Assessing the Financial Impact of an Inpatient Acute Palliative Care Unit in a Tertiary Care Teaching Hospital

Abstract

Background:

Evidence suggests palliative care consult services yield cost improvements; few studies have examined the impact of an inpatient palliative care unit on hospital costs.

Objective:

This study estimates the cost avoidance of a single hospital's acute palliative care unit (APCU), building upon previous studies (1) by limiting pre-APCU costs to two days pre-APCU transfer, thereby minimizing bias from higher-cost first days of admission, and (2) by not limiting the study to cancer patients or patients who died, thereby presenting more comprehensive APCU costs.

Design:

This retrospective study compares direct costs of care on an APCU with costs pre-APCU transfer from general medical units, intensive care units (ICU), and the emergency department (ED), and compares the direct costs of APCU patients with those of control patients. The data were entered into an SPSS^® 17.0 (SPSS Inc., Chicago, IL) statistical software database. Paired and independent samples t-tests were conducted to test cost differences.

Setting/Subjects:

Study patients were admitted or transferred to the APCU from October 2008 through January 2009. Control patients were inpatients during the same time period and met several matching criteria. Measurements: The hospital's finance department provided direct costs, case mix index (CMI), and geometric mean length of stay, and the Department of Quality and Resource Management provided patients' demographic and administrative data. Results: Of 209 patients transferred to the APCU, 50% transferred from a medical unit, 32% from an ICU, and 18% from the ED. Annualized, the total cost avoidance realized by transfers to the APCU was $848,556, over half of which came from ICU to APCU transfers.

Conclusions:

Cost avoidance is realized when patients transfer to an APCU even when conservative pre-APCU cost measures are used and when patients with varying diagnoses and discharge outcomes are included. This study demonstrates a replicable methodology for estimating the financial impact of an APCU.

Introduction

Many U.S. hospitals have instituted palliative care programs to better meet the interdisciplinary needs of patients with advanced or terminal illness. The percentage of hospitals with 50 or more beds having palliative care (PC) programs increased by 138.3%, from 24.5% in 2000 to 63.0% in 2009.¹ This growth in PC programs has been accompanied by studies of their impact on quality of care and hospital costs, with mounting evidence that inpatient palliative care consult services (PCCS) yield quality and cost improvements.^2–6 Few studies have analyzed the financial impact of inpatient PC units,^7–10 only two of which measured the direct costs of patient care on a dedicated PC unit.^9–10 Jung and colleagues⁷ retrospectively compared hospital charges for cancer patients six months prior to death on the PC unit with those for patients who died on all other units. Elsayem and colleagues⁸ compared reimbursement rates and mean daily charges on a PC unit in a cancer center with those of all inpatient discharges. While both studies found the PC unit to be financially viable, their results may not be applicable to many U.S. hospitals that are financed primarily by Medicare Diagnostic Related Group (DRG) payments.^7,8 Further, direct cost measures are more widely used in studies of PC programs, since the association between charges and reimbursements is inconsistent regardless of actual patient care. Smith and colleagues⁹ and White and colleagues¹⁰ demonstrated a cost avoidance model for measuring the direct cost outcomes of providing end-of-life care on an inpatient PC unit in the first six months⁹ and then over the first four years of its operation.¹⁰ They compared costs before and after transfer to a PC unit and the costs for PC patients with those of medical-surgical control patients matched by age and diagnosis. Results showed lower costs and charges after transfer to the PC unit and compared to those of control patients.^9,10 Additionally they found lower overall costs for patients admitted directly to the PC unit from the emergency department (ED).¹⁰

This study employs similar methodologies to those of Smith and colleagues⁹ and White and colleagues¹⁰ estimating the cost avoidance of a single hospital's 12-bed acute palliative care unit (APCU) in a setting that includes an integrated system consisting of a PCCS, an outpatient PC clinic, and a home hospice program. Cost avoidance is estimated by comparing four months of APCU patient direct costs with control patients on other units and comparing direct costs pre- and post-transfer to the APCU. Unlike previous studies, pre- and post-APCU comparisons are limited to the two days pre-APCU transfer; this minimizes potential bias from higher costs of diagnostics and differing goals of care during the first days of admission. Further, this study is not limited to patients with cancer or deceased patients, providing a broader perspective of APCU costs. Results will provide additional evidence of the financial impact of an APCU using more conservative measures of cost and data inclusive of patients discharged alive.

