Improving Access to Palliative Care through an Innovative Quality Improvement Initiative: An Opportunity for Pay-for-Performance

Abstract

Background:

Improving access to palliative care is an important priority for hospitals as they strive to provide the best care and quality of life for their patients. Even in hospitals with longstanding palliative care programs, only a small proportion of patients with life-threatening illnesses receive palliative care services. Our two well-established palliative care programs in large academic hospitals used an innovative quality improvement initiative to broaden access to palliative care services, particularly to noncancer patients.

Methods:

The initiative utilized a combination of electronic and manual screening of medical records as well as intensive outreach efforts to identify two cohorts of patients with life-threatening illnesses who, according to University HealthSystems Consortium (UHC) benchmarking criteria, would likely benefit from palliative care consultation. Given the differing cultures and structure of the two institutions, each service developed a unique protocol for identifying and consulting on suitable patients.

Results:

Consultation rates in the target populations tripled following the initiative: from 16% to 46% at one hospital and from 15% to 48% at the other. Although two different screening and identification processes were developed, both successfully increased palliative care consultations in the target cohorts.

Conclusion:

Quality improvement strategies that incorporate pay-for-performance incentives can be used effectively to expand palliative care services to underserved populations.

Introduction

Palliative care is increasingly recognized as an evidence-based intervention that improves outcomes such as pain control,¹ quality of life,^2,3 patient and family satisfaction,^4,5 communication,⁵ and bereavement morbidity.⁶ In addition, palliative care is associated with more efficient resource use and therefore decreased medical costs.^7–10 Recent studies suggest that use of palliative care may even prolong life.^2,3

Despite increasing efforts to improve access to palliative care services, particularly to those with noncancer diseases,^11,12 significant numbers of patients still have unmet end-of-life medical needs.^13–16 Many patients who would benefit from palliative care are not referred or are referred too late to take full advantage of such services.^17–19

In 2007, the University HealthSystem Consortium (UHC), an association of 110 academic health centers, conducted the Palliative Care Benchmarking Project, a retrospective analysis of 44 hospitals measuring the percentage of inpatients seen by palliative care that the UHC considered “high risk,”²⁰ as defined below. The UHC audit showed that only 16% of patients at Massachusetts General Hospital (MGH) and 15% of patients at Brigham and Women's Hospital (BWH) who fell into the UHC-defined high-risk cohorts received a palliative care consultation—figures considerably lower than some institutions with similarly established palliative care services. Both hospitals also fell below the benchmark published in a 2008 consensus statement by the Center to Advance Palliative Care, identifying a target goal of 2.0 new referrals per staffed hospital bed per year for established programs in existence for 5 years or more²¹: in 2007, MGH saw 1.2 referrals per staffed hospital bed per year and BWH saw 1.0 referrals per staffed hospital bed per year. Consequently, palliative care leaders and hospital administration set a goal of improving access to palliative care services.

Partners HealthCare System, which includes MGH and BWH, joined with an insurer to select improved palliative care metrics as pay-for-performance quality improvement measures for the period 2007–2009. One of the performance measures chosen by MGH and BWH was the percent of patients seen by the hospital's palliative care service who, in retrospect, were identified as falling into a UHC high-risk cohort. Partially supported by an internal grant, the MGH and BWH palliative care programs collaborated with each other and their respective quality improvement offices to develop screening procedures that identified patients suitable for consultation.

Methods

Table 1 outlines the characteristics of the two hospitals, both of which are large urban academic medical centers. The Partners Institutional Review Board approved the study. The target population was patients who met the UHC criteria for high-risk patients, namely, inpatients over the age of 18 with a length of stay of 4 days or longer who either met: (1) acute cohort criteria—patients who had not been admitted to the hospital within the previous 12 months but died during the target hospitalization or (2) chronic cohort criteria—patients with 2 admissions in the previous 12 months for whom those admissions and the target admission were assigned 1 of 26 UHC-designated, high-risk diagnosis-related groups (DRGs),a regardless of whether the patient survived or died on the target admission. Although the insurer provided funds for the study, the patient population was not limited to the insurer's patients and could include patients with other or no insurance.

