Evaluation of Collaborative Therapy Review to Improve Care of Heart Failure Patients

, , , ,
The American Journal of Managed Care, October 2014, Volume 20, Issue 10

We implemented and evaluated a collaborative therapy review process aimed at optimizing heart failure therapy among patients managed by their primary care providers.



As more demands are placed on primary care providers, new innovative models are required to optimize heart failure (HF) care. The purpose of this study was to evaluate a collaborative therapy review (CTR) program that was implemented to improve guideline-based therapy among HF outpatients.

Study Design and Methods

We screened patient lists of 18 PCPs at the Portland Veterans Affairs Medical Center to identify patients with an ICD-9 code for HF. The charts of patients with ejection fractions (EFs) <40% were then abstracted in more detail. The CTR team reviewed each patient and provided specific guideline-based recommendations. The team then gave specific recommendations to providers through the electronic medical record system. We categorized recommendations relating to drug or device therapies, or need for laboratory testing, and calculated provider acceptance rates by recommendation type.


Of the 641 patients reviewed, 156 patients had detailed chart reviews. We found opportunities for improvement in care in 70 (45%) patients who received 100 recommendations. Among the 100 recommendations, 62 (55%) were for guideline-based drugs, 12 (17%) were for consideration of device therapy, and 26 (24%) were to update lab tests or echocardiograms. Eighty percent of the recommendations were acted on within 90 days.


The CTR program was able to facilitate guideline-based management for HF patients by identifying treatment gaps and making specific guideline-based recommendations to PCPs. While further evaluations are needed, this approach may serve as an efficient method of leveraging the expertise of specialty-trained clinicians to optimize patient care.

Am J Manag Care. 2014;20(10):e425-e431

  • Primary care providers manage the majority of heart failure (HF) patients; efforts to improve care should target the primary care delivery system.
  • Leveraging the expertise of specialty-trained clinicians can help in the management of HF patients without placing major new demands on busy primary care providers.
  • The collaborative medication review process can improve the quality of patient care by maximizing resources.

While there have been many therapeutic advances in treating heart failure (HF) that have decreased morbidity and mortality, HF continues to be a major public health burden that consumes significant resources. Treatment of HF requires intensive medical, dietary, behavioral, and lifestyle modification to achieve improved quality of life, fewer hospitalizations, and decreased mortality.1 Evidence from randomized clinical trials has clearly demonstrated the benefits of angiotensin converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), beta-blockers, aldosterone antagonists, implantable cardioverter-defibrillators (ICDs), and cardiac resynchronization therapy for improving survival in patients with left ventricle systolic dysfunction.1 Despite the strength of evidence for these therapies, implementation is still suboptimal, particularly for aldosterone antagonists and device therapy (ie, cardiac resynchronization and ICD).2,3

In the US healthcare system, the collaborative care model represents an evolving philosophy that integrates the expertise of healthcare providers across different disciplines to provide efficient high-quality patient care. The accumulating body of evidence supports the theory that collaborative care models can overcome treatment gaps for patients with chronic diseases.4-8 At the Portland Veterans Affairs Medical Center (PVAMC), we have developed a collaborative care model to optimize HF management. Under this model, a collaborative therapy review (CTR) team was formed that consisted of HF specialty-trained clinicians, including a pharmacist, cardiologist, hospitalist, and senior internist from the Primary Care Division of PVAMC. These providers regularly staffed the HF clinic and thus had extensive HF experience.

Our methods were patterned after a recently published study that employed a collaborative care model to optimize care of patients with ischemic heart disease.6 Specifically, we implemented a CTR program to review charts of outpatients with HF and an ejection fraction (EF) ≤40%. Our objective was to increase the proportion of patients receiving guideline-directed therapy by evaluating all medications and monitoring parameters that could potentially influence the ability to titrate or initiate life-prolonging medications. After assessment, recommendations were made for the initiation or up-titration of life-prolonging medications, updating laboratory values, and/or referral for ICD. In this study, we report the prevalence of treatment gaps identified by the CTR team, describe the recommendations made, and document the proportion of recommendations accepted by providers in addressing these gaps.