Setting

The APCU at Summa Health System's Akron City Hospital opened in 2006 within an integrated system of hospice and palliative care in a 545-bed, tertiary care teaching hospital. The rapid growth of the hospital's home hospice program (established in 1999) and PCCS (established in 2002, now averaging over 140 consults per month) provided justification for the dedicated APCU.¹¹ The PCCS responds within 24 hours to requests from physicians and nursing unit patient care coordinators. Continued growth supported the establishment of an outpatient PC clinic and a nursing home PCCS in 2008. This spectrum of PC services operates with a coordinated, interdisciplinary team of PC providers in medicine, nursing, social work, pharmacy, pastoral care, nutrition/dietetics, and psychology. The APCU is a 12-bed unit designed to meet the needs of PC and hospice patients and families. A robust three-to-one patient to nurse staffing ratio is maintained on this unit. Rooms are designed to accommodate families for overnight stays; a common kitchen area provides simple meals and snacks. Patients may be admitted to the APCU from home or through the ED, may transfer to the APCU from a medical/surgical or critical care bed, or may transfer and enroll as a new hospice patient. All patients have a PCCS consult prior to admission or transfer to the APCU. A protocol is followed to prioritize patient need for an APCU bed when demand exceeds capacity. Mean daily occupancy rate on the APCU is 89%.

Methods

The sample

This is a retrospective study comparing direct costs of care on an APCU, general medical units, and intensive care units (ICUs). Three separate cost comparisons were conducted. First, patients' direct costs were compared pre- and post-transfer to the APCU. Second, direct costs of patients who were admitted to the APCU from the ED were compared with those of patients prior to APCU transfer from other units. Third, direct costs of patients who died on the APCU were compared with a control group of patients who died on other units.

Patients included in the study were admitted or transferred to the APCU between October 1, 2008 and January 31, 2009 and were not hospice patients. Hospice patients were excluded, because most enroll in hospice (under a new funding source) after transfer to the APCU. The only nonhospice patients excluded from the study were those directly admitted to the APCU from home; typically there are too few of them to be analyzed as a separate group. A sample of control patients were obtained from an administrative database. Control patients had to meet the following criteria to match the characteristics of deceased APCU patients: inpatients between October 1, 2008 and January 31, 2009; not enrolled in hospice; deceased at discharge; at least 55 years old; a 3M^™ All Patient Refined Diagnostic Related Group (APR DRG) severity score of at least three; and hospital length of stay of at least two days for cost comparisons and no more than 20 days to minimize the impact of outliers.

Data

The hospital's Department of Quality and Resource Management provided administrative data including primary diagnosis, age, gender, race, Medicare status, length of stay (LOS) on specific units and in total, and APR DRG severity score.¹² The PCCS provided dates of PC consults. Daily and total direct costs (including those for pharmacy, respiratory care, and unit costs) were provided by the finance department. Direct costs include resources used directly on or for the patient, such as patient care supplies, salaries, and drugs, and include the costs of operating the hospital departments associated with a patient's treatment, such as fixed labor costs, depreciation, benefits, support costs, and direct overhead. Hospital overhead (or indirect cost) was excluded from the study; as a fixed percentage of direct costs it is not controllable by nursing unit or providers. The finance department also provided patients' case mix index (CMI) and Geometric Mean Length of Stay (GM-LOS). Both CMI and GM-LOS are based on Medicare Severity Diagnostic Related Group (MS-DRG). The CMI is a weighted DRG that can be used as an indirect measure of severity and resource consumption.¹³ The GM-LOS is a marker for expected length of stay beyond which patient costs may exceed DRG payment.

