Improvements in 1-Year Cardiovascular Clinical Outcomes Associated with a Hospital-Based Discharge Medication Program

Context Despite evidence to support the effectiveness of a variety of interventions for the secondary prevention of cardiovascular disease, many eligible patients do not receive these interventions. Contribution Compared with a baseline period, patients hospitalized with cardiovascular disease who received an intervention that focused on discharge medications had higher rates of prescription of aspirin, -blockers, statins, angiotensin-converting enzyme inhibitors, and warfarin at hospital discharge. The risk for death and readmission was lower in the intervention period than in the baseline period. Cautions The prepost design of this study does not permit conclusions about a causal relationship between the intervention and the observed improvements. The Editors Cardiovascular disease remains the most common source of morbidity and mortality in western countries (1). During the past 2 decades, a variety of significant medical advances have been made in the treatment and prevention of complications associated with major cardiovascular disorders, including coronary artery disease, congestive heart failure (CHF), and atrial fibrillation. Many of these advances in secondary prevention relate simply to the appropriate use of certain medications, for example, aspirin and statins for coronary artery disease, -blockers and angiotensin-converting enzyme inhibitors for CHF, and warfarin for prevention of stroke and other embolic events in patients with atrial fibrillation. Each of these medical therapies has been proven in large, multicenter, randomized, double-blind, placebo-controlled trials to save lives under appropriate circumstances (2-7). However, despite the overwhelming evidence supporting the benefit of these medications, many studies have documented a significant treatment gap. Often, far fewer than 50% of potential beneficiaries actually receive treatment (8-13), and studies from our institution and others have shown that many patients do not receive indicated medications at hospital discharge (14, 15). On the other hand, the chances of long-term adherence are significantly higher when medications are provided at hospital discharge, and this difference is associated with decreased mortality rates (13, 16). Hospital discharge is a critical juncture in the process of care. Patients are available for consultation and are often more disposed to adopt health care recommendations. A hospital dischargebased intervention could be more easily implemented, more effectively managed and measured, and more cost-effective than other potential outpatient intervention strategies. Consequently, there is substantial justification for focusing on hospital discharge as the ideal time to improve prescription of medications for secondary prevention of cardiovascular disease. In the current study, our objectives were to 1) determine the feasibility of implementing a comprehensive quality improvement initiative in a large multihospital network to significantly increase rates of prescription at discharge of evidence-based, secondary prevention medications for life-threatening cardiovascular diseases; 2) ascertain the effect of such a program on long-term clinical outcomes; and 3) determine whether good adherence to such an initiative is sustainable. Methods Intermountain Health Care (IHC) is a nonprofit, integrated health care system including 20 hospitals, a system of health plans, and both employed and affiliated physicians. The approximately 400 employed physicians mainly practice primary care, and the approximately 2500 affiliated physicians are mostly specialists. Intermountain Health Care serves approximately 60% of the population of Utah and southern Idaho. The system's 10 largest hospitals were included in the quality improvement initiative, accounting for more than 90% of the total population of cardiovascular patients. The major intent of this initiative was to ensure that appropriate secondary prevention medications were prescribed at discharge to all patients (excluding those with documented contraindications) who were hospitalized with a principal cardiovascular diagnosis of acute myocardial infarction (MI), coronary heart disease (CHD), CHF, or atrial fibrillation. Table 1 shows the guidelines that were developed for each of the diagnostic categories, and Appendix Figure 1 shows the reference card that was developed to aid physicians and clinical staff. All of the guidelines were based on the current class IA American Heart Association/American College of Cardiology recommendations, except for the recommendation of a statin in all patients with CHD, including those with a low-density lipoprotein cholesterol level less than 2.59 mmol/L (<100 mg/dL). Our research (17) showed that a statin provided significant benefit in this population, a notion that has since been supported by a large clinical trial (18). -Blocker therapy for patients with CHF was deferred until after discharge. Table 1. Discharge Medication Guidelines for the Secondary Prevention of Cardiovascular Disorders Design and Implementation of the Quality Improvement Program In 1998, before the new discharge medication program (DMP) was initiated, a new institution-wide database was developed to assist in its implementation and long-term management. Simultaneously, all participating institutions began prospectively tracking prescription of applicable discharge medications for all cardiac patients discharged from their respective cardiovascular departments (Appendix Figure 2). The DMP was implemented on 1 January 1999, and tracking of discharge prescriptions continued. Maintenance of the discharge medication database is ongoing, and information through February 2002 was included in this study. Appendix Figure 2. Data collection form for the discharge medication program. Implementation of the DMP required support and commitment from every level of the health care system, including administration, physicians, nurses, and other staff. To gain this support, the IHC Cardiovascular Clinical Program leadership conducted an extensive education campaign in all participating hospitals. Before the actual initiation of the DMP, this group conducted an extensive tour to engage physicians, residents, and nurses with the program, its goals, and its rationale. The actual DMP implementation and documentation process varied from hospital to hospital but always included the essential core elements. The appropriate indication for each medication was printed directly on the patient discharge form, so physicians needed only to check the correct box or record the specific contraindication (Appendix Figure 3). This form served as a project management and data collection tool. When an appropriate medication was not prescribed, the discharge-planning nurse contacted the attending physician or resident directly, after which the missing medication could be added to the discharge prescriptions or an appropriate contraindication for its use could be documented. All information was entered into a computerized database for tracking. Other than the additional resources for data management, the DMP used existing hospital personnel. No organized effort was made to guarantee long-term adherence to prescribed medications after discharge. Appendix Figure 2. Discharge orders form.. Documentation of Short-Term Success of the Program To measure and manage the short-term success of the program, monthly reports were generated showing the proportion of cardiovascular patients discharged with prescriptions for the indicated secondary prevention medications at each of the 10 hospitals. These reports were shared extensively with all participating health care providers from each contributing institution in an effort to further increase adherence. Long-Term Follow-up To monitor the effect of DMP implementation on long-term clinical outcomes, the baseline characteristics (age, sex, and discharge diagnosis) of each cardiovascular patient discharged from the 10 participating facilities were compiled from the centralized IHC informatics database. Each patient was followed prospectively for up to 1 year for hospital readmission due to a cardiovascular indication or for death. If no evidence of hospital readmission or death could be found for a patient, we assumed that the patient had not experienced either event. Information was included from 1996, 3 years before DMP initiation, to February 2002, more than 3 years afterward. Rehospitalizations (cardiovascular only) were determined through the IHC informatics system's electronic data warehouse, and death was ascertained by using the IHC system and the Social Security Administration's death records. Such electronic follow-up, in our experience (16), has proven more thorough and accurate than telephone surveys. Although a limited proportion of patients may have been readmitted to other hospitals for subsequent care, they were probably randomly distributed between the pre-DMP and DMP samples. Characteristics of patients admitted before (1996 to 1998) and after (1999 to 2002) DMP implementation were evaluated separately and compared. Statistical Analysis The chi-square test was used to evaluate differences in the proportions of patients receiving appropriate discharge prescriptions in the pre-DMP and DMP groups. To evaluate the effect of the DMP, we formed 4 mutually exclusive diagnostic or procedural categories for patients with the following index admissions: CHD without CHF, MI, or coronary artery bypass grafting (CABG); CHF without MI or CABG; MI without CABG; and CABG. The category of atrial fibrillation was not mutually exclusive of the other 4 categories. These categories reflected different patient samples and were not designed to evaluate groups on the basis of indications for each separate medication, since more than 1 medication could have been indicated in any given patient. Differences between baseline characteristics of the pre-DMP and DMP groups were eva