Integrating Case Management and Palliative Care - CiteSeerX

JOURNAL OF PALLIATIVE MEDICINE Volume 7, Number 1, 2004 © Education Development Center, Inc.

Promoting Excellence in End-of-Life Care

Integrating Case Management and Palliative Care DIANE E. MEIER, M.D.,1 WILLIAM THAR, M.D.,2 ASHBY JORDAN, M.D.,3 SUZANNE L. GOLDHIRSCH, M.A.,1 ALBERT SIU, M.D.,1 and R. SEAN MORRISON, M.D.1

ABSTRACT Most seriously ill Americans live at home under the care of their primary physician and with the support of family caregivers. To reduce costs while simultaneously improving the quality of patient care, insurers have increasingly turned to the concept of case management. While case management is targeted to individuals with life-threatening illnesses, palliative care assessment and interventions are typically not included in the management protocols. An academic/care management/health plan partnership between Mount Sinai School of Medicine, Franklin Health, a care management organization, and South Carolina Blue Cross Blue Shield, was formed in 1998 to test the utility of integration of case management with formal palliative care assessment, feedback and recommendations to treating physicians, and ongoing support for implementation of a palliative care plan. The goal of the project was to ensure identification and optimal care of seriously ill patients’ complex needs, while facilitating doctor-patient continuity, improving patient/family/physician communication, providing assistance with decision-making, ensuring quality care at home, and promoting efficient use of health care resources. Care management nurses were randomly assigned to a control (usual care) group or to the intervention (palliative care) group. Intervention nurses were trained in formal palliative care assessment and interventions, supported by treatment protocols and communication strategies with treating physicians. Measurements included symptom burden, prescribing practices, advance care planning status, satisfaction, and health care utilization. These results are pending completion of study run-out and analysis. Preliminary programmatic results indicate that combining palliative care with the case management approach is a logical, feasible, and effective strategy to improve the care of seriously ill patients living in the community. Franklin Health has offered the program to their entire client base because they feel that the integration of palliative care into their case management program improved the standard of patient care. Blue Cross Blue Shield of South Carolina has also chosen to sustain this enhanced model of care management for seriously ill patients. INTRODUCTION

W

50% of Americans die in hospitals, this statistic masks the fact that the overwhelming majority of terminally ill paH ILE MORE THAN

tients spend most of their last year of life at home being cared for by their families without the assistance of formal home care services.1 This small group of complex and seriously ill patients (approximately 1% of the commercially insured pop-

1 Hertzberg Palliative Care Institute; Department of Geriatrics and Adult Development, and Mount Sinai School of Medicine, New York, New York. 2 Franklin Health, Inc. 10 Mountainview Road, Upper Saddle River, New Jersey. 3 Blue Cross Blue Shield of South Carolina, Columbia, South Carolina.

119

120

MEIER ET AL.

ulation and 5% of Medicare-covered lives) drives one third to more than one half of all health care expenditures because of high hospitalization rates and costly treatment needs (Fig. 1).2 In an effort to not only improve care but also to minimize spiraling health care costs, insurers have turned to case management as a key component in the care of seriously ill patients with complex needs (Fig. 2). In one study of 18 large Medicare health management organizations (HMOs), virtually all included case management programs of varying degrees of intensity, which conducted screening, assessment, care planning, implementation, and ongoing monitoring.3 Although case management programs are often directed at patients with serious and life-threatening illness, they typically focus primarily on service coordination, to a lesser extent on medical evaluation and decision making, and only rarely incorporate standard palliative care assessment and interventions in their role and functions.4–9 Conversely, whereas palliative care assessment and feedback to physicians has been shown to be both

Patient Type Worried well Self-resolving illness Low grade acute illness

feasible and to improve outcomes in settings such as hospice, the hospital, and in long-term care settings,10–15,17–21 such assessment rarely occurs in community settings where most seriously ill persons live. Why are persons with serious illness in need of community-based palliative care services when hospice care is widely available in the United States? At present, only approximately 20% of all adult decedents in the United States are served by hospice.11,19 The median length of stay in hospice has steadily declined in recent years to a median of 25 days11 and one third of hospice patients receive services for less than a week before death. Many patients and families in need of palliative care fail to access hospice because of regulatory restrictions requiring a prognosis of 6 months or less and patient agreement to give up insurance coverage for disease-modifying treatments in order to be eligible for hospice benefits.18,22,23 As a result, for the 77% of persons who die of chronic illnesses other than cancer, and for those with serious illness of any

100

% Claimants

Cost per Claimant

Management Approach

Low

Demand Management

Medium

Disease Management

90 80

Chronic diseases Moderate to severe acute illness

70 60

Complex Patients

50

Significant diagnosis

40

Multiple co-morbidities

30

Often terminal

20

Several providers of care

10

Psychological/social/financial upheaval

FIG. 1.

Care management focuses on the needs of patients.

Case Management

High

Complex Care Management

121

Resources

CASE MANAGEMENT AND PALLIATIVE CARE

Demand Management

CM/DM

CCM-palliative care

NEEDS

CM/DM = case management/disease management CCM = complex case management

FIG. 2.