Analysis

The data were entered into an SPSS^® 17.0 statistical software database for descriptive and comparative analyses.¹⁴ APCU patient data were analyzed in groups based on location pre-transfer or admission to the APCU: admissions from the ED, transfers from a medical unit, transfers from an ICU. Paired samples t-tests were conducted on mean daily direct costs of the two days prior to transfer and mean daily direct costs of the entire APCU LOS. Independent samples t-tests were conducted on the mean direct costs of patients admitted to the APCU from the ED and the mean direct costs of patients before transfer to the APCU. Pearson correlations were conducted to identify associations between delays in PC consult requests and costs and between delays in transfer to the APCU and costs. The criterion for statistical significance was set a priori at p≤0.01 for all statistical analyses.

Control patient data were analyzed in groups based on their unit at death (medical or ICU). Independent samples t-tests were conducted to assess comparability of control patients who had a PC consult with those who did not on age, APR DRG severity score, CMI, GM LOS, hospital LOS, and direct costs. Subsequent independent samples t-tests were conducted to assess comparability of control and study patients on age, APR DRG severity score, CMI, GM LOS, and hospital length of stay. Independent samples t-tests were conducted to compare mean daily APCU direct costs with control patients' mean daily direct costs of the last two days of admission.

Cost avoidance realized by transferring patients to the APCU was calculated based on the following straightforward formula: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} &\hbox{cost avoidance} = (\hbox{pre-transfer unit last two-day}\\& \quad\hbox{mean daily direct cost} - \hbox{APCU mean daily direct cost}) \times \\& \quad\hbox{number of patients transferred} \times \hbox{mean LOS on the APCU} \end{align*} \end{document}

This formula was used to estimate the cost avoidance realized by admitting patients directly to the APCU from the ED. Two calculations were performed based on the theoretical assumption that ED to APCU patients would have been transferred to either a medical unit or an ICU in the same proportions as study patients who transferred to the APCU (61% and 39%, respectively). Estimates of cost avoidance are multiplied by three to annualize the four months of data included in the study.

Results

Transfers to the APCU

During the study period, a total of 209 nonhospice patients transferred to the APCU, 50% from a medical unit, 32% from an ICU, and 18% from the ED. An additional 84 hospice patients excluded from the study were admitted to the APCU during the study period. Table 1 compares patients' demographic, administrative, and direct cost data by unit of transfer to the APCU. ICU transfer patients had higher mean direct costs, CMI, and APR DRG severity score than medical unit transfer patients. ED to APCU patients had a significantly lower mean hospital LOS, CMI, APR DRG severity score, and GM-LOS than medical unit and ICU transfer patients (p≤0.001 for each unit comparison). ED to APCU patients had a higher mean daily APCU direct cost than patients who transferred to the APCU from medical units (p≤0.004) and ICUs (p≤0.01). However, ED to APCU patients' mean APCU LOS and mean total APCU direct costs were not significantly different from those of patients transferred from medical units or ICUs (p≥0.01). LOS from admission to PC consult was positively associated with mean daily hospital costs over the entire hospital stay for APCU patients who transferred from a medical unit (p≤0.001). LOS between PC consult and transfer to the APCU was positively associated with mean daily APCU costs for patients who transferred from a medical unit (p≤0.001).

Table 1.