Table 1.

Hospital Characteristics (2009 Data)

	MGH	BWH
Total beds	907	775
CAPC beds^a	771	657
Annual admissions	47,756	47,707
Annual deaths	1211	1025

Excludes pediatrics, obstetrics, and psychiatry beds, as per CAPC guidelines

CAPC, Center to Advance Palliative Care; MGH, Massachusetts General Hospital; BWH, Brigham and Women's Hospital.

MGH's goal was to increase the consultation rate in this population from 16% to 23% while BWH's target was from 15% to 25%. The insurer withheld funds representing a material amount of dollars that were returned to the hospital if targets were met. The palliative care teams did not benefit directly from returned funds. For the purposes of the pay-for-performance audit, the measurement period was October 1 to December 31, 2009, compared to October 1 to December 31, 2007.

Screening procedures

Figure 1 outlines the general screening and identification protocols. Due to the distinctive hospital structures and cultures at MGH and BWH, as well as the unique characteristics of the two palliative care services, similar but different staffing and screening strategies were used at the two sites. The main difference was the use of an registered nurse or physician at MGH and a social worker at BWH, and a heavier reliance on initial computerized screening at MGH. Table 2 shows the staffing and resources available at the two sites.

FIG. 1.

Initial screening protocol for high-risk patients. *To minimize the risk of patients being discharged before being seen, computer screening of patients was performed on day 2 of admission. **At MGH, computer-based screening flagged patients whose medical records indicated a high risk of dying as evidenced by presence of key words in free text medical records indicative of serious illness, e.g., “critically ill,” “congestive heart failure.”

Table 2.

Staffing and Resources

	MGH	BWH
P4P project staffing	0.4 FTE MD	0.3 FTE MD
	1.0 FTE RN	0.5 FTE LICSW
Administrative support	Center for Quality and Safety	Center for Clinical Excellence
Information technology resources	0.2 FTE System development consultant	IT consultants on “as needed” basis
	Queriable Patient Inference Dossier (“QPID”): a medical record search engine that can search both structured and free text documents	Report central: computer system designed to report quality measures for outpatient providers
Palliative Care daily consultation team staffing^a	3 MDs	2 MDs
	1–2 palliative care fellows	1–2 palliative care fellows
	2 nurse practitioners	1 social worker
	1 social worker

BWH staffing does not include Intensive Palliative Care Unit staffing.

MGH, Massachusetts General Hospital; BWH, Brigham and Women's Hospital; FTE, full-time Equivalent; LICSW, licensed clinical social worker.

Acute cohort

At MGH, patients were screened electronically by a medical record search engine, QPID.²² QPID searched the records for potentially eligible patients using the above-mentioned UHC criteria, and also narrowed the list by searching for patients whose electronic medical records contained keywords indicating that the patient was seriously ill (e.g. “congestive heart failure,” “multiorgan failure,” or “critically ill”). This list of words was regularly revised as part of the Plan-Do-Check-Act process discussed below. Using both the UHC criteria and the keyword search QPID produced a daily list of 100–200 potential acute patients a day, roughly 30 of whom were newly admitted. A palliative care registered nurse/physician would screen the electronic record of the new patients, and also review records of those patients still in the hospital on days 7, 10, 14, 21, and 31 of their admission. The paper records of those patients identified as high risk for dying were also reviewed. The study nurse also made regular visits to cardiac floors to provide ongoing informal palliative education to nursing staff as well as to further discuss potential study patients.

At BWH, the study team selectively targeted intensive care unit (ICU) patients and patients with long length of stay (>14 days); these patients were screened weekly. In addition, the study social worker or physician made daily visits to the ICU and cardiac floors to ask nursing staff about potential referrals. Finally, the study social worker developed a significant source of referrals from an informal network of social workers, nurses, chaplains, and care coordinators who identified high-risk patients.

