariates) was performed for surgical and sepsis patients due to small/unbalanced sample size and inequity of variance. Furthermore, Chi-square test for categorical variables was performed to compare patient disposition outcome between early and late palliative care group.

Results

Specific patient demographics are summarized in Table 1.

Table 1.

Study Population by Days of PC Consultation

Variables	≤3 days N = 228 n (%) or mean	>3 days N = 221 n (%) or mean	p
Age group (years)			0.001
<65	24 (10.35%)	35 (15.84%)
65–79	43 (18.86%)	68 (30.77%)
>79	161 (70.61%)	118 (53.39%)
Avg. age	83.4%	78.5%	<0.001
Gender			0.708
Female	135 (59.21%)	127 (57.47%)
Male	93 (40.79%)	94 (42.53%)
Race			0.046
White	222 (97.37%)	204 (92.31%)
Black	4 (1.75%)	9 (4.07%)
Other	2 (0.88%)	8 (3.62%)
Religion			0.077
Christian	200 (87.72%)	192 (86.88%)
None	10 (4.39%)	19 (8.60%)
Other	18 (7.89%)	10 (4.52%)
Case type			<0.001
Medical	213 (93.42%)	166 (75.11%)
Surgical	15 (6.58%)	55 (24.89%)
Insurance			0.188
Medicare	203 (89.04%)	193 (87.33%)
Medicaid	13 (5.70%)	21 (9.50%)
Other	12 (5.26%)	7 (3.10%)
Disease categories			0.001
Cancer	86 (37.72%)	56 (25.34%)
Dementia	46 (20.18%)	31 (14.03%)
Cardiovasc	34 (14.91%)	18 (8.14%)
COPD	28 (12.28%)	38 (17.19%)
Sepsis	21 (9.21%)	38 (17.19%)
Other	13 (5.70%)	40 (18.10%)

COPD, chronic obstructive pulmonary disease; PC, palliative care.

Statistical analysis of the study group demonstrated that the patients seen in the early intervention cohort tended to be slightly older. Late intervention with palliative care consultation was noted to have a higher number of surgical admissions. There were only minor differences in race and no differences in gender or insurance coverage. There was significant heterogeneity with respect to disease categories. Cancer and dementia tended to dominate the early intervention patients. Patients with sepsis and COPD exacerbation were also prominent in the late intervention group along with cancer and dementia.

The palliative care service started measuring the PPSv2 on patients in 2016 (Table 2). Although most referring physicians had a general idea of the patient's functional status at the time of admission, most were unfamiliar with the PPSv2 measurement, which was highlighted in each consult. Our analysis indicated that the PPSv2 score as a measurement of functional status did not have an initial influence on whether or not to request a palliative care consult.

Table 2.

PPSv2 (2016 Data Only)

Score	≤3 days, (N = 99), n (%)	>3 days, (N = 98), n (%)	p
≤30	15 (15.15)	12 (12.24)	0.255
40–60	62 (62.62)	54 (55.10)
70–100	22 (22.22)	32 (32.65)

Clinical outcomes

Final dispositions of the patients in our study are summarized in Table 3.

Table 3.

Outcomes

	≤3 days	>3 days	p
Total expired inpatients	93 (40.26%)	107 (48.20%)	0.089
Expired inpatient (non hospice)	73 (32.02%)	72 (32.58%)	0.899
Expired acute hospice	20 (8.77%)	35 (15.84%)	0.021
Discharged outpatient hospice	59 (25.88%)	33 (14.93%)	0.005
SNF admission (non hospice)	21 (9.21%)	20 (9.05%)	0.953
Subacute rehab	28 (12.28%)	27 (12.22%)	0.984
Other (home)	27 (11.84%)	34 (15.38%)	0.273

SNF, skilled nursing facility.

Early intervention with palliative care did not have an adverse effect on inpatient mortality. Forty percent of the early intervention group expired in the hospital. There was a 48% mortality rate in the late intervention group, which was not statistically significant. Sixteen percent of the late consultation group was admitted to inpatient hospice in contrast to 9% acute hospice admission with early the intervention group. This was statistically significant (p value 0.021). Early intervention demonstrated a tendency to admit to outpatient hospice either skilled nursing facility (SNF) or home, as compared with late intervention (p 0.005). Eighteen percent of all the patients (9% in each group) were transferred to a nursing home without hospice.