Matching resources to needs.

kind whose prognosis is uncertain, or who wish continued access to life-prolonging treatments, the current hospice benefit model is a poor fit. For these reasons, alternate models of palliative care delivery are needed for a large proportion of community-dwelling persons with serious, chronic, and complex illness. In 1998, an academic/case management/health plan partnership was formed between Mount Sinai School of Medicine; Franklin Health, Inc., a case management organization serving complex and high cost patients; and South Carolina Blue Cross Blue Shield, an insurance company contracting with Franklin for case management services. Funding was obtained from Promoting Excellence in End-of-Life Care National Program Office of The Robert Wood Johnson Foundation to design, pilot, and evaluate a new model of health care delivery with goals involving integration of routine complex case management with formal palliative care assessment, feedback to treating physicians, and support for implementation of treatment. Because community-based complex case management is a rapidly growing and increasingly available means of delivering care to a vulnerable class of patients with costly healthcare needs, this collaborative academic/industry initiative to

study the effectiveness of delivering palliative care through case management programs has the potential for broad generalizability. The true measure of the feasibility and effectiveness of this model of will be whether the study intervention delivers improved patient care without significantly increasing utilization and costs. The claims data required to evaluate these outcomes are not available at this time and will be published at a later date. This report, therefore, focuses on the initial study design and methods, implementation of the intervention, barriers encountered, subsequent modifications in the study design, and the response of the case management company and the purchasing insurer to the project.

PATIENT POPULATION AND SETTING Patients targeted by this project were seriously ill, commercially insured HMO members who entered into a complex case management program through Blue Cross Blue Shield of South Carolina indemnity plans and their affiliate HMO, Companion Health Care. Patients were screened by Franklin Health and identified as appropriate for case management through a proprietary method. Criteria for entry into the Complex Care Man-

122

agement Program include specific diagnostic categories (cancer, heart disease, chronic lung, kidney, liver disease, major trauma, degenerative neurologic disease) and resource utilization data such as hospitalization or home-care service use indicative of complexity and service needs. Blue Cross Blue Shield of South Carolina, headquartered in Columbia, South Carolina, has more than 1 million members in South Carolina alone. Through government contracts and additional business, the company serves an additional 21 million customers nationwide, making it one of the largest Blue Cross Blue Shield programs in the United States. Its leadership has made a firm commitment to offering complex case management services to their members in South Carolina, and was willing to serve as a test site for the palliative care intervention. Franklin Health, Inc. is a for-profit national community-based complex case management organization, with offices located in Upper Saddle River, New Jersey (,www.franklinhealth.com.). With a large and successful program of patientcentered, complex case management for the highrisk seriously ill patient in a managed care setting, Franklin’s patient base extends to all 50 states. A computerized information system enables detailed tracking of outcomes both for both individual patients and groups. In addition to South Carolina Blue Cross Blue Shield, Franklin has multiple corporate clients including managed care plans, health systems, and large self-insured corporations. Franklin Health provides an intensive and highly standardized system of complex case management delivered to patients and their families primarily in the home, but which remains in effect regardless of patient venue. Approximately 25%–30% of Franklin’s patient population dies or is referred to hospice while receiving case management. Most of the remaining patients are discharged when their medical condition stabilizes, or resolves. Average length of case management services is 5 months. The Complex Care Management Team (Fig. 3) includes a centrally located (in Upper Saddle River, New Jersey) physician case manager employed by Franklin, who provides medical direction for team nurses and communicates with the treating physician; a centrally located (also in Upper Saddle River, New Jersey) nurse clinical account manager who is responsible for case selection, supervision and support of nurse care coordinators, and coordination of the

MEIER ET AL.

team and team conferences; and typically four local community-based nurse care coordinators, who provide local community-based coordination of services and direct (initially in-person and subsequently telephonic) support and advice to patients and their families. Because the care coordinator nurses are geographically based, they are able to identify and implement community supports for each patient and family unit. Weekly (or more frequent) team meetings are held by telephone and review of individual cases occurs throughout the course of care. Day-to-day communication occurs both by telephone and through electronic information systems. The goal of the care provided by the complex care management team is to try to improve the likelihood that the patient will receive high-quality and cost-effective care in accordance with his or her informed wishes. Franklin’s leadership considers the highly structured and consistent management framework to be the key to its success: individualized high-quality patientlevel services are provided in a uniform and standardized manner; via a team structure and process; with a high degree of accountability and supervision; and utilizing weekly telephonic team meetings and daily information system reporting. The result is that no single individual on the team makes patient care decisions alone. Global satisfaction of Franklin recipients has been high (95% of recipients would recommend or highly recommend the program to others), suggesting that patients and families feel their needs are appropriately met under Franklin’s care. The Complex Care Management (CCM) program has consistently demonstrated a return on investment of 1.8 to 1, which is realized largely through reductions in hospitalizations and procedures.24 These data underscore the attraction of the case management model of care for the sickest and costliest patients to health plans, corporate insurers, and Medicare providers, and supports the logic of building upon the existing structure of case management programs as an efficient and appropriate locus for application of a palliative care intervention.