Characteristics and Outcomes of APCU Patients by Unit of Transfer: Medical, ICU, and the ED

Patient demographics and hospital admission and cost outcomes	Med. unit transfers to the APCU N=104	ICU transfers to the APCU N=67	ED transfers to the APCU N=38
Mean age	77	75	74
% male	45	52	50
% white	83	82	90
% Medicare	79	87	82
Most common diagnoses, %	Cancer 19	Septicemia 22	Cancer 24
	Resp. dis. 14	Resp. dis. 14	Resp. dis. 16
	Septicemia 9	ICH/CVA 7	Septicemia 13
	ICH/CVA 9	Heart failure 4	Heart failure 8
	Heart failure 8	Cancer 4	Renal failure 8
Discharge outcome
Deceased, %	30	55	45
Hospice, %	58	36	34
Total LOS mean (range)	9.8 (2–30)	9.9 (2–30)	2.8 (1–8)
APCU LOS mean (range)	3.0 (1–15)	2.5 (1–12)	2.8 (1–7)
^aLOS prior to PCCS consult mean (range)	5.1 (0–27)	5.8 (0–28)	0.13 (0–1)
^bLOS PCCS to APCU transfer mean (range)	1.2 (0–17)	1.5 (0–14)
GM-LOS (expected)	5.7	7.1	4.3
Mean CMI	1.8740	3.1495	1.2710
Mean APR DRG severity scores (1–4)	3.3	3.7	2.7
Mean total hospital LOS direct costs	$10,054	$16,720	$2387
Mean two-day pre-APCU direct costs	$904	$1735	----
Mean total APCU LOS direct costs	$2207	$1787	$2151
Mean APCU direct costs per day	$691	$701	^c$826

APCU, acute palliative care unit; APR DRG, All Patient Refined Diagnostic Related Group; CMI, case mix index; ED, emergency department; GM LOS, Geometric Mean Length of Stay; ICH/CVA, intracranial hemorrhage/cerebrovascular accident; ICU, intensive care unit; PCCS, palliative care consult services; resp dis, respiratory disease.

Significantly correlated with mean daily hospital costs over the entire LOS for medical unit transfer patients only, p≤0.001.

Significantly correlated with mean daily costs over the APCU LOS for medical unit transfer patients only, p≤0.001.

Significantly less than pre-APCU, ICU mean daily costs ($1735) p≤0.001; not significantly less than pre-APCU, medical unit mean daily costs ($904) p≤0.162.

The mean direct cost reduction from medical unit transfers to the APCU was $213 per day and for ICU transfers, $1034 per day; both were significant, as noted in Table 2. Pharmacy, respiratory care, and unit direct costs were the most commonly incurred costs across all patients. Mean daily pharmacy, respiratory, and unit direct costs prior to APCU transfer were significantly higher for ICU patients but not for medical unit patients.

Table 2.

Medical Unit and ICU Patient Costs Compared with Post-Transfer APCU Costs

	Mean medical unit costs two days pre-APCU compared with mean APCU daily costs N=104			Mean ICU costs two days pre-APCU compared with mean APCU daily costs N=67
	Pre-APCU on med unit	APCU^a	Paired samples t-test significance	Pre-APCU in ICU	APCU^a	Paired samples t-test significance
Pharmacy costs	$142	$87	n.s	$205	$116	p≤0.001
Respiratory costs	$97 (N=58)	$77 (N=56)	n.s.	$145 (N=55)	$ 71 (N=53)	p≤0.001
Unit costs	$506	$509	n.s.	$1088	$492	p≤0.001
Mean daily focused and direct costs	$904	$691	p≤0.001	$1735	$701	p≤0.001

APCU, acute palliative care unit; ICU, intensive care unit; n.s., not significant, p≥0.01.

Based on the mean of the entire APCU length of stay.

The calculations estimating the total cost avoidance realized by transferring patients to the APCU are presented in Table 3. Estimates for medical unit and ICU transfer patients are based on the differences between pre-APCU mean daily direct costs and APCU mean daily direct costs. Estimates for ED patients are based on the differences between their APCU mean daily direct costs and transfer patients' pre-APCU mean daily direct costs proportionate to their location (ICU or medical unit). Cost avoidance is calculated by multiplying these differences by the number of patients and their mean APCU LOS for each transfer group. Total cost avoidance realized by ED to APCU admissions and inpatient transfers to the APCU was $282,852; annualized, the total estimated cost avoidance was $848,556.