Chronic cohort

Computer programs at both institutions identified patients admitted to the hospital twice within the preceding 12 months and who had been discharged with the UHC-designated high risk DRGs. The most common DRGs that triggered consults were: heart failure and shock; respiratory infections and inflammations with complications; and chronic obstructive pulmonary disease. Typically, the teams at both hospitals were alerted to 2–10 potential chronic patients per day. Informal referrals from floor staff resulting from an increased presence on wards, also contributed to additional potentially suitable patients.

Approval by attending

After a study team member reviewed the patient's electronic and paper records and deemed the patient likely to be suitable for a palliative care consultation, the inpatient attending was contacted and informed that the patient had been identified as a high-risk patient and potentially suitable for a referral. At MGH, attendings were contacted via email from a member of the study team. At BWH, the social worker would make contact either in person or via email. Patients were only seen if the attending agreed that a consultation would be appropriate. However, at BWH, medical intensive care unit and hospitalist leadership granted blanket approval for referrals.

Patient consultation

At MGH, all patients were seen by palliative care clinicians and a full consultation was performed; these consults were termed “Supportive Care” rather than palliative care, a term preferred by the program's advisory group.²³ At BWH, acute patients were seen by the inpatient team; chronic patients were seen by a social worker who completed the consult and referred to the inpatient palliative care team only if patients had uncontrolled symptoms. At BWH, the term “Supportive Care” consultation was used for cardiology patients only.

Plan-Do-Check-Act cycle

The intervention was a quality improvement project that employed a traditional Plan-Do-Check-Act cycle, an iterative four-step process that allows for continuous incremental changes each cycle.²⁴ An essential part of the planning process involved getting early input and strong support from the leadership of both hospitals. The chief medical officers and the quality and safety units at both hospitals were intimately involved with the project and invested in its success. Buy-in from the hospital administration was critical in gaining access to hospital departments, particularly those less familiar with or supportive of palliative care. Early planning for the project also included establishing clinical advisory boards, engaging clinicians likely to care for target patients, and establishing buy-in from the palliative care teams.

The quality improvement cycle of Plan-Do-Check-Act enabled teams to respond quickly to problems that arose. Although outcomes were only formally measured quarterly, both project teams initially met on a weekly basis to review the accuracy and reliability of the screening procedures, discuss progress, and identify problems and possible solutions, thus allowing for continual and timely improvement. Serial changes were made to the electronic screening tools at both hospitals. At MGH, for instance, where initial screening protocols for the acute cohort proved to be labor intensive with relatively low yield, the screening criteria were refined significantly to increase accuracy. Advice was sought from clinicians who had been contacted about referrals and different methods of contacting physicians were trialed in an effort to decrease the non-response rate. Feedback from members of the palliative care teams regarding the appropriateness of consults for the screened patients was also continually elicited.

As familiarity developed between the palliative care services and other hospital units, and as personal relationships strengthened, collaboration improved such that clinicians would proactively identify potential patients for the project. At MGH, the research team's strong relationships with medical intensive care units (MICU) and cardiology were particularly helpful in identifying high-risk patients, and there was a shift away from relying solely on the computer-generated lists. BWH began with intensive outreach to the MICU and cardiology as well as developing an informal network for identifying patients at high risk of death; these relationships and networks expanded over the course of the project.

The palliative care services also recognized that certain teams or referring physicians routinely refused a suggested consultation, and during the intervention, limited resources were directed to more fertile fields.

Results

At MGH, in the final quarter of 2009, the Palliative Care Service consulted on 46% of the patients in the target cohorts, compared to 16% in the final quarter of 2007. At BWH, 47% of the target patients were seen compared to 15% two years previously (Fig. 2 and Table 3). These rates were well above the pay-for-performance targets of 23% and 25% for MGH and BWH respectively. In addition to the increase in the target cohorts, overall consult volumes increased at both hospitals (Table 4).