PROGRAM DESIGN Intervention The initial project design established four specific objectives: (1) To develop training modules

123

CASE MANAGEMENT AND PALLIATIVE CARE Care Coordinator nurse – Primary contact for patients/family/providers/plan case managers – Talks to patient before they make care plan choices to help them make the best choice for themselves – Lives in/near patient’s community (9 Care Coordinators in South Carolina; 50 across continental US) – Meets every patient & family at home/facility and provides additional visits as needed – Case load: 20 patients – 35–40 hours per case (5 month average case length)

Clinical Account Manager – Provides case direction/collaboration to CC and acts as a daily resource to ensure the big picture is in view – Represents client’s business perspective – Interfaces with on-site plan clinical liaison – Assesses financial impact and measurement – Case load: 80 patients

FHI Account Physician FHI Clinical Account Manager

Physician Manager – Performs weekly case reviews with team to anticipate issues – Communicates with treating physicians and plan physicians as needed to offer support and solutions – Provides medical perspective as only an actively practicing clinician can – Case load: 100 patients

FIG. 3.

Patient & Treating Physician

FHI Care Coordinator

The Franklin Health Clinical Team.

for case management nurses on palliative care assessment, identification and implementation of treatment interventions, and approaches to physician feedback; (2) To provide formal palliative care assessment with structured feedback and recommendations for physicians whose patients are enrolled in the complex care management program; (3) To compare outcomes for intervention patients receiving palliative care assessment plus case management versus control patients receiving standard case management alone; and (4) To assess the feasibility and acceptability of this model to patients, family caregivers, health professionals, and industry. These four objectives were implemented in three phases over a period of 39 months (October 1, 1998 to December 31, 2001). Phase 1: (October 1, 1998). Case-based training module materials and in-service practice sessions were developed to ensure the expertise of those Franklin care coordinator nurses who were assigned to the intervention group. The training focused on the conduct, delivery, and implementation of palliative care assessment and treatment protocols (Fig. 4).25,26 Computerized protocols for the assessment instruments, subsequent treatment recommendations, and guidelines for pro-

viding feedback and recommendations to physicians were developed, reviewed by focus groups, and piloted. Explicit stepped-care pathways or treatment alternatives were also developed for each physical symptom assessed, for psychosocial, spiritual, or existential concerns identified on the Quality of Life Index,27 and for identified communication and advance care planning preferences. These treatment and follow-up protocols were reviewed and modified with Franklin nursing staff as part of the development and piloting process in phase 1. The treatment protocols were placed on the Franklin Web-based information system with appropriate Web links. For example, if severe pain is identified, the nurse is brought to the protocol section focused on choices of opiate analgesics and routes of administration; and can click on links to opiate conversion tables and standard bowel regimens. Instruments for evaluation of acceptability and utility of the intervention to patients, family, nurses, and treating physicians were also developed during phase 1. Phase 2: (November 1, 1999 to December 31, 2001). All patients (cared for by control and intervention nurses) were selected for the CCM program through a multiple step screening process (Fig. 5). The process of entering the Franklin Complex

124

MEIER ET AL.

FIG. 4.

Training program content: 1.5 days.

Care Program was the same for both control and intervention patients. Initial visits by care coordinators took place primarily at home. The case management system was explained in detail, and a voluntary patient informed consent was obtained for participation in both the Franklin Complex Care Program and the study. After a standardized baseline assessment (Fig. 6), an initial evaluation summary was prepared, and a care plan was developed and discussed with the patient/family. After documentation of the patient’s and family’s choices, a treatment plan was formulated based upon patient goals, actions needed, outcomes, providers, medications, and services.

FIG. 5.

A total of 321 Franklin complex case management patients were enrolled in the study and received either usual case management (169 control) (Figs. 3, 5, and 6) or case management enhanced by palliative care assessment and physician feedback/recommendation protocols (152 intervention; Fig. 7). Care Coordinator nurses were randomly assigned to provide either usual case management (4 nurses) or the palliative care enhanced intervention (5 nurses) Clinical assessments for all patients were tracked using the Franklin Personal Path Care Management System (PCMS), which tracks information related to a patient’s enrollment, assessments and outcomes.

Patient identification and selection.

125


FIG. 6.

Franklin Health baseline assessment instrument.

This database is one of the largest of its kind in the United States and uses a sophisticated system of issues identification, goal setting, action tracking, and outcomes assessment. Additional assessment data for the study were also tracked (Fig. 8). For intervention cases only (Fig. 7), the following additional assessment instruments were used: (1) The Franklin Health Communication Instrument (questions derived from the Toolkit28 ), was used to assess patients’ interest in discussing end-of-life care with their family and physicians,

FIG. 7.

and in appointing a health care agent; (2) The Edmonton Symptom Assessment Scale (ESAS),29 modified from a 10-cm visual analogue scale to a four-point numeric rating scale (none to severe), served to assess symptoms (activity, anxiety, appetite, constipation, depression, physical discomfort, dyspnea, fatigue, pain, nausea, and well being). Symptom assessment was conducted every 2 weeks for patients who were stable and not experiencing distressing symptoms (defined as moderate or severe). If moderate or severe symptoms were identified, assessment was carried out

Intervention group.

