Focus on Cardiovascular Disease

Improving Member Outcomes With a Congestive Heart Failure Health Management Program*

CHF affects nearly 5 million people in the United States and is the primary or contributory cause of almost 300,000 deaths each year.[5] There is also ample evidence from the literature that, despite major therapeutic advances, the management of CHF is far from optimal.[5]

The health plan featured in this quality improvement activity (QIA) profile found that CHF was the 20th most prevalent medical condition in its commercial population and the third most prevalent condition in its senior population. With further analysis, the health plan determined that a large portion of the $140 million expended annually on CVD care was related to the management of CHF. The high impact of CHF on utilization of services is illustrated by data showing that, among patients with CHF, there were 515 admissions and 3,270 inpatient days per 1,000 members with CHF per year compared with only 35 admissions and 151 inpatient days per 1,000 members per year for non-CHF patients, excluding maternity patients.

According to the literature, one intervention effective in managing CHF was the use of multidisciplinary health management programs. These programs significantly reduced recurrent hospitalizations and utilization of services while improving patient health status, quality of life, and compliance with medications and diet.[6]

All health plan members, who had International Classification of Disease, Ninth Revision codes related to CHF (primary codes 428 and 425) on claim forms submitted between November 1997 and October 1998, had been continuously enrolled in the health plan for 12 months, and were at least 18 years of age, were included in the population for the QIA. The study was longitudinal, so the health plan identified a group of 1,700 patients for the baseline measurement and continued to assess this same cohort with subsequent remeasurements. This approach enabled the health plan to directly assess the impact of interventions on the same group of patients over time. The 1,700 members with CHF identified during the baseline period represented 100% of the eligible population for this study. During subsequent years, additional members with CHF were identified and offered the interventions described in this study, but were not included in this study.

To monitor the impact of the interventions, the health plan chose measures related to hospital use, including admissions, inpatient days, and emergency department (ED) visits (visit rates normalized to use per 1,000 members with CHF per year), as well as the average days’ supply of angiotensin-converting enzyme (ACE) inhibitors and beta-blockers per year (Table 2).

The included hospital measures correlated with enhanced clinical management in the ambulatory setting. Fewer trips to the ED and a reduction in hospital admissions and inpatient days were indications that the patient was not experiencing an exacerbation of CHF and was being managed well in the ambulatory setting. In addition, data were readily available in administrative databases without the need for a chart review.

The average days’ supply of ACE inhibitors and beta-blockers gave the health plan information on patient compliance with medication treatment. The measure was calculated by taking the total number of days of medication dispensed by pharmacists to all patients and dividing the number of patients included in the study.

The ideal average days’ supply would be 365 days per member per year for each medication, if all patients were fully compliant. On the other hand, the number of days’ supply would be less than 365 days per member per year if some patients were noncompliant. The lower the average days’ supply, the poorer the compliance of the group as a whole. At the time that the study was done, there were no available HEDIS® measures related to CHF.

Medical claims/encounter data were used to calculate the inpatient admission rate, inpatient days rate, and ED visit rate, and pharmacy data were used to determine the medication supply. Inpatient admission, inpatient days, and ED visits for any reason (not just CHF-related admissions) were counted. The health plan took this approach because it was felt that CHF always plays a significant role in the patient’s health status, even in situations where the primary reason for the hospitalization was a condition other than CHF. Pharmacy data were checked to confirm that all patients diagnosed with CHF had received a prescription for an ACE inhibitor and/or a beta-blocker. The average days’ supply of medication took into account differences in dosing regimens (e.g., once daily, twice daily).

The health plan measured baseline data from November 1997 to October 1998 for the 1,700 identified patients. Analysis of the baseline data revealed, as expected, a very high rate of hospital and ED use.

An outcomes committee comprising a medical director, a QI manager, a medical projects specialist, a business analyst, and an information technology specialist conducted a barrier analysis based on the baseline data and a review of the published literature. Barriers identified included lack of:

• Patient education on CHF
• Intensive monitoring of symptoms
• Timely provider feedback concerning patients who were experiencing exacerbations of their CHF
• Physician education related to the clinical practice guideline

The barrier analysis was reviewed by a strategic planning committee comprising professionals from medicine, nursing, psychiatry, health education, public health, data management, and compliance. Based on the identified barriers, the strategic planning committee determined the best approach would be implementation of a CHF Health Management Program. Goals set for the intervention were a 5% decrease from baseline or the previous remeasurement for inpatient admission rate, inpatient days rate, and ED visit rate, and a 5% increase for the average number of days’ supply of ACE inhibitors or beta-blockers received by patients with CHF.