FIG. 2.

High-risk patients seen by the Palliative Care Service.

Table 3.

Results: P4P Consultation Rates on High-Risk Patients

	MGH	BWH
Total number of high-risk patients identified by UHC in 2009^a	968	886
Number of high-risk patients seen by palliative care team	324	292
Identified via P4P screening and identification protocols	134	108
Referred to palliative care service via normal referral channels^b	190	184
Number of high-risk patients not seen by palliative care team^c	644	594
Number of patients additionally seen as part of P4P screening and identification protocols who did not ultimately meet target population group criteria	199	305

The insurer retrospectively provided the hospitals with a list of the patients who fell within the acute and chronic criteria at the end of each quarter.

Given BWH informal referral system, identifying which patients were seen through the QI effort and which came through regular referral channels was not always clear.

These patients were not seen either because they were never identified by the screening protocols, or they were identified but the treating team did not require a palliative care consult.

MGH, Massachusetts General Hospital; BWH, Brigham and Women's Hospital; UHC, University Health System Consortium.

Table 4.

Results: Change in Clinical Consult Volume

MGH
Palliative care consults	2005	2006	2007	2008	2009
Oncology	450	552	822	930	1,023
Nononcology	229	237	272	360	878
Total	679	789	1094	1290	1901
Consults/CAPC bed^a	0.9	1.0	1.2	1.7	2.5
BWH
Palliative care consults	2005	2006	2007	2008	2009
Oncology	413	618	596	596	1116
Nononcology	26	59	86	214	597
Total	429	677	682	810	1713
Consults/CAPC bed^a	0.7	1.0	1.0	1.2	2.6

Number of palliative care consults per staffed hospital bed (excludes pediatrics, obstetrics, and psychiatry beds, as per CAPC guidelines) per year.

MGH, Massachusetts General Hospital; CAPC, Center to Advance Palliative Care; BWH, Brigham and Women's Hospital.

At MGH, 51% of e-mail requests resulted in a consult. For those that did not result in a consult, 68% did not respond to the e-mail, while 32% wrote back and declined a consult. While the majority of the rejection e-mails did not provide a reason, common responses were that the attending did not feel that palliative care was appropriate at the time of request or that the patient was being discharged that day.

At BWH, the nonresponse/refusal rate was 23%. Two reasons may have contributed to BWH's lower rate: (1) more of the consults were requested in person by the social worker and (2) the medical ICU and hospitalist leadership had granted blanket permission for their patients to be seen by the service.

Discussion

In this initiative, a quality improvement effort that involved screening records and offering consults proved an effective strategy for expanding access to palliative care in a targeted population of high-risk patients. Consultations in this retrospectively defined, high-risk population tripled at both hospitals over a 3-year period, with overall referral rates increased significantly, most dramatically among nononcology patients.

The opportunity for this quality improvement project was provided by the participation in a pay-for-performance initiative and the choice of palliative care metrics as the key measure. Pay-for-performance programs that link financial payments with achievement of high quality care are increasingly being utilized in health care in an effort to improve patient outcomes as well as encourage appropriate resource utilization.^23,25–28

Factors contributing to increased consultation rates

Each of the stakeholders in the pay-for-performance initiative was motivated to participate. For the insurer, previous studies reporting palliative care consultations leading to improved care and cost savings drew attention. For the hospital administration, the most important motivator was “doing the right thing” and improving patient care for those suitable for palliative care, with additional hopes of reducing inappropriate resource utilization. Commitment from hospital leadership ensured sufficient financial backing, including resources for significant information technology support that was necessary to successfully carry out a project of this scope. Supported time from the developers of QPID allowed us to experiment with and create a user-friendly computer-based screening tool that facilitates identification of a variety of patient populations. For the referring clinicians, improved care for patients and potentially decreased workload for themselves were thought to be prime factors.