Table 3.

Estimated Savings Based on Admissions and Transfers to the APCU

Patient admissions and transfers to the APCU 10/1/08 to 1/31/09 N=209	Cost avoidance=mean daily cost difference×# patients×mean APCU LOS	Cost avoidance annualized
Medical unit transfers	$ 213×104×3.0=$ 66,456	$199,368
ICU transfers	$1034×67×2.5=$173,195	$519,585
Admits from the ED
(medical unit averted)	$ 78×23×2.8=$ 5023	$15,069
(ICU averted)	$ 909×15×2.8=$ 38,178	$114,534
Total	$282,852	$848,556

APCU, acute palliative care unit; ED, emergency department; ICU, intensive care unit; LOS, Length of Stay.

Patients who transfer to and expire on the APCU compared with control groups

Independent samples t-tests revealed no significant differences (p≥0.01) between medical unit control patients who had a PC consult and those who did not. ICU control patients who did not have a PC consult had a significantly higher MS-DRG LOS compared to those who did (9.5 and 6.5 respectively; p≤0.007). Control patients with and without a PC consult did not differ significantly from like unit APCU transfer patients on any comparability parameters: age, hospital LOS, MS-DRG LOS, APR-DRG severity score, and CMI (p≥0.01). Table 4 provides comparative results of medical and ICU transfer patients' end-of-life care direct costs on the APCU with those of like unit control patients. Mean daily direct costs of APCU patients who transferred from a medical unit were lower but not significantly different from those of medical unit control patients ($683 versus $793, respectively). Mean daily direct costs of APCU patients who transferred from an ICU were significantly lower than those of ICU control patients ($669 versus $2135, respectively).

Table 4.

Characteristics and Outcomes of Deceased APCU and Control Patients by Unit of Care

Deceased APCU and control patients	Transfer to APCU from med unit N=31	Med unit control patients N=19	Transfer to APCU from ICU N=37	ICU control patients N=36
Mean age	76	83	76	77
% Male	52	47	54	64
% White	88	95	84	94
% Medicare	73	90	84	97
Most common diagnoses, %	Heart fail 12	Septicemia 21	Septicemia 22	Septicemia 22
	Septicemia 9	Cancer 16	ICH 11	Pneumonia 11
	Lung ca 9	CVA 11	Heart fail 8	Resp fail 11
	CVA 9	Heart fail 11	Cardiac arrest 5	ICH 11
Hospital LOS, mean (range)	8.1 (2–26)	8.1 (3–18)	8.8 (2–30)	10.7 (3–20)
APCU LOS, mean (range)	1.9 (1–8)	-	1.6 (1–3)	-
GM LOS	5.4	5.7	7.6	8.6
Mean APR DRG severity score (1–4)	3.4	3.7	3.8	3.9
Mean CMI	1.9225	1.9983	3.5241	3.7576
Mean total hospital direct costs	$9669	$8767	$17,112	$24,629
^aControl patients' mean daily direct costs		^b$793		^c$2315
^dAPCU patients' mean daily direct costs	^b$683		^c$669

APCU, acute palliative care unit; APR DRG, All Patient Refined Diagnostic Related Group; ca, cancer; CMI, case mix index; CVA, cerebrovascular accident; GM LOS, Geometric Mean Length of Stay; heart fail, heart failure; ICH, intracranial hemmorhage; ICU, intensive care unit; resp fail, respiratory failure.

Mean daily cost based on the last two days of admission.

t-test for independent samples was not significant, p≤0.164.

t-test for independent samples was significant, p≤0.001.

Mean daily cost based on the entire APCU length of stay.