126

MEIER ET AL.

weekly (or more often) at the discretion of the intervention care coordinator nurse and the clinical account manager. Treatment recommendations based on the computerized protocols were telephoned in to the primary physician’s office for symptoms ranked as moderate or severe, or if a patient expressed a desire to discuss goals of care with the physician. In the case of specific medical treatment recommendations (such as a new prescription medication), either the nurse care coordinator or the Franklin physician case manager made a direct call to the treating physician. For control patients, the initial assessment includes questions regarding general health state and advanced care planning. These queries, did not use the formalized assessment instruments used by the intervention nurses, nor were they repeated on a regular basis. Treatment of any symptoms identified by control nurses were left to the clinical judgment of the nurse and CCM team, and were not directed by the specific treatment protocols developed for the intervention group. All patients were followed until death or case closure. Fig. 9 describes patient and data flow through the study. Within 2–7 working days after a case was opened, all patients (control and intervention) received a brief telephone assessment of symptom burden using the modified

FIG. 8.

ESAS only. This survey was conducted by study staff blinded to the assignment group of the patient. This blinded telephone assessment of symptoms was repeated for all subjects 6–8 weeks after the initial baseline telephone assessment. Patient and family satisfaction was assessed 3–8 weeks after case closure (or death) using Franklin Health’s standard instrument (,www.morpace.com/customer-satisfaction.htm.) with the addition of global quality of life questions in the physical and emotional domains. Phase 3. The study is currently in phase 3, completing run-out of subjects still in case management in the intervention and control groups, and analyzing phase 2 data in terms of both patientcentered outcomes (pain and symptoms, quality of life, communications about treatment preferences, satisfaction) and utilization criteria (hospital days, case management days, in-patient and out-patient services, location of death, hospice referral) (Fig. 8).

Innovation This project was conceived in response to the pressing need for the integration of palliative care into the services offered to a seriously ill

Patient and data flow.

127


FIG. 9. Outcome measures.

population living at home or being cared for across multiple settings. While our approach was derived from preexisting models of complex case management and hospital-based palliative care, we discuss here innovative three aspects of the project: Industry/academic partnership. For Franklin Health Inc., partnership with Mount Sinai presented an opportunity to develop a better model of care, quantify and report on the improvement, and use the results to attract a larger client base. For Mount Sinai’s Palliative Care Program, the project was a vehicle to export palliative care out of the hospital or office, and into the home setting where the majority of seriously ill patients reside. Each partner relied on the skills and resources of the other to achieve shared objectives. Franklin Health provided the existing platform of their CCM model, a connection to a patient population with serious illness and complex needs, and an insurance client, Blue Cross and Blue Shield of South Carolina, who was willing to participate in the project. The investigators at Mount Sinai School of Medicine contributed the educational, clinical and research expertise needed to design, implement and evaluate the intervention. Training of CCM nurses. A second distinctive component of this project was the training offered to the nurse care managers concerning (1)

administration of formal assessment instruments; (2) use of the assessment results to develop appropriate treatment recommendations; and (3) employment of a variety of strategies to communicate effectively with busy, treating physicians (Fig. 4). Not only were the nurse care managers empowered to identify patient distress in this study, but they were also given responsibility to do something about it. In other case management models, unmet patient needs may or may not be identified. When a need is identified, nurses may not have the knowledge, the confidence, nor the authority to devise an appropriate intervention. The palliative care training of Franklin’s nurses, and the project design, explicitly set out to: (1) formally and repeatedly screen for a wide range of symptom distress and (2) provide nurses with detailed computerized treatment protocols so that a symptom, once identified, could be effectively addressed by the nurse, with the support and guidance of the FHI clinical account manager and physician case manager, and in discussion with the patient’s treating physician. Encouragement and facilitation of physician involvement. The complex case management plus palliative care program encouraged a strong working relationship with the treating physician, focusing on supporting and implementing the physician’s efforts to provide good care for their

128

MEIER ET AL.

complex patients, and creating an incentive for physicians to actually make use of the results of the palliative care assessment protocol. The nurse’s role was conceptualized as a support for both patient and family and the treating physician. Through the routine identification of important clinical issues and the hands-on support for whatever intervention the physician decided on, the nurse enhanced the physician’s ability to provide meticulous, high-quality care to seriously ill persons living at home. None of the case management nurses conceptualized their role as primarily focused on cost savings, but rather worked together to achieve a common goal—optimal care of the patient and family.

PROGRAM EVOLUTION Challenges encountered/strategic responses As the study progressed, a series of challenges prompted modification of the design and implementation of the study, as follows: 1. Finding the right partnership: The first major challenge was to put together a viable research partnership comprised of an academic institution, a commercial health insurance plan, and a case management service organization. In the rapidly changing health care environment, it was important to identify a partnership that would be stable over the period of the research project. The original proposal paired Mount Sinai School of Medicine with Oxford Health Plans and Franklin Health, Inc. But because of economic difficulties, Oxford abandoned the business of Medicare Managed Care targeted by this study and, in addition, ended many vendor contracts, including that with Franklin Health, Inc. After Oxford withdrew from the proposed study, Franklin reviewed their client list and proposed replacing Oxford with Blue Cross Blue Shield of South Carolina, and this insurer was willing to participate in the research study. 2. Adjustment in Patient Population: The original proposal targeted a Medicare-risk population. The new project partner, Blue Cross Blue Shield of South Carolina, had a commercially insured, employed population that met the criteria for the study, but with one major differ-