Benefits to the palliative care services

For both services, motivators included increased visibility and hospital-supported outreach as well as improving end-of-life care for patients. The project was responsible for a number of positive outcomes. Teams were involved earlier in admissions and could have significant impact by attending to the patient's symptoms and the psychosocial needs of both the patients and their families. There were opportunities to access new patient populations. For instance, new patients identified from the acute cohort were often transferred from outside hospitals with complex diseases, high symptom burden, and a need for intensive medical intervention, but this set of patients were usually not referred to the palliative care or were consulted in the very last hours of life. The project also enabled the services to increase contact with units that did not usually utilize palliative care. There were many opportunities to further educate other providers about the management of end-of-life symptoms, psychosocial support, and communication about end-of-life issues. The use of QPID at MGH introduced the palliative care team to a valuable methodology for identifying patients for other projects. At BWH, another important outcome was an increased appreciation of the ability of nonpalliative care social workers to accurately identify patients appropriate for consultation.

The overall consult volume increased significantly for both services from 2005 to 2009. While some of this growth is accounted for by the quality improvement project, we believe that much it was a consequence of the services having higher visibility during the project, both from the increased contact with individual physicians solicited because of the project and increased outreach efforts. In 2010, a year after the intervention was completed, consult volumes declined by 30% in nononcology populations at MGH and by 50% at BWH.

The profile of the palliative care services at the higher levels of hospital management was also significantly raised. Partners HealthCare System deepened its commitment to palliative care, providing support for a leadership retreat as well as a palliative care strategic plan. Subsequently, both services have been expanded throughout the network, have received funding for additional staff and increased resources for palliative care education. As a direct result of the project, both hospitals are now involved in projects working on increasing awareness of the relevance and utility of palliative care for end-stage heart failure patients. Most importantly, a gradual culture change occurred; palliative care grew to be viewed as a standard facet of good medical care for patients with life-limiting illnesses, particularly in the units where efforts were concentrated.

Challenges

Both palliative care teams faced challenges during this project. In targeting the acute patients, patients at risk for death in the hospital, often in their final days, were targeted. Some felt that we were undermining our message of “consult us early.” The costs required to support clinical and administrative study staff were not insubstantial. There have been arguments in the literature that money used to administer pay-for-performance projects may be better used to directly benefit patient care.²⁵ In this project, developing a database to consistently and accurately identify suitable patients was resource and time-intensive. At MGH, despite significant changes, QPID still identified a high number of patients clearly not suitable for a palliative care consultation, which required additional screening by the registered nurse/physician.

Screening for the UHC cohorts sometimes identified patients for whom the reason for consultation was nebulous or who gained marginal benefit from the consult. When the census was high, some of the staff felt that they were devoting too much effort to reaching the quality improvement goal rather than focusing on more needy patients and families. At times, this significantly impacted upon staff morale.²⁶ One solution was to develop new models for caring for the study population, particularly those in the chronic cohort. For example, although the initial plan at MGH was that all study patients would receive a full palliative care consultation, for those in the chronic cohort without uncontrolled symptoms but more in need of psychosocial support, patients were seen less frequently (e.g., three times a week rather than daily).

Limitations

This study had several limitations. The two hospitals are major medical centers with mature palliative care services, and thus generalizability may be limited for other types of sites and programs. We measured consultation rates, but are aware that pay-for-performance programs that are associated with improvements in such process of care measures have had less of an impact on patient outcomes or efficiency of care.^{22,25,27–30} Unfortunately, limited resources prevented us from obtaining data on secondary outcomes such as length of stay, readmission rates, use of intensive care, hospital costs, hospice referrals, or family satisfaction, so we cannot claim that patient care was improved or that there was a more efficient use of scarce resources. Finally, the particularities of the measured outcome—consultation on a select group of high-risk patients, as determined retrospectively by UHC criteria—limit the applicability of the results to other defined populations.