METHODSStudy Patient Population

The study was conducted in an academically affiliated PVAMC healthcare system. The PVAMC Institutional Review Board approved this investigation as a quality improvement and assessment study. At PVAMC, an outpatient heart failure clinic has been in operation since 2006. The clinic is staffed by a team that consists of a cardiologist, HF nurse practitioners, and an HF-trained pharmacist. Since its inception, the clinic has offered specialized training in HF management to PVAMC primary care providers (PCPs). The PCPs participate in a 3-month HF practicum in which they attend weekly half-day clinics and receive individualized one-on-one teaching and practical experience in HF management.


The CTR process started in August 2009 and continued until August 2010. At the time of the study’s initiation, 18 PCPs had participated in the 3-month practicum. Since these PCPs knew the HF clinic staff and were supportive of its quality improvement goals, the CTR team asked to review the charts of their patients. All 18 PCPs assented to this review. Consequently, each of their patient panels was searched for any patients who had received a treatment at the clinic for an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code for HF in the past 2 years (n = 641). The team pharmacist then reviewed each patient’s chart and those with an EF ≤ 40% were included in the sample targeted to receive CTR. Patients who had died (n = 51), had no EF recorded (n = 62) in the VA records or scanned documents, or were no longer under the participating PCP’s care (n = 10) were excluded. Of the remaining patients, 173 (34%) had an EF ≤40%. Because 17 of the 173 patients with EF ≤40% were already being actively followed by the HF clinic, they were also excluded. Thus, the CTR intervention group included 156 patients (30%) ().

Collaborative Therapy Review

To conduct the CTR, a team was formed, consisiting of the HF clinic cardiologist and pharmacist, a hospitalist who saw newly discharged HF patients in the clinic, and a senior internist from the Primary Care Division, who had oversight over the PCPs and worked regularly in the clinic. To assist with the initial detailed chart extracting and review, 3 to 4 rotating PharmD students also contributed to the team. Extracted data included at least 1 year of trends in vital signs and serum creatinine levels and 5-year trends in serum potassium levels. Both Veterans Affairs (VA) and any available non-VA medical records were reviewed for any documented EFs. Patients who had not had an EF within the previous 3 years received an echocardiogram or multigated acquisition (MUGA) scan recommendation. Information about medications and active medical problems were also collected.

Appendix A

In assessing each patient’s optimal HF treatment, the CTR team considered 3 key elements: 1) Can new life-prolonging medications be initiated safely? 2) Can lifeprolonging medications be up-titrated safely to target doses? and 3) Is the patient a candidate for a primary prevention ICD? We developed a standardized data collection tool to extract data (, available at www. ajmc.com). The team assessed the possibility for initiation or up-titration of beta-blockers, ACE inhibitors, ARBs, and spironolactone using vital signs and laboratory data to ensure that the patients were likely to tolerate the new doses without complications. Target doses of life-prolonging medications included lisinopril 40 mg daily, metoprolol succinate 200 mg daily, carvedilol 25 mg twice daily for patients weighing <85 kg and 50 mg twice daily for patients weighing >85 kg, and spironolactone 25 mg daily. If the chart reflected that a patient was taking the highest dose tolerated without side effects, the patient was classified as being at the target dose.

Appendix B

The team pharmacist then made recommendations to the patient’s PCP via a brief note in the electronic medical record. When applicable, pre-prepared electronic orders were also included; the PCPs could choose to accept these orders based on clinical judgment (see , available at www.ajmc.com). The number of recommendations was deliberately kept to a minimum and staggered over time to distribute the PCPs’ workload evenly.

Data Analysis

We summarized patient demographics, comorbidities, and medication utilization using frequencies. We classified recommendations made to PCPs into 3 general categories: guideline-based medications, laboratory testing, or device therapy consultation. Medication and laboratory recommendations were then subclassified to describe drug and laboratory test categories (eg, ACE inhibitors, ARBs, a basic metabolic panel, vitals). The frequency of PCP acceptance of CTR recommendations was calculated overall and stratified by recommendation class. We also assessed the time taken to accept these recommendations. All analyses were performed using SAS version 9.2 (Cary, North Carolina).


