DESIGN: A disease management approach for heart failure based on guidelines from the Agency for Health Care Policy and Research (AHCPR). The key elements in this approach included program preparation, clinical education, protocol implementation, outcomes measurement, and continuous quality improvement. Specially trained consultant pharmacists received approval for the heart failure program from the administration and clinical staff at each long-term care facility and provided in-service programs for the nursing staff. Finally, recommendations were made to attending physicians to improve the pharmacotherapy for heart failure patients within the facility.

SETTING: Skilled long-term care facilities in Ohio.

PARTICIPANTS: Consultant pharmacists, nursing staff, nursing home administrators and clinical staff, attending physicians.

MAIN OUTCOME MEASURES: Hospitalizations for heart failure (primary or contributing diagnosis).

RESULTS: For every $1 invested in this program, $2.44 could be saved in heart failure-related hospitalization costs.

CONCLUSIONS: Results indicate that this program is cost effective in this setting.

KEYWORDS: Clinical guidelines, Congestive heart failure, Disease management.

J Managed Care Pharm 1999: 516-520

In a letter to providers dated November 8, 1995, the Ohio Department of Human Services outlined why heart failure is a major and growing public health problem:

• Congestive heart failure (CHF) is the only major cardiovascular disease whose incidence is increasing.
• In this country, $7 billion is spent annually for CHF hospitalizations.²
• Poor patient compliance with preventive programs and therapy (i.e., diet and exercise) tends to greatly increase morbidity and mortality.
• Significant variations in the treatment of heart failure has led to the need for widely accepted guidelines for the management of this disease.

The American Medical Directors Association (AMDA) re-leased Clinical Practice Guideline for Heart Failure in December 1996. Its goal was a simpler, user-friendly guideline that focused on application in the long-term care institutional setting.³

At the time the AHCPR guidelines were issued, NCS HealthCare was an independent provider of pharmacy services to approximately 25,000 patients in the Midwest, primarily in Ohio. Because CHF is a significant problem for its primarily elderly patient population, NCS decided to develop and implement a comprehensive, interdisciplinary disease-management program that would include education of the patient-care team, treatment recommendations to improve therapy, monitoring of clinical and economic outcomes, and continuous quality improvement. This process was developed to be driven by consultant pharmacists working with the patient-care team in each facility.

Plan Development

Next, NCS established criteria for patient inclusion in the program. Any patient with CHF, history of myocardial infarction, and/or new symptoms of systolic heart failure � left ventricle (LV) enlargement; ejection fraction <40%; and physical signs and symptoms of heart failure, including orthopnea, edema, and nocturnal dyspnea � was included. NCS decided to exclude patients with known contraindications to ACE inhibitor therapy such as serum creatinine >3.0 mg/dL, serum potassium >5.5 mEq/L, and systolic blood pressure <100 mm Hg. The program's primary objective was to use a disease-focused approach based on consultant pharmacist recommendations in the treatment of CHF to reduce hospitalizations related to heart failure.

A timeline was developed beginning with pharmacist education. NCS thought it essential to train consultant pharmacists not only about clinical guidelines but also how best to communicate this information to other health care professionals. Another important part of this pharmacist education program is periodic competency testing to ensure a uniform level of knowledge and skills. Information about the proposed program is then communicated by each pharmacist to the nursing facilities they serve during quality assurance meetings and directly to the administrator, director of nursing, and medical director.

The next step was to develop an in-service training program for nurses, nurses' aides, and other direct care providers. This program included practical information related to the dis-ease process, patient assessment, drug therapy, dietary concerns, and desired goals (outcomes) of treatment. The program was approved for one contact (0.1 CEU) for nurses by the Ohio Nurses Association and presented by consultant pharmacists.

The primary goal of physician education is to provide information about the AHCPR guidelines and their clinical rationale. NCS also addressed the specific educational needs of attending physicians who are primarily trained in family practice or internal medicine.

An important phase in program development involved creating intervention recommendations to be used by consultant pharmacists to request diagnostic testing, laboratory monitoring, and drug therapy additions or changes. Each recommendation was thoroughly researched and referenced. Experience showed that legible, concise, and referenced recommendations have a much higher physician acceptance rate. Finally, these recommendations were reviewed by an interdisciplinary team for accuracy and appropriateness for the patient population.

Outcomes goals and measurement instruments were developed that were easy to collect and interpret. The primary purpose of any outcomes measure should be to demonstrate that the intervention or therapy has value. NCS chose to determine value from the payor perspective. The primary goal was to reduce hospitalizations related to heart failure, especially rehospitalizations. A reduced hospitalization rate would have obvious value for both the payor and the facility by improving census and for the patient by improving quality of life and maintaining function.

The Ohio Department of Human Services is the primary payor for about 60% of the residents NCS serves in Ohio. In a medical assistance letter mailed to all Ohio physician providers in November 1995, the department stressed the need for widespread adoption of the AHCPR guidelines. This letter provided treatment guidelines and a drug therapy algorithm for the treatment of congestive heart failure.² Of note was the department's willingness to pay for diagnostic testing (including screening echocardiograms) and drug therapy that has been shown to be cost effective in such trials as SAVE,⁴ SOLVD,⁵ and CONSENSUS.⁶

The final step was to develop a mechanism to complete the disease management loop by using outcomes data to provide a continuous quality improvement process. By knowing which parts of the process were the most cost effective, NCS could concentrate its finite resources in those areas.