ence: according to Franklin’s prior experience in the field, the number of commercially insured individuals requiring complex case management was one tenth that seen in the Medicare population, and a higher proportion of those targeted for complex care management recovered sufficiently from the illness to no longer require it. As a result, the number of patients participating in this commercial plan who were terminally ill was much lower than that originally projected for a Medicare population. The focus of the project was adjusted to compensate for this change in the diagnoses and life expectancy rates of the new target population. The emphasis was shifted from demonstrating the impact of palliative care on the care of terminally ill patients to delivering palliative care to seriously ill patients independent of prognosis. 3. Difficulty reaching and communicating with physicians: Intervention nurses reported persistent problems with reaching treating physicians directly to discuss symptom management protocols. In response to these encounters, a CDROM entitled A Practical Guide to Communication Skills in Clinical Practice30 was given to each program case management nurse. A review of the literature was also conducted to develop evidence-based one-page physician information sheets on palliative care symptom management for anorexia/cachexia, dyspnea, constipation, fatigue, and nausea and vomiting. The purpose of these information sheets was to enrich the support given to intervention nurses as they attempted to make recommendations to primary physicians about symptom management. These one-page referenced treatment summaries were distributed by fax to the physician’s office at the time of the telephone call. The goal of this educational strategy was to encourage and support physicians in the use of sometimes unfamiliar therapies (such as dexamethasone or methylphenidate for fatigue). 4. When necessary, and upon request of the case management nurses, the Franklin Physician Case Manager telephoned directly to South Carolina treating physicians to discuss new therapeutic recommendations. 4. Accrual of study subjects took longer than anticipated: Because of slower than anticipated accrual of new patients in the South Carolina


Blue Cross Blue Shield and Companion Health Services client base, the enrollment of the 322 patients included in this study took longer than expected and required extension of the subject entry period. 5. Difficulty obtaining program evaluations from primary physicians: The program evaluation staff had difficulty contacting the physician directly for their evaluation of Franklin’s service to their patients. Staff were often directed by office nurses and managers to fax the form to the physician’s office, an action which only rarely resulted in a returned and completed survey. Specific impediments to a better physician response rate were not documented. It is likely that, in the busy medical practices we contacted, survey completion was given a low priority.

Evaluation Initial evaluation design and plan. This project was designed to test the hypothesis that incorporation of routine palliative care assessment and support for patient family centered care planning and decision making, with practical feedback and support for physicians regarding implementation, can be a feasible and effective component of home-based (as opposed to office- and institutional-based) complex case management programs. The objective of our evaluation plan was to examine the effectiveness of this intervention by comparing clinical and utilization outcomes to a control group of patients receiving case management only. Although it would have been methodologically ideal to randomize individual patients, the fact that Franklin Health was already using a CCM model affected how the intervention and control group were devised. Early in the planning stages of the project, for example, a decision was made that the study would not compare the intervention to no case management at all. Subsequently the investigators devised a comparison between the existing Franklin CCM program (CCM control) and Franklin Plus (CCM plus palliative care assessment and feedback: CCM1). The evaluation itself focuses on three critical areas: patient centered outcomes, provider outcomes, and utilization of medical resources. Primary outcome measures selected for patients and families include: (1) percent of eligible patients

129

who agreed to participate and in whom the intervention assessment process could be completed; (2) results of the assessment instruments; (3) the feasibility and acceptability of the process to the participants; and (4) the impact on patient care of the intervention nurse feedback to physicians concerning assessment results, treatment recommendations, and support for implementation of a palliative care plan. Assessment instruments and outcome measures. The original evaluation plan included four assessment instruments employed to measure patient/family outcomes (Fig. 9): Modified Edmonton Symptom Assessment Scale, 28 converted from a visual analogue scale to a numeric rating scale (0, none to 3, severe) to facilitate telephonic use. The ESAS assesses pain, shortness of breath, appetite, nausea, sense of well-being, activity level, fatigue, anxiety, and depression. Five questions were added to the original ESAS (pain now; pain over last 3 days; acceptability of pain control; level of constipation; time of last bowel movement). Missoula-VITAS 16-Item Quality-of-Life Index: A 16-item survey with a 5-point Likert scale that assesses quality of life across 5 domains.27 Morpace Satisfaction Instrument: Telephonic satisfaction survey of individuals whose care management plan has been completed. Surveys are conducted by independent market research and consulting organization, Morpace International, Inc.16 Advance Care Planning Questionnaire: An eight-question instrument modified from the Toolkit28 that was used to assess patient’s interest both in discussing end-of-life care with their family and physicians and in appointing a health care agent. Questions included yes/no queries about specific wishes for care; prior communication with family or physician; desire for communication with family or physician; prior completion of an advance directive; desire to discuss the advance care planning process. The provider outcome measures explored the acceptability and feasibility of this intervention for the nursing and medical staff, and industry participants. Physicians, for example, were queried for their opinion on the utility and ben-