The team identified 641 patients as potential targets for CTR based on ICD-9-CM codes for HF in the past 2 years. This represented 3% to 6% of each PCP’s patient panel. A total of 18 panels were reviewed. On average, 10 patients per panel received a detailed chart review and each review took approximately 30 to 40 minutes. Overall, the extraction time was about 6 hours per panel (or approximately 30 minutes per patient), or approximately 100 hours in total. In addition to the time members spent individually reviewing charts, the CTR team met for 1.5 to 2 hours every 1 to 3 weeks to discuss the reviews of approximately 12 to 14 patients and to collectively decide on final recommendations. The CTR team limited the recommendations to 3 to 4 per week to any 1 provider. Overall, the time spent per patient was approximately 45 minutes.

Table 1

The baseline characteristics for the 156 patients are shown in . They were predominantly male (99%) and their median age was 68 years (range: 46-90 years). The primary etiology for patients’ HF was ischemic (72%), and the most common comorbidities were hypertension (81%), coronary artery disease (68%), and diabetes (53%). The majority of patients were receiving a beta-blocker (90%) and an ACE inhibitor (85%) at the time of CTR. However, at least half of these patients had not reached target doses for either of those medication classes (Table 1).

Table 2

Among the 156 patients reviewed, 86 (55%) were found to be on maximum tolerated doses of life-prolonging medications, had current laboratory and EF evaluations, and had either already received or declined ICD therapy. For the remaining 70 (45%) patients, a total of 100 recommendations (mean: 1.4 recommendations per patient; range: 1-4) were made. Among these recommendations, 62 were for guideline-based drugs, 11 for updating laboratory tests, 15 for EF measurements, and 12 for consideration of ICD therapy ().

For guideline-based drugs, 27 recommendations were made for up-titration and 35 for initiation of therapy; the majority of drug recommendations were for lisinopril and metoprolol succinate. For patients receiving up-titration, the median (interquartile range, [IQR] doses for these drugs prior to up-titration were 10 mg (IQR: 3-20 mg) for lisinopril and 50 mg (IQR: 50-100 mg) for metoprolol succinate. After up-titration, the median doses were 20 mg (IQR: 10-28 mg) and 100 mg (IQR: 75-200 mg), respectively. Some of the recommendations for drugs included a 2-step titration to target doses based on patient tolerability. However, the recommendations were considered accepted after the first step titration was performed.

Twelve of the 15 patient recommendations for MUGA, 5 of 11 recommendations for updated labs, and 10 of 12 recommendations for discussion of ICD implants were accepted. Among those who had a recommendation for either an updated MUGA or ICD discussion, 10 had an improved EF >35% and were thus no longer eligible for an ICD, 5 refused an ICD, 2 had an ICD placed, 4 moved and did not have recommendations acted on, and 3 were lost to follow-up. Only 12 (60%) recommendations for beta blockers were accepted. There were 2 recommendations for initiating digoxin in patients who had concomitant atrial fibrillation and had elevated heart rates despite being on target doses of beta-blockers. Overall, 70% of all recommendations were carried out and the majority (80%) were acted on within 90 days.


To our knowledge, this is the first evaluation to be performed of a collaborative chart review and feedback strategy to optimize the care of HF patients managed in an outpatient primary care setting. Our data demonstrated that in 516 ambulatory VA HF patients who had complete data for review, 66% had an EF of 40% or greater and thus were not candidates for further recommendations for life-prolonging therapies because no data exist to support such recommendations in this population of patients. However, among the remaining one-third of patients who had low EFs, substantial opportunities existed to enhance HF management. At our institution we identified 45% of such patients whose HF management could be improved. PCPs in our study accepted and acted on 70% of the CTR team recommendations.