Financial analysis

Table 4 summarizes the results of the financial analysis.

Table 4.

Financial Outcomes

Variables	≤3 days	>3 days	p^a
All patients	N = 228	N = 221
Total LOS	6.09	16.5	0.000
Avg. revenue/case	$10,428	$24,543	0.006
Avg. cost/case	$6544	$22,626	0.012
Avg. contribution margin	$3884	$1917	0.078
Total contribution margin	$882,735	$411,450
Adjusted for CMI	N = 215	N = 153
Total LOS	5.8	13.3	0.000
Avg. revenue/case	$9144	$12,729	0.007
Avg. cost/case	$5898	$14,190	0.000
Avg. contribution margin	$3246	($1416) loss	0.000
Total contribution margin	$695,069	($223,483) loss
Medical cases	N = 213	N = 166
Total LOS	6.0	14.3	0.000
Avg. revenue/case	$9615	$16,258	0.003
Avg. cost/case	$6602	$16,286	0.000
Avg. contribution margin	$3413	($28.0) loss	0.024
Total contribution margin	$724,095	($16,829) loss
Surgical cases	N = 15	N = 55	t test, unadjusted^b
Total LOS	7.32	22.9	0.004
Avg. revenue/case	$21,981	$49,549	0.757
Avg. cost/case	$11,405	$41,762	0.679
Avg. contribution margin	$10,586	$7787	0.61
Total contribution margin	$158,428	$428,279
Sepsis cases	N = 34	N = 27	t test, unadjusted^b
Total LOS	7.32	14.85	0.001
Avg. revenue/case	$11,011	$19,087	0.021
Avg. cost/case	$7603	$20,434	0.00
Avg. contribution margin	$3408	($1348) loss	0.001

Adjusted for age, gender, race, religion, case type (medical vs. surg.), and PPS score.

Due to small sample size.

CMI, case mix index; LOS, length of stay.

The results of the above analysis showed that early intervention with palliative care had a significant influence on the LOS regardless of the CMI, medical versus surgical admission, or DRG. The reduced LOS had a significant effect on cost savings for the medical patients and those with similar CMIs.

For those patients with similar CMIs, the average contribution margin was $3246 in the early intervention group as opposed to a $1416 loss in the late intervention group. In aggregate, there was a profit of $695,069 with early intervention versus $223,483 loss with the late intervention group. An analysis of medical admissions (not adjusted for CMI) shows similar trends ($3413 average profit vs. $102 loss). The aggregate contribution margin for medical admissions was $724,095 profit in the early group versus a $16,829 loss in the late.

An unadjusted simple t test was used in the analysis of surgical patients due to the relatively small number of referrals for palliative care in that group. With surgical patients, who were not adjusted for CMI, the cost savings were not as pronounced, although an average profit of $10,576 was seen in early intervention as opposed to $7787 with late intervention. This difference was not statistically significant since the late intervention surgical group was much larger than the early intervention group. The surgical patients had a higher CMI reflecting a much greater DRG with potential for a higher reimbursement. Roughly 70% of the surgical cases were nonelective or urgent in nature.

For those patients who were admitted with severe sepsis, the average contribution margin was $3408 profit in the early intervention group versus a loss of $1348 in the late intervention group. In aggregate, there was a $36,386 loss for those patients in the late intervention group in contrast to a $115,873 profit for the early intervention group. As with the surgical patients, due to the small numbers, we could not perform a complete covariate analysis. Table 4 does not include the patients with other DRGs (congestive heart failure and COPD) since the numbers were quite low and statistical significance could not be achieved.

Discussion

Early intervention with outpatient palliative care in cancer patients has influenced the approach to end-of-life care even after patients have been admitted to the hospital. Earlier inpatient utilization of hospice and less implementation of traditional care has been demonstrated in that setting.²⁰ Ciemins et al. found a significant reduction in costs post-palliative care intervention highlighting the importance of early intervention. They also found significant improvement in the management of pain and dyspnea with these patients.¹¹ This study was done in a large academic center. There is a paucity of studies that has demonstrated the outcomes of early inpatient palliative care intervention for nononcologic inpatients in community-based hospitals. Reyes-Ortiz et al. saw a positive effect with early intervention (<3 days) for elderly frail patients. They demonstrated a shortened LOS post-palliative care intervention and greater referral to hospice services.²¹ The actual cost savings benefit in this study was not calculated.

Patients in the current study included those of all ages with multiple diagnoses. Early referral to inpatient palliative care correlated with a significant cost benefit as well as appropriate transition to either comfort care or outpatient hospice care. With late referral, the transition to end-of-life care or acute hospice was often delayed until the very end of the hospitalization. The financial benefit was most likely attributable to the significant reduction in the LOS.

As with most inpatient palliative care services, our program is dependent on a consultation request from the attending physician. This necessity tends to highlight an inherent barrier to earlier timing of palliative care intervention. Many of our physicians, especially surgeons, voice that they are not comfortable requesting a palliative care consultation and hence reluctant to request a service that is not involved with traditional, curative care. This sentiment is often seen with patient families and nursing staff as well.

One of the reasons why we chose to measure the PPSv2 score was to investigate whether or not the patient's functional status before admission influenced the decision by a physician to pursue palliative care intervention. As an independent variable, a low PPSv2 score in our patients did not demonstrate a major influence on the physician's willingness to have early palliative consultation and thus pursue the usual traditional care of these patients. As part of a process to foster early intervention with palliative care, education is being provided to the hospitalists, nurses, and case managers to utilize the PPSv2 score as a tool to help trigger appropriate cases for palliative care intervention. This has been specifically instituted in the intensive care unit as part of a nursing protocol. As seen in our sepsis patients, early intervention with palliative care, when appropriate, could have a major financial impact in the intensive care unit. This will likely be a subject for further study at our hospital.

There are some limitations to the analysis of this study. The economic benefits of early consultation were demonstrated for patients with similar CMIs. However, this measurement is dependent on external factors such as coding variability. Also, this analysis cannot account for all of the potential endogenous differences in either patient preference or physician preference that could lead to whether or not an early or late consult was requested and thus influence the results of the study. The use of covariate analysis of the variables within the patient population should help mitigate some of concerns related to patient heterogeneity. Physician referral preferences did not seem to be influenced by the patient's prior functional status. Their concern was usually focused on the medical problem at hand.

This study corroborated many earlier studies highlighting that the implementation of a palliative care program can potentially realize financial gains for the hospital. As demonstrated in our hospital, this does not have to be done at great expense. Perhaps, just as relevant were the benefits seen with patient care. The palliative program fostered greater alignment between patient and family goals and the final treatment plan, which was tempered by more realistic objectives. The economic benefits achieved were in a sense, an outcome of this change in focus and not necessarily the primary intent. By utilizing our nurse champions, hospitalists, social work, and case management, there has been greater acceptance of this approach hospital wide.

Conclusion

In general, intervention with palliative care has shown to provide benefits for hospitals in terms of cost savings and improvement in quality for patient care. This analysis demonstrated that when patients are involved with early palliative care, the positive effects of this intervention are amplified. In addition to the improvement in the financial bottom line for the hospital, patients who received early intervention were placed into a more appropriate level of care much earlier. There was no negative impact on other patient outcomes such as mortality. These gains were accomplished with minimal expense in a medium-sized community-based hospital. Similar results may be achieved in comparable settings.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Appendix 1

Palliative Medicine Consult Worksheet

Name _____________________ MRN: ___________ Age: _______ Date: ___________

PALLIATIVE MEDICINE FOLLOWING? YES / NO (please circle one)

If Palliative Medicine is following, do not complete this form & return form to collection box.

Please check ONLY the criteria that apply to your patient.

Appendix 2.

Palliative Performance Scale (PPSv2) Version 2

PPS level	Ambulation	Activity and evidence of disease	Self-care	Intake	Conscious level
100%	Full	Normal activity and work No evidence of disease	Full	Normal	Full
90%	Full	Normal activity and work Some evidence of disease	Full	Normal	Full
80%	Full	Normal activity with effort Some evidence of disease	Full	Normal or reduced	Full
70%	Reduced	Unable normal job/work Significant disease	Full	Normal or reduced	Full
60%	Reduced	Unable hobby/house work Significant disease	Occasional assistance necessary	Normal or reduced	Full or confusion
50%	Mainly sit/lie	Unable to do any work Extensive disease	Considerable assistance required	Normal or reduced	Full or confusion
40%	Mainly in bed	Unable to do most activity Extensive disease	Mainly assistance	Normal or reduced	Full or drowsy±confusion
30%	Totally bed bound	Unable to do any activity Extensive disease	Total care	Normal or reduced	Full or drowsy±confusion
20%	Totally bed bound	Unable to do any activity Extensive disease	Total care	Minimal to sips	Full or drowsy±confusion
10%	Totally Bed Bound	Unable to do any activity Extensive disease	Total care	Mouth care only	Drowsy or coma±confusion
0%	Death	—	—	—	—

References

Dumanovski

, Augustin

, Rogers

, et al.: The growth of palliative care in US hospitals: A status report. J Palliat Med, 2016; 19:8–15.

May

, Normand

, Morrison

: Economic impact of hospital inpatient palliative care consultation: Review of current evidence and directions for future research. J Palliat Care Med, 2014; 17:1593–1598.

Smith

, Cassel

: Cost and non-clinical outcomes of palliative care. J Pain Symptom Manage, 2009; 38:32–44.

Cardona-Morrell

, Kim

JCH

, Turner

, et al.: Non-beneficial treatments in the hospital at the end of life: A systematic review on extent of the problem. Int J Qual Health Care, 2016; 28:456–469.

Seaman

, Barnato

, Sereika

, et al.: Patterns of palliative care service consultation in a sample of critically ill ICU patients at high risk of dying. Heart Lung, 2017; 46:18–23.

Puntillo

, Arai

, Cohen

, et al.: Symptoms experienced by intensive care unit patient at high risk of dying. Crit Care Med, 2010; 38:2155–2160.

Penrod

, Deb

, Dellenbaugh

, et al.: Hospital-based palliative care consultation: Effects on the hospital cost. J Palliat Med, 2010; 13:973–979.

Morrison

, Dietrich

, Ladwig

, et al.: Palliative care consultation team cut hospital cost for Medicaid beneficiaries. Health Aff (Millwood), 2011; 30:454–463.

Morrison

, Penrod

, Cassel

, et al.: Cost savings associated with US hospital palliative care consultation programs. Arch Intern Med, 2008; 168:1783–1790.

10.

Smith

, Coyne

, Cassel

, et al.: A high volume specialist palliative care unit and team may reduce inpatient end-of-life care costs. J Palliat Care Med, 2003; 6:699–705.

11.

Ciemins

, Blum

, Nunley

, et al.: The economic and clinical impact of an inpatient palliative care consultation service: A multifaceted approach. J Palliat Med, 2007; 10:1347–1355.

12.

Penrod

, Deb

, Luhrs

, et al.: Cost and utilization outcomes of patients receiving hospital-based palliative care consultation. J Palliat Care Med, 2006; 9:855–860.

13.

Whitford

, Shah

, Moriarty

, et al.: Impact of a palliative care consult service. Am J Hosp Palliat Care, 2014; 31:175–182.

14.

Albanese

, Radwany

, Mason

, Gayomali

: Assessing the financial impact of an inpatient acute palliative care unit in a tertiary care teaching hospital. J Palliat Med, 2013; 16:289–294.

15.

Teno

, Gozalo

, Bynum

, et al.: Change in end-of-life care for Medicare beneficiaries: Site of death, place of care, and healthcare transitions in 2000, 2005, and 2009. JAMA, 2013; 309:470–477.

16.

Gade

, Venohr

, Conner

, et al.: Impact of an inpatient palliative care team: A randomized control trial. J Palliat Med, 2008; 11:180–190.

17.

May

, Garrido

, Cassel

, et al.: Prospective cohort study of hospital palliative care teams for inpatient with advanced cancer: Earlier consultation is associated with larger cost-saving effect. J Clin Oncol, 2015; 33:2745–2752.

18.

Anderson

, Downing

, Hill

, et al.: Palliative performance scale: A new tool 1996. J Palliat Care, 1996; 12:5–11.

19.

Palliative Performance Scale (PPSv2). Medical Care of the Dying, 4th ed; p. 120. Copyright Victoria hospice Society, 2006. WW.Victoriahospice.org/Ed publications.HTML (Last assessed January 8, 2008).

20.

Amano

, Morita

, Tatara

, et al.: Association between early palliative care referrals, inpatient hospice utilization, and aggressiveness of care at the end of life. J Palliat Med, 2015; 18:270–273.

21.

Reyes-Ortiz

, Williams

, Westphal

: Comparison of early versus late palliative care consultation in end-of-life care for the hospitalized frail elderly patients. Am J Palliat Care, 2015; 32:516–520.