130

efit of the intervention for their patients. Five questions were included in the physician survey: 1. Did you find the assessments and recommendations helpful? 2. Did you follow any of them? 3. Did the case management team help in their implementation? 4. Did you perceive those recommendations you followed to improve the care of your patient? 5. Would you like to have future patients enrolled in this type of program? Finally, measures of utilization were tracked during the study: hospital days, intensive care unit (ICU) days, emergency department use, physician visits, length of stay in complex care management, physician inpatient and outpatient relative value units (RVUs), hospice referral rate, home care services used, analgesic/anxiolytic/ antidepressant prescribing, and site of death. Adjustment of evaluation plan and instruments. The evaluation plan required adjustment in the assessment tools. The Missoula-VITAS Quality-of-Life Index was eliminated as a result of feedback from nurses during the piloting of the assessment package. The nurses felt that the instrument was most appropriate in a terminally ill population who had acknowledged the fact that they were dying. Only 30% of patients in this project died during case management. Most recovered sufficiently from their illness to no longer need complex case management or an adequate long-term management plan was put in place. As a result, routine questions focused on facing death were neither appropriate nor well-tolerated by the majority of study subjects. Instead, two questions were added which assessed Franklin Health’s performance in terms of patient/family perception of improvement in their quality of life as the result of the intervention. Overall, has Franklin Health improved your/the patient’s quality of life? What could we do to make the Franklin Health Program more helpful to improve your/the patient’s quality of life? A second minor adjustment involved the modified ESAS. Because of the prevalence of constipation as a side effect of the use of opiates, two

MEIER ET AL.

questions about constipation as a symptom were added. In addition, patients were asked three questions about pain (as opposed to the single question included on the unmodified ESAS): current levels of pain, pain over the last three days, and acceptability of pain control. Randomization. Because Franklin Health was already using a CCM model, the risk that care of intervention patients would influence or contaminate the care of control patients treated by the same clinical coordinator nurse was high. This issue affected how the intervention and control groups were devised. To reduce the risk of contamination, two completely separate CCM teams with separate physician case managers and clinical account managers were created. The approximately 10 South Carolina clinical coordinator nurses were then randomly assigned to the control team (usual CCM) or intervention team (CCM 1 palliative care assessment and feedback). Final evaluation of nurses’ assessment of the program (in both intervention and control groups) will be conducted once run-out of the final study patients is complete. Exit interviews with industry (South Carolina Blue Cross Blue Shield) staff will also be conducted once run-out is complete.

DISCUSSION Results Demographics of study subjects may be found in Table 1. The study is currently in the evaluation and analysis phase, and therefore final clinical and utilization results are not currently available. The ensuing discussion of the project is based upon preliminary data, interviews with stakeholders, and field observations. Feasibility of the programmatic intervention: Industry. Complex patients typically represent a subset within the top 1% of high utilizers of health care resources. The ability to manage the care of these high-risk patients carefully, thereby exerting some control over the costs involved, is part of responsible management of health care and is critical to the fiscal viability of the health care industry (Figs. 1 and 2). Prior to this study, Franklin Health had implemented and refined a CCM program to work with this group of outlier patients. It was relatively easy for the research

131


team to integrate a palliative care focus into this preexisting, intensive, and highly structured case management platform. The resulting delivery model has now been field tested and refined as a product ready to be utilized by other health plans and their case management services. Because of their experience with this model, and Franklin’s belief in the value-added dimension of the palliative care component, the program has been rolled out to Franklin’s entire client base. This decision was based on two factors: (1) preliminary evaluation of the study data showed a trend toward improved patient care and patient satisfaction without an increase in program costs and (2) increasing industry demand for case management services incorporating a palliative care focus. Patients and families. The Franklin CCM program was designed to function in an advocacy mode, incorporating the following benefits to patient and family: 1. Provides information and professional support for patient decision-making; 2. Supports and guides the patient and family in their relationship to the health care system; 3. Offers case management services to secure access and oversee the quality of services delivered within a physician-directed plan of care; 4. Remains available to patients and families through the entire course of the illness, even if there is a change in setting or geography. Because of the addition of routine and formal palliative care assessment and feed-back to their physician in the areas of pain and symptom management and psychosocial support, patients and families in this study received health care that was calibrated to match the intensity and complexity of their needs. We hope to demonstrate that it is also effective in terms of clinical and utilization outcomes. Clinicians. Materials developed for this project are simple to use, feasible in the real clinical setTABLE 1.

DEMOGRAPHICS

Average age Number (%) female Number (%) with cancer

OF

STUDY SUBJECTS

Control n 5 169

Intervention n 5 152

48.6 91 (54) 124 (73)

47.6 87 (57) 118 (78)

ting, and of specific value to the doctors and nurses actually taking care of the patient. This emphasis on relevance and practicality was intended to maximize the chances that recommendations based on the assessment protocol would actually be implemented by the patient and physician, with the support of the case management nurse. The nurse case managers were enthusiastic about the palliative care training, describing themselves as empowered by the new knowledge, the requirement for conduct of regular symptom assessment, and an enhanced role definition that included taking responsibility for initiating action when a problem was identified. With the special training, computerized treatment guidelines, and the Franklin approach (field nurses are supported in their work by the weekly team case discussions and constant team contact through Franklin Health Information Systems), nurses could engage in comprehensive clinical assessment because they knew with confidence what to do in response to a problem identified as a result of their assessment and specifically, what interventions to recommend when they contacted the treating physician by telephone. Acceptability and sustainability of the intervention among the “key stakeholder groups” (patients, families, clinicians, institutions). Preliminary reports have indicated that this model is feasible, and that it increases the level of patient and family satisfaction with care. The actual assessment data from the patient/family and physician surveys is currently being gathered and analyzed. There is evidence from other sources, however, which demonstrates a strong positive reaction among key stakeholders. Blue Cross Blue Shield of South Carolina, for example, has consistently demonstrated an enthusiasm for the program and an interest in continuing their participation in the project. Franklin Health, Inc.’s decision to integrate the palliative care intervention module into the program offered to their entire client roster nationwide is, in fact, the strongest evidence of feasibility, acceptability and perceived value of this intervention, suggesting that the palliative care intervention has enhanced the ability of Franklin Health to market its services to prospective consumers.

Generalizability The CCM approach described here is flexible and can be applied to a range of patient populations in a variety of plan arrangements. The pre-

132

dictable instability of the health care market does not preclude the use of this clinical model. Because the model depends on existing nursing staff evaluations, communications with treating physicians, implementation of treatment plans, and the availability of computerized information support systems, it can be also integrated into a variety of other types of care systems including traditional visiting nurse home care programs, PACE programs, fee-for-service, and medical staff or provider-based organizations. This protocol will be posted on the website of the Center to Advance Palliative Care (,www. capc.org.). The palliative care assessment tools, training modules for nurse case managers, protocol guidelines for treatment recommendations, and patient and professional satisfaction instruments will also be made available.

Implications Significant instability exists in the health care area in general—on the policy level, in the field, and in the insurance and managed care industry itself. This instability has a direct impact on the effort to conduct the type of rigorous research needed to improve the quality of care. The second insight gained from this project focuses on the importance of empowering health professionals to act on behalf of their patients. This observation was prompted by the unexpected enthusiasm and positive reaction of the team nurses, who despite their training in complex care management, gained confidence and initiative in the management of patient distress as a result of the palliative care assessment and intervention modules. Over a relatively short period of time, they collectively progressed from some skepticism and insecurity about these assessments and interventions to an attitude of enthusiasm and excitement about the effectiveness of their new clinical skills. The investigators theorize that even though the team nurses had been routinely confronted with palliative care issues, they had not had the opportunity to implement a formal, structured palliative care assessment and feedback approach at this level of intensity. More importantly, they had not previously been in the position of making the specific recommendations to the treating physicians that they were trained to provide in this program. As a result of this project, the nurses were

MEIER ET AL.

empowered to be more assertive and directive as advocates for their patients who were suffering— and they were given the information necessary to back up their observations and their recommendations. This model is powerful because it enables physicians and nurses to do a better job, reinforced by their perception that the patients actually benefit as a result.

Limitations Limitations of this study include the fact that satisfaction rates with the Franklin Care Management Program were extremely high (95%) at baseline, thus reducing the chances that a statistically significant difference (at least in a global satisfaction measure) will be detected. Second, although there may be significant differences between the randomized nursing groups in terms of experience, training and attitudes, their level of education was similar and nurses in both groups had at least 8 years of prior nursing experience. Third, biases may result from nonrandomization of individual patients.

New directions This project has the potential to affect the delivery of care significantly near the end of life because it exploits a preexisting and successful structure for delivering care to the sickest and most complex patients. Because of the steady growth of case management as a means of caring for complex-high risk patients while controlling resource utilization, this model is the logical place to apply the well-developed processes of palliative care assessment, while utilizing the resulting data to stimulate and maintain the ongoing involvement of the treating physician in the care of the patient. This model does not require fundamental changes in the reimbursement structure of medicine, nor in the venue where care is delivered, and could be locally adapted and widely implemented in a short period of time. If, as we anticipate, this project yields solid evidence of both feasibility and effectiveness in improving patient outcomes, the integration of case management with palliative care assessment and feedback to treating physicians could become an important approach to community care of the seriously ill outside of formal hospice programs. The project is innovative because of its simplicity and potential for broad applicability, utilizing a


widely available programmatic approach enhanced by palliative care.

ACKNOWLEDGMENTS This work was supported by Promoting Excellence in End-of-Life Care, a National Program Office of The Robert Wood Johnson Foundation. Drs. Meier, and Morrison are Faculty Scholars of the Project on Death in America, New York. Dr. Meier is recipient of an Academic Career Leadership Award (K07 AG00903) from the National Institute on Aging, Bethesda, Maryland. Dr. Morrison is the recipient of a Mentored Clinical Scientist Development Award (K08 AG00833) from the National Institute on Aging, Bethesda, Maryland, and is a Paul Beeson Faculty Scholar in Aging Research.

REFERENCES 1. Emanuel EJ, Fairclough DL, Slutsman J, Alpert H, Baldwin D, Emanuel L: Assistance from family members, friends, paid caregivers, and volunteers in the care of terminally ill patients. N Engl J Med 1999; 341:956–963. 2. Thar W. Cost of health care for complex case management clients. Franklin Health Incorporated, 1999. ,www.franklinhealth.com. (Last accessed May 15, 2003). 3. Pacala J, Boult C, Hepburn K, Kane R, Kane R: Case management of older adults. J Am Geriatr Soc 1995; 43:538–542. 4. Boult C, Kane RL, Brown R: Managed care of chronically ill older people: The US experience. BMJ 2000; 321:1011–1014. 5. McCorkle R, Strumpf NE, Nuamah IF, Aaler DC, Cooley ME, Jepson C, Lusk EJ, Torosian M: A specialized home care intervention improves survival among older post- surgical cancer patients. J Am Geriatr Soc 2000;48:1707–1713. 6. Nordmann A, Heilmbauer I, Walker T, Martina B, Battegay E: A case-management program of medium intensity does not improve cardiovascular risk factor control in coronary artery disease patients: The Heartcare I trial. Am J Med 2001;110:543–550. 7. Boult C, Boult L, Pacala JT: Systems of care for older populations of the future. J Am Geriatr Soc 1998; 46:499–505. 8. American Geriatrics Society: Care management position statement. J Am Geriatr Soc 2000;48:1338–1339. 9. HMO Work Group on Care Management: Essential components of geriatric care provided through health maintenance organizations. J Am Geriatr Soc 1998;46: 303–308.

133 10. National Hospice Organization: Family Satisfaction Survey. National Hospice Organization, 1997. ,www. nhpco.org. (Last accessed May 15, 2003). 11. National Hospice and Palliative Care Organization: Stats/Research/Resources. Alexandria, VA; 2001. ,www.nhpco.org. (Last accessed May 15, 2003). 12. Keay TJ, Fredman L, Taler GA, Datta S, Levenson SA: Indicators of quality medical care for the terminally ill in nursing homes. J Am Geriatr Soc 1994;42: 853–860. 13. Keay TJ, Taler GA, Fredman L, Levenson SA: Assessing medical care of dying residents in nursing homes. Am J Med Qual 1997;12:51–56. 14. Keay TJ, Schonwetter RS: The case for hospice care in the long term care environment. Clin Geriatr Med 2000;16:211–223. 15. Meier DE: Planning a hospital-based palliative care program: A primer for Institutional leaders. In: Center to Advance Palliative Care, eds. Technical Assistance Series 1. New York: 2000. ,www.capc.org. (Last accessed May 15, 2003). 16. Morpace Customer Satisfaction Telephonic Survey. ,www.morpace.com/customer-satisfaction.htm. (Last accessed May 15, 2003). 17. Morrison RS, Siu AS, Leipzig RM, Cassel CK, Meier DE: The hard task of improving quality measurement at the end of life. Arch Intern Med 2000;160:743–747. 18. Meier DE, Morrison RS, Cassel CK: Improving palliative care. Ann Intern Med 1997;127:225–230. 19. Mor V, Masterson-Allen S: A comparison of hospice vs conventional care of the terminally ill cancer patient. Oncology 1990;4:85–91; discussion 94, 96. 20. Du Pen SL, Du Pen A, Polissar N, Hansberry J, Kraybill BM, Stillman M, Panke J, Everly R: Implementing guidelines for cancer pain management: Results of a randomized controlled clinical trial. J Clin Oncol 1999;17:361–370. 21. O’Boyle CA, Waldron D: Quality of life issues in palliative medicine. J Neurol 1997;244(Suppl 4):S18–25. 22. Medicare Payment Advisory Commission: Improving care at the end of life. Report to the Congress: Selected Medicare Issues. Washington, D.C.: 1999, pp. 117–132. 23. Institute of Medicine. Approaching Death: Improving Care at the End of Life. Washington, D.C.: National Academy of Sciences, 1997. 24. Franklin Health: Program Outcomes. ,www. franklinhealth.com, pro_out.htm. (Last accessed May 15, 2003). 25. Wrede-Seaman L: Symptom management algorithms for palliative care. Am J Hosp Palliat Care 1999;16: 517–526. 26. Wrede-Seaman L: Symptom Management Algorithms for Palliative Care. Yakima, WA: Intellicard, 1996. 27. Byock IR, Merriman MP: Measuring quality of life for patients with terminal illness: the Missoula-VITAS quality of life index. Palliat Med 1998;12:231–244. 28. Teno JM, Okun SN, Casy V, Welch LC: Toolkit of Instruments to Measure End of Life Care (TIME). Resource Guide: Achieving Quality at Life’s End.

134 ,www.chcr.brown.edu/pcoc/toolkit.htm. (Last accessed May 15, 2003). 29. Chang VT, Huang S, Feuerman M: Validation of the Edmonton Symptom Assessment Scale. Cancer 2000;88:2164–2171. 30. Medical Audio Visual Communications, Inc., Toronto, Ontario: 1998. ,www.mavc.com. (Last accessed May 15, 2003).

MEIER ET AL.

Address reprint requests to: Diane E. Meier, M.D. Box 1070 Mount Sinai School of Medicine New York, NY 10029 E-mail: [email protected]