Methodology

Implementation
NCS consultant pharmacists were thoroughly trained and provided with the resources needed to go into nursing facilities and train other members of the patient-care team. These pharmacists provided in-service training in each client facility to the direct care staff, including nurses, nurses' aides, therapists, and others. These training programs emphasized the symptoms and physical findings related to heart failure. Early detection and treatment of volume overload is a critical step in reducing hospitalization related to heart failure.

Nine specific, referenced recommendations or interventions for heart failure detection, diagnosis, monitoring, and treatment, based primarily on the AHCPR guidelines (see Table 1), were provided to the consultant pharmacists for use during their drug regimen review process. Physicians were encouraged to communicate their rationale for not accepting the recommendations, to help the pharmacist and nursing staff understand the physician perspective, and to help target patients who might benefit the most from future recommendations (i.e., patients with comorbidities, such as chronic renal disease, or other contraindications to a certain drug therapy recommended for CHF). Consultant and dispensing pharmacists also shared the goal of preventing drug-drug and drug-disease interactions, as might be the case with the addition of a potassium-sparing diuretic to a regimen that already included an ACE inhibitor and a potassium supplement.

Outcome Measurement
NCS consultant pharmacists collected data on physician acceptance or rejection of pharmacist recommendations, direct clinical outcomes, and hospitalizations identified in the discharge record as heart-failure related. The direct clinical outcomes include physical findings, physical symptoms, adverse reactions, and changes in diagnostic testing (see Table 2). This information was collected from the clinical record, nursing and patient interviews, and hospital admission and discharge records. Hospitalizations were recorded over a six-month period following the acceptance or rejection of the pharmacist recommendation.

Surrogate outcomes markers are indirect measures of effect often tied to changes in treatment, cost of care, and quality of life. These factors include changes in drug therapy (including oxygen), changes in activities of daily living, exercise tolerance, and hospitalizations (see Table 3). In the long-term care setting this information is readily available as part of the care plan and Minimum Data Set (MDS). These surrogate endpoints are markers of a clinical outcome. Hospitalization obviously reflects a serious negative clinical event such as pneumonia. However, pneumonia as an acute disease is difficult to value, whereas a hospitalization related to pneumonia can be readily assigned a cost. Similarly, in a long-term care facility, there is a definite assignable cost associated with residents' inability to feed themselves (i.e., the salary and benefits of the person who must feed them).

The selection of those surrogate markers that have high predictive value based on sound clinical trials is important. The addition of a drug known to have benefit in heart failure patients (i.e. an ACE inhibitor) or an increase in the dose of such a drug toward target daily doses would be such a marker.

Results

The risk reduction achieved in this program means that for every 11 patients treated, one hospitalization related to heart failure can be avoided. NCS estimated a cost per intervention of $205, which includes pharmacists' labor and drug costs for a six-month period. Thus, the total cost to prevent one hospitalization is $2,225 ($205 per intervention x 11 interventions). At an average direct cost per hospitalization of $5,500, based on 1998 data from the Health Care Financing Administration, the NCS program yields a benefit/cost ratio of 2.44 to 1, or $2.44 in hospital costs avoided for every $1 spent on intervention.

The primary reason for physician rejection, when noted, was that the patient was stable and additional therapy was not needed. The typical recommendation to add an ACE inhibitor called for lisinipril 2.5 mg daily. If the patient was already on an ACE inhibitor, the typical recommendation was to increase the dose of the current therapy by 50%�100% (i.e., captopril 12.5 mg bid to 12.5 mg tid or 25 mg bid). This therapy was generally well tolerated with a dropout rate of <5%.

Discussion and Limitations
The lack of comparative group demographics presents one limitation in this study's results. Was the recommendation-rejected group sicker than the other group and therefore more likely to have an increased rate of hospitalization? The NCS program instructs the pharmacists to identify those patients most likely to benefit from a change in therapy. Often these patients suffer unstable heart failure as indicated by recent hospitalization or fluid overload requiring physician intervention. It is surprising to note that physicians most often rejected these recommendations because they considered their CHF patients stable, with no need to change therapy.

Continuous Quality Improvement
Disease management is the process of improving clinical outcomes. Therefore, disease management could also be referred to as "outcomes management." In order to manage and improve outcomes, the process must be continually monitored and revised. This effort requires an active intervention to improve physician response. The facility quality assurance team and especially the medical director can be invaluable assets in this effort. Their buy-in to the disease management process and guidelines can give a local legitimacy similar to that provided by specialist committees in acute care. Long-term care facilities also are very interested in global facility outcomes data to show the value of their care. Positive outcomes related to specific diseases provide the facility with a significant marketing advantage over a facility with outcomes data that are not conclusive or, worse yet, do not exist. From the facility perspective, predictable and measurable outcomes have definite value.

Conclusion

Education and training are critical to provide the rationale for the disease management process. Provider acceptance of the guidelines and treatment algorithms is essential to improving patient outcomes. Education and training must also be ongoing. New developments must be communicated to the front-line practitioners in a timely manner.

For physicians, facility staff, patients, and their families to accept a disease management program, the development and maintenance of trust is necessary. Sound recommendations based on accepted clinical guidelines help to ensure that trust. Programs in which one player wins at the expense of the others will not succeed. Successful disease management programs must always be a win-win proposition for all involved.

With the success of these early efforts, NCS is currently implementing this heart failure management program in all its client facilities, which serve more than 220,000 residents in 32 states. Although this task will be formidable, the basic principles of planning, implementation, outcomes measurement, and continuous quality improvement remain the same.

  References