The project team felt that this degree of change would be associated with both a statistically and clinically significant result.

The health plan launched the CHF Health Management Program by mailing an information packet to internal medical directors and QI coordinators (N=300) and to all network clinicians who were likely to treat patients with CHF (N=14,000).

In order to match interventions with needs, the health plan stratified patients into high- and low-risk groups. High-risk patients were defined as patients with CHF who had visited the ED in the past year, were hospitalized in the past year, or had a comorbid condition identified based on a prespecified list. Low-risk patients were defined as patients with CHF who had only visited physicians in their offices in the past year and did not have a comorbid condition. In general, interventions in the high-risk group were interactive (e.g., phone calls initiated by the management team), while those aimed at the low-risk group were more informational (e.g., educational mailings).

• Interactive interventions for high-risk patients included:
• Telemonitoring for all high-risk patients by an external disease management vendor (see appendix 1)
• Enrollment in telemonitoring program solicited by contract nonclinical staff (patient contacted first, then physician, because the reverse order could delay intervention and reduce timeliness of care)
• Weekly telephone calls by nurses to track patient weight, shortness of breath, swelling, and chest pain (weight scale provided to patient if needed)
• Physician contacted if patient was decompensating
• Physician intervened as needed to prevent further decompensation
• Postdischarge in-home or telephonic education (about a 50/50 split) by external contract nurses for members recently discharged from the hospital with a CHF diagnosis code
• Treatment compliance
• Dietary restrictions
• Medication schedule
• Informational interventions for high- and low-risk patients included:
• Monthly educational mailings on medication compliance, weight monitoring, nutrition, exercise, and stress management (see appendices 2-6)

Increasing Physician and Patient Cooperation The health plan felt that the biggest challenge it faced with regard to the successful completion of the intervention was the possibility of substantial patient and physician noncompliance with or resistance to the program. When communicating with patients, the health plan emphasized the program’s collaborative nature, reassuring them that it would not interfere with the patient-doctor relationship. In addition, the CHF Health Management Program was carefully designed to be useful for and acceptable to physicians. For example, it was a value-added patient service that required minimal time and involvement on the physician’s part, except when appropriate (eg, when on of their patients with CHF was found to be decompensating). Communications with physicians also stressed the collaborative and worthwhile nature of the program. After a careful implementation period, most physicians and patients accepted the program with enthusiasm.

The first remeasurement, based on data from November 1998 to October 1999 from both high- and low-risk patients, showed improvements that exceeded goals on all measures -- in some instances by a substantial amount (Table 3). Inpatient admission decreased by 44% and inpatient days decreased by 49%, with a decrease in ED visits of just under 7%. The average days’ supply of ACE inhibitors and beta-blockers rose by 15% and 28%, respectively, indicating better compliance with medications known to be effective in treating CHF. All improvements in performance on the measures, except ED visits, were statistically significant.

To determine where interventions could be modified to provide further change, the health plan analyzed first-year results. Although adherence to ACE inhibitors and beta-blockers had improved - and exceeded goals - the measured average days’ supply (101 and 59 days, respectively) still fell far short of the ideal level of 365 days per member. The health plan decided to focus its next set of efforts on education physicians about the importance of these medications in the treatment of CHF, since previous interventions had focused on members. Subsequently, a set of CHF treatment guidelines and practice tools were sent to physicians who were likely to treat CHF patients (see appendix 7).