Discussion

Since the APCU opened, efforts to assess its financial impact have been a work in progress. With each additional study, new methodologies were entertained and elaborated upon. Building upon these, the present study examined the impact of admissions and transfers to the APCU on direct costs, estimating a cost avoidance of $848,556 annually. Results provide evidence of the economic importance of timely identification of patients desiring and appropriate for PC. These results have been used in the justification of maintaining the robust nurse to patient ratio on the APCU and increasing the PC presence in the ED and outpatient cancer center.

The APCU is part of an integrated program of palliative and hospice services that supports continuity of care and the ability to monitor costs across settings. The methodologies and other results emerging from this and previous studies are laying the groundwork for more extensive study of the impact of this continuum of services on health care system costs.

Limitations

The APCU is part of a larger and growing system of palliative and hospice care; and the resources needed to collect and merge data from three data sources are limited. As a result, the decision was made to compare direct costs differences within shorter periods of time over several years to track long-term trends, recognizing that these estimates of cost avoidance may contain some seasonal trend bias.

Measuring direct costs before and after transfer to an APCU and calculating differences to estimate the extent of cost avoidance is practical but not without bias. Pre-APCU direct costs may be inflated by necessary tests and treatments, reflecting differing goals of care. Alternatively, pre-APCU direct costs may be deflated by the implementation of PCCS recommendations, reflecting changing goals of care in the days prior to transfer.⁷ In this study, length of PCCS involvement prior to APCU transfer was not associated with pre-APCU costs, though the study may not have been powered to demonstrate differences. APCU direct costs may be inflated by carryover costs from the previous unit when nonbeneficial lab or medication orders are not discontinued prior to transfer. To minimize the impact of these potential sources of bias, we calculated the difference between the mean daily direct cost of the last two days pre-APCU and the mean daily direct cost of the entire APCU stay. The latter was also used to minimize biases in costs over a relatively short APCU mean length of stay (three days or less for nonhospice patients). In this study, length of PC involvement on a medical unit prior to APCU transfer was positively associated with mean daily costs on the APCU. Further study is needed to explore causes of the cost impact of delay in transfer to the APCU.

Estimating cost avoidance realized by admitting patients to the APCU from the ED is limited by lack of certainty about which type of unit the patient would have been admitted to if the APCU had not been available. The percentage of patients used to reflect medical unit and ICU avoidance from the ED were based on the percentage of inpatients that transferred to the APCU. Estimates of cost avoidance from ED to APCU admissions are theoretical and may not reflect actual cost savings. Nonetheless, involved PC clinicians believed that a greater proportion of ED to APCU patients would have been admitted to an ICU in spite of a lower CMI. More extensive study is needed to clarify ED to medical unit, ICU, and APCU admission decision making and subsequent cost and clinical outcomes.

Identifying a control group for comparison to APCU patients is challenging even in a large, 545-bed hospital. While a group of control patients were identified with similar demographic, severity, hospital stay, and discharge outcome markers, uncertainty about their appropriateness for PC or as a valid control group remain.

Finally, this study does not account for additional revenue generated when inpatients change to a hospice payment source. The 84 hospice patients admitted to the APCU had an average LOS on the APCU of 4.6 days. At our current hospice rate, revenue from these patients was $191,975; annualized inpatient hospice enrollments generate approximately $575,925.

Conclusions

PCUs reduce costs through their impact on resource utilization consistent with patient goals of care.^15,16 Costs are avoided by the elimination of nonbeneficial tests, pharmaceuticals, treatments, and transfers from higher-cost ICUs.¹⁷ Findings of this study are consistent with those of previous studies.^7,9,16 However, in this study, patients with varying diagnoses and discharge outcomes were included, and pre-APCU direct cost comparisons were limited to two days prior to transfer; and still, significant cost reductions were found. The results demonstrate a replicable methodology for estimating the financial impact of a PC unit. Hospitals with a PCCS and in the planning stages of establishing a PC unit might consider the results when determining the bed capacity that will maximize benefits to PCCS patients and to hospital costs.