Conclusion

This quality improvement initiative was successful in increasing access to palliative care services in high-risk populations, especially non-oncology patients. Effective screening and identification methods were developed, and the pay-for-performance initiative appeared to motivate clinicians to engage with the palliative care team.

Footnotes

Acknowledgments

The MGH team acknowledges the diligent efforts of Mitchell Harris PhD and Michael Zalis MD with QPID. BWH team thanks Suzanne McHale, Susan Wante and the Center of Clinical Excellence staff for their enthusiastic support.

Author Disclosure Statement

No competing financial interests exist.

Appendix A.

University Health Consortium-Defined High-Mortality Diagnosis-Related Groups

1.	DRG 010	Nervous system neoplasms with complications
2.	DRG 014	Specific cerebrovascular disorders except transient ischemic attack
3.	DRG 079	Respiratory infections and inflammations, age >17 with complications
4.	DRG 082	Respiratory neoplasms
5.	DRG 088	Chronic obstructive pulmonary disease
6.	DRG 089	Simple pneumonia and pleurisy, age >17 with complications
7.	DRG 092	Interstitial lung disease with complications
8.	DRG 127	Heart failure and shock
9.	DRG 144	Other circulatory system diagnosis with complications
10.	DRG 172	Digestive malignancy with complications
11.	DRG 174	Gastrointestinal hemorrhage with complications
12.	DRG 202	Cirrhosis and alcoholic hepatitis
13.	DRG 203	Malignancy of hepatobiliary system or pancreas
14.	DRG 205	Disorders of the liver except malignancy, cirrhosis and alcoholic hepatitis with complications
15.	DRG 239	Pathologic fractures and musculoskeletal and connective tissue malignancy
16.	DRG 296	Nutritional and miscellaneous metabolic disorders, age >17 with complications
17.	DRG 316	Renal failure
18.	DRG 395	Red blood cell disorders, age >17
19.	DRG 398	Reticuloendothelial and immunity disorders with complications
20.	DRG 403	Lymphoma and non-acute leukemia with complications
21.	DRG 416	Septicemia, age >17
22.	DRG 423	Other infectious and parasitic disease diagnoses
23.	DRG 473	Acute leukemia without major OR procedure, age >17
24.	DRG 475	Respiratory system diagnosis with ventilator support
25.	DRG 489	HIV with major related condition
26.	DRG 492	Chemotherapy with acute leukemia as secondary diagnosis

a

The list of DRGs is available from upon request.

References

Higginson

, Hearn

. A multicenter evaluation of cancer pain control by palliative care teams. J Pain Symptom Manage, 1997; 14:29–35.

Bakitas

, Lyons

, Hegel

, Balan

, Brokaw

, Seville

et al. Effects of a palliative care intervention on clinical outcomes in patients with advanced cancer: the Project ENABLE II randomized controlled trial. JAMA, 2009; 302:741–749.

Temel

, Greer

, Muzikansky

, Gallagher

, Admane

, Jackson

et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med, 2010; 363:733–742.

Ringdal

, Jordhoy

, Kaasa

. Family satisfaction with end-of-life care for cancer patients in a cluster randomized trial. J Pain Symptom Manage, 2002; 24:53–63.

Casarett

, Pickard

, Bailey

, Ritchie

, Furman

, Rosenfeld

et al. Do palliative consultations improve patient outcomes? J Am Geriatr Soc, 2008; 56:593–599.

Gelfman

, Meier

, Morrison

. Does palliative care improve quality? A survey of bereaved family members. J Pain Symptom Manage., 2008; 36:22–28.

Fromme

, Bascom

, Smith

, Tolle

, Hanson

, Hickam

et al. Survival, mortality, and location of death for patients seen by a hospital-based palliative care team. J Palliat Med, 2006; 9:903–911.

Morrison

, Penrod

, Cassel

, Caust-Ellenbogen

, Litke

, Spragens

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

Penrod

, Deb

, Luhrs

, Dellenbaugh

, Zhu

, Hochman

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

10.