Cause and Effect How can you increase your level of confidence that the improvements in performance were due to program interventions and not other causes? This health plan decided to measure the same outcomes in a control group (N=649) of patients who were not initially asked to participate in the CHF program due to staggered enrollment. This analysis confirmed that the interventions had impacted the results, as medical utilization (inpatient admissions, inpatient days, and ED visits) in the control group increased, while the medical utilization in the test group decreased. Control Group Test Group Measure Baseline Remeasurement 1 Baseline Remeasurement 1 Inpatient admission rate* 656 773 515 286 Inpatient days rate* 3166 3644 3270 1657 Inpatient days rate* 676 741 552 515 *Per 1,000 members

Quality of Life The health plan also wanted to know if the program was having a positive effect on patient quality of life (QOL). A validated survey that focuses on treatable physical and emotional impairments attributed to CHF (the Minnesota Living With Heart Failure Questionnaire) was administered during the first remeasurement period to high-risk members during the telemonitoring intervention enrollment call (before interventions were implemented) and was readministered after six and 18 months. This analysis revealed that the program interventions were having a positive effect on patient QOL. At six months, patients experienced a statistically significant 31% improvement in the total QOL score compared with baseline, and at 18 months they experienced a statistically significant 36% improvement. Along with the positive impact of interventions on utilization measures, this finding supported the health plan’s decision to continue the program.

The health plan also reviewed the telemonitoring intervention enrollment rate, an important process measure. Anticipating a 60% enrollment rate based on the vendor’s historical call center performance, the plan was disappointed to learn that its enrollment rate was only 31%. A focus group was conducted with the telemonitoring staff and call center management to identify barriers to enrollment. Member barriers were considered to be the most important (Figure 8); two of these member barriers were later validated using quantitative data collected during the enrollment process. To overcome the barriers, contract nurses were used for enrollment instead of nonclinical staff, starting in the second remeasurement period. The nurses had greater freedom to build awareness and problem-solve barriers during the enrollment call.

Figure 8. Analysis of Barriers to Telemonitoring Program Enrollment

After further adjustments and changes in the program, the health plan again checked performance using the measurement set in the 1,457 patients who remained in the program. Based on data from November 1999 to October 2000, the second remeasurement showed statistically significant improvements that met goals for all measures, except ED visits (Table 4). From remeasurement 1 to remeasurement 2, inpatient admissions decreased an additional 9%, inpatient days decreased 8%, ED visits decreased 3%, and the average days’ supply of ACE inhibitors rose 8%, while the average days’ supply of beta-blockers rose 32%.

However, since ACE-inhibitor and beta-blocker use was still well below the optimal 365 days’ supply, the health plan decided to implement several further interventions to try to increase medication use. Physicians who treated high-risk CHF patients or whose patients were not adherent to ACE-inhibitor and/or beta-blocker therapy (N=1,100) received academic detailing on the health plan’s CHF program and nationally recognized CHF treatments. The health plan hoped that a positive physician response to this intervention would result from the practice tools distributed during these visits.

The health plan also found that the enrollment rate for the telemonitoring intervention increased from 31% to 51% as a result of the revised enrollment procedure. Seeking to enhance enrollment even further, the health plan modified the telemonitoring database to enable quantification of the precise reasons for member refusal of high-risk intervention. In addition, members who refused weekly telemonitoring calls were offered the option of monthly telemonitoring calls, which addressed the patient-cited barrier of inconvenience.

As a result of these additional efforts, the health plan continued to see improvement during the third remeasurement period in 1,054 patients who remained from the original cohort (November 2000 through October 2001, Table 5). From remeasurement 2 to remeasurement 3, inpatient admission decreased another 8%, inpatient days decreased 6%, ED visits decreased 11%, the average days’ supply of ACE inhibitors rose 1%, and the average days’ supply of beta-blockers rose 12%. All changes met goals, except for average days’ supply of ACE inhibitors, and all changes that met goals were statistically significant, except for the decrease in ED visits.

At the completion of the study, the health plan had achieved statistically significant improvement for all measure except for the ED visit rate (Table 6). From baseline to remeasurement 3, inpatient admissions decreased 53%, inpatient days decreased 56%, ED visits decreased 19%, the average days’ supply of ACE inhibitors rose 25%, and the average days’ supply of beta-blockers rose 89%.

Since the QIA was submitted to National Committee for Quality Assurance (NCQA) for review, the CHF Health Management Program has continued but has undergone some changes. The health plan now conducts only telephonic discharge education, citing the difficulty of administering in-home education, the delay in getting a nurse out to the patient, and a lack of compelling evidence that in-home education was superior to telephonic intervention. In addition, the health plan revised its CHF treatment guidelines and made them available on the Internet. To strengthen the educational components of the program and to increase and enhance interactions between nurses and patients, the health plan decided to bring the health management program for CHF in-house. The health plan now employs a large multi-disciplinary staff of nurses, health educators, registered dieticians, exercise physiologists, respiratory therapists, and outreach specialists to deal with all aspects of the members’ conditions, including comorbidities. Although this requires more resources, the health plan believes the return on investment will be positive, and the health plan now has more control over patient interactions. The plan attributes the success of this activity to the careful identification of barriers, including patient need for education and physician need for notification of decompensation, and the successful design and implementation of appropriate targeted interventions. Other factors that contributed to the success of this project were the persistence of the work, the frequent reassessment and setting of new goals, and the analysis of the effectiveness of multiple interventions and adjustments in response to data from carefully selected measures. The health plan is now poised to take the next step in the evolution of its disease management programs, with the imminent launch of a comprehensive CVD management program, which will identify and manage members with hyperlipidemia, hypertension, CAD, and CHF.

Another Plan's Approach Technology in Disease Management Programs Although many health plans successfully use outbound telephonic monitoring by a nurse in their disease management programs, others use sophisticated technology to achieve impressive results. Another health plan used an outside vendor who placed a patented biometric measurement device with an easy-to-use interactive display and communications appliance in the patient’s home or other care setting to gather daily information on weight and symptoms. The scale weighed the patient within one tenth of a pound, and if it wasn’t used for two consecutive days, a cardiac nurse called the patient to determine the reason. After weighing him- or herself, the patient pressed the YES or NO keys on the interactive display in response to audible or visual questions on medication use and symptoms such as shortness of breath. Weight and symptoms were then automatically transmitted via the patient’s regular phone line to the vendor’s cardiac nurses, who reviewed the patient’s past and present weight and symptom information, medical history, comorbidities, past and present medications, and demographic information. Data from all monitored patients were organized in order of patient need, with those patients most in need being presented first. After reviewing the data against parameters set by the physician, the cardiac nurse alerted the physician promptly about any abnormal results. This allowed the physician to intervene before symptoms escalated and hospitalization became necessary. For examples of patient education components that complemented this QIA, please see appendices 8-10.

An excellent example of the type of CVD QIA currently being implemented by many health plans, the QI initiative profiled here illustrates how a health plan acted upon its realization that CHF was not being managed optimally. Through its QIA the plan was able to decrease inpatient admissions, inpatient days, and ED visits, and increase the use of ACE inhibitors and beta-blockers in patients with CHF. These were significant achievements in a difficult-to-manage patient population, and the interventions probably saved lives. Although the plan benefited from the lower resource utilization, the patients with CHF - who experienced a positive effect on QOL as a result of the interventions - were the ultimate beneficiaries.

To achieve optimal cardiovascular care, however, we need to balance our efforts in patients with established disease with an equally strong focus on prevention through risk factor management. When it comes to CVD management, there is still a gap between how risk factors should be managed and how they are being managed today. Having begun the process of improving the quality of cardiovascular care, plans and clinicians are ready to move the quality needle even further by improving the management of overall cardiovascular risk, not just after, but also before, a first cardiovascular event.

[1] - The National Committee for Quality Assurance. The State of Health Care Quality: 2003. Washington, DC: The National Committee for Quality Assurance; 2003.

[2] - American Heart Association. Heart Disease and Stroke Statistics - 2003 Update. Dallas, Tex: American Heart Association; 2002.

[3] - American Heart Association. Risk factors and coronary heart disease: AHA scientific position. Available at: www.americanheart.org/presenter.jhtml?identifier+4726 Accessed August 21, 2003.

[4] - . Sipkoff M. Once shunned, disease management programs are now winning favor. Available at: www.premierhealthcare.com/cgi-bin/article.cgi?article_id=869. Accessed September 4, 2003.

[5] - Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult: executive summary. Circulation. 2001;104:2996-3007.

[6] - Rich MW. Heart failure disease management: a critical review. J Card Fail. 1999;5:64-75.

Return to top