Footnotes

Acknowledgments and Author Disclosure Statement

This study was funded by the Summa Health Foundation through a grant to Summa's Health Services Research and Education Institute and by the Department of Internal Medicine, Summa Health System. The authors have no competing financial interests to disclose.

References

Center to Advance Palliative Care: Palliative Care in Hospitals Continues Rapid Growth for 10th Straight Year, According to Latest Analysis. 2011. www.capc.org/news-and-events/releases/07-14-11. 2011 August 4.

O'Mahony

, Blank

, Zallman

, Selwyn

. The benefits of a hospital-based inpatient palliative care consultation service: Preliminary outcome data. J Palliat Med, 2005; 8,5:1033–1039.

Penrod

, Deb

, Luhrs

, Dellenbaugh

, Zhu

, Hochman

et al. Cost and utilization outcomes of patients receiving hospital-based palliative care consultation. J Palliat Med, 2006; 9,4:855–860.

Gade

, Venohr

, Conner

, McGrady

, Beane

, Richardson

et al. The impact of an inpatient palliative care team: A randomized control trial. J of Palliat Med, 2008; 11,2:180–190.

Hanson

, Usher

, Spragens

, Bernard

. Clinical and economic impact of palliative care consultation. J Pain Symptom Manage, 2008; 35,4:340–346.

Morrison

, Penrod

, Cassel

, Caust-Ellenbogan

, Litke

, Spragens

et al. Cost savings associated with US hospital palliative care consultation programs. Arch Intern Med, 2008; 168,16:1783–1790.

Support Care Cancer Online First™. Jung

, Kim

, Heo

, Baek

. Health economics of a palliative care unit for terminal cancer patients: A retrospective cohort study. 2010. www.springerlink.com/content/0941-4355/preprint/?sort=p_OnlineDate&sortorder=desc&o=170. 2011 February 10.

Elsayem

, Swint

, Fisch

, Palmer

, Reddy

, Walker

et al. Palliative care inpatient service in a comprehensive cancer center: Clinical and financial outcomes. J Clin Oncol, 2004; 22,10:2008–2014.

Smith

, Coyne

, Cassel

, Penberthy

, Hopson

, Hager

. A high-volume specialist palliative care unit and team may reduce in-hospital end-of-life care costs. J of Palliat Med, 2003; 6,5:699–705.

10.

White

, Stover

, Cassel

, Smith

. Nonclinical outcomes of hospital-based palliative care. J Healthcare Manage, 2006; 51,4:260–274.

11.

Radwany

, Mason

, Clarke

, Clough

, Sims

, Albanese

. Optimizing the success of a palliative care consult service: How to average over 110 consults per month. J Pain Symptom Manage, 2009; 37,5:873–883.

12.

Agency for Healthcare Research and Quality. Rockville, MD: Hughes J: Development of the 3M™ All Patient Refined Diagnostic Related Groups (APR DRGs). Mortality Measurement, 2009. www.ahrq.gov/qual/mortality/Hughes.htm. 2010 November 30.

13.

HCPro. What does case-mix index mean to you? 2010. www.hcpro.com/HOM-250674-5728/What-does-casemix-index-mean-to-you.html. 2010 November 30.

14.

Statistical Package for the Social Sciences. SPSS 17.0: Chicago, IL, 2003.

15.

Smith

, Cassel

. Cost and non-clinical outcomes of palliative care. J Pain Symptom Manage, 2009; 38,1:32–43.

16.

Edens

, Harvey

, Gilden

. Developing and financing a palliative care program. Am J Hosp Palliat Med, 2008; 25,5:379–384.

17.

Davis

, Walsh

, Nelson

, Konrad

, LeGrand

. The business of palliative medicine: Management metrics for an acute-care inpatient unit. Am J Hosp Palliat Care, 2001; 18,1:26–29.