Zhang

, Wright

, Huskamp

, Nilsson

, Maciejewski

, Earle

et al. Health care costs in the last week of life: Associations with end-of-life conversations. Arch Intern Med, 2009; 169:480–488.

11.

Bradley

, Weaver

, Brasel

. Addressing access to palliative care services in the surgical intensive care unit. Surgery, 2010; 147:871–877.

12.

Widera

, Pantilat

. Hospitalization as an opportunity to integrate palliative care in heart failure management. Curr Opin Support Palliat Care, 2009; 3:247–251.

13.

Davison

. End-of-life care preferences and needs: Perceptions of patients with chronic kidney disease. Clin J Am Soc Nephrol, 2010; 5:195–204.

14.

Goodlin

. Palliative care in congestive heart failure. J Am Coll Cardiol, 2009; 54:386–396.

15.

Yohannes

. Palliative care provision for patients with chronic obstructive pulmonary disease. Health Qual Life Outcomes, 2007; 5:17.

16.

Mitchell

, Teno

, Kiely

, Shaffer

, Jones

, Prigerson

et al. The clinical course of advanced dementia. N Engl J Med, 2009; 361:1529–538.

17.

Ferrell

. Late referrals to palliative care. J Clin Oncol, 2005; 23:2588–2589.

18.

Costantini

, Toscani

, Gallucci

, Brunelli

, Miccinesi

, Tamburini

et al. Terminal cancer patients and timing of referral to palliative care: A multicenter prospective cohort study. Italian Cooperative Research Group on Palliative Medicine. J Pain Symptom Manage, 1999; 18:243–252.

19.

Osta

, Palmer

, Paraskevopoulos

, Pei

, Roberts

, Poulter

et al. Interval between first palliative care consult and death in patients diagnosed with advanced cancer at a comprehensive cancer center. J Palliat Med, 2008; 11:51–57.

20.

www.uhc.edu

21.

Weissman

, Meier

, Spragens

. Center to Advance Palliative Care palliative care consultation service metrics: Consensus recommendations. J Palliat Med, 2008; 11:1294–1298.

22.

Alkasab

, Harris

, Zalis

, Dreyer

, Rosenthal

. A Case tracking system with electronic medical record integration to automate outcome tracking for radiologists. J Digit Imaging(in press).

23.

Fadul

, Elsayem

, Palmer

, Del Fabbro

, Swint

, Li

et al. Supportive verus palliative care: What's in a name? A survey of medical oncologists and midlevel providers at a comprehensive cancer center. Cancer, 2009; 115:2013–2021.

24.

Langley

, Nolan

, Norman

, Provost

. The Improvement Guide: A Practical Approach to Enhancing Organizational Performance, 2nd. San Francisco: Jossey-Bass, 2009.

25.

Fonarow

, Peterson

. Heart failure performance measures and outcomes: real or illusory gains. JAMA, 2009; 302:792–794.

26.

Werner

, Asch

. The unintended consequences of publicly reporting quality information. JAMA, 2005; 293:1239–1244.

27.

Campbell

, Reeves

, Kontopantelis

, Sibbald

, Roland

. Effects of pay for performance on the quality of primary care in England. N Engl J Med, 2009; 361:368–378.

28.

Grumbach

, Osmond

, Vranizan

, Jaffe

, Bindman

. Primary care physicians' experience of financial incentives in managed-care systems. N Engl J Med, 1998; 339:1516–1521.

29.

Glickman

, Schulman

, Peterson

, Hocker

, Cairns

. Evidence-based perspectives on pay for performance and quality of patient care and outcomes in emergency medicine. Ann Emerg Med, 2008; 51:622–631.

30.

Conrad

, Perry

. Quality-based financial incentives in health care: Can we improve quality by paying for it? Annu Rev Public Health, 2009; 30:357–371.