Since the HF clinic has limited specialty clinic availability, the goal was to target a group of untreated HF patients through collaborative care with PCPs. A recent study by Lee et al demonstrated that patients who received assessment by both a primary care provider and cardiologist post discharge were more likely to receive evidence-based medicine, evaluation of left ventricle function, and revascularization. 9 In this study, collaborative care significantly reduced mortality compared with primary care alone.9 Collaborative care was also superior to primary care alone for the composite end points of all cause hospitalization.9 This evidence further supports the contention that collaborative follow-up improves the quality of care and outcomes for HF patients.9 Our collaborative care model identified gaps in HF care and made recommendations. This resulted in improvement of guideline-directed medical therapy.

Previous research suggests that the opportunities for improving HF management identified by our CTR team are not unique to our patient population. While most patients in our study received an ACE inhibitor and a beta-blocker, approximately 50% were not on target doses. Only 27% of patients were receiving target doses of an ARB. This is consistent with prior observations in other patient populations.10,11 Furthermore, implementation of both aldosterone antagonists and device therapy has been slow in clinical practice.2,12 Heywood and colleagues report that only 36% of HF patients are prescribed an aldosterone antagonist in the outpatient cardiology practice setting.11 This is likely due to safety concerns.13 In our study, the CTR team made recommendations pertaining to initiation or dosing of aldosterone antagonists in only 12% of patients reviewed. This relatively low percentage is not surprising, since many PVAMC HF patients had renal insufficiency and hyperkalemia, which made it challenging to initiate aldosterone antagonists.

Treatment of heart failure is an “add-on phenomenon” requiring sequential addition of life prolonging medications. However, in clinical practice the majority of eligible patients may not be adequately treated with target doses of ACE inhibitors and beta-blockers because of intolerance of these medications. Renal dysfunction, hypotension, and hyperkalemia are the most common reasons for not prescribing or titrating ACE inhibitors and ARBs.14 Similarly, titration of beta-blockers is limited by bradycardia, hypotension, and respiratory disease.14 Other reasons, such as patient preferences and nonadherence, are also identified as barriers to optimizing guideline-directed therapy.14

For example, in certain patients (n = 106) up-titration of doses or the addition of beta-blockers and/or ACE inhibitors or ARBs was challenging due to the presence of 1 or more comorbidities, such as hypotension (16%), bradycardia (4%), worsening renal function (54%), and hyperkalemia (45%). The CTR team evaluated at least 5 years of consecutive readings for critical laboratory values and vitals, and if safety was a concern, the team chose not to make drug recommendations. In the most complicated HF cases, the CTR team advised that patients be referred to the HF clinic for further evaluation. Nevertheless, it was encouraging to see a 74% (n = 14) acceptance rate for spironolactone recommendations, which suggested that the CTR process recommended aldosterone antagonist in appropriate patients. In other cases, providers may have acted on only 1 drug recommendation for evidence-based therapy if 2 or more were suggested. This may have been the reason that 40% of beta-blocker recommendations (8 of 20) were not accepted.

We did not ask providers directly why they did not accept 30% of our recommendations; however, the clinical details of patients provide insight. For some patients, a new EF measurement and possible consideration of ICD therapy was suggested. However, after receiving a repeat MUGA, 42% of patients had an improved EF and thus were no longer eligible for an ICD. In other patients who were eligible for an ICD, certain barriers such as cultural beliefs or fear of device therapy and surgical procedures may have hindered acceptance of the recommendations.15

Similarly, 6 of 11 recommendations for updated laboratory values were not acted on. Some of these patients were co-managed by providers outside the VA and may have received their tests at non-VA laboratories. Further, some patients were hospitalized immediately post recommendation and lost to follow-up. For some patients, the PCPs waited until the next scheduled follow-up appointment to get labs.

Our study has several limitations. First, we did not evaluate clinical patient outcomes. However, the intent of the project was not to demonstrate the link between guideline-based use of HF medications and patient outcomes, as this has been clearly established by multiple randomized clinical trials that inform current HF guidelines. Our focus was process improvement, specifically ensuring guideline-adherent care for all HF patients with low ejection fractions. Second, the providers who participated in our study were not randomly selected. The fact that they had previously spent some time in our clinic may have made them more predisposed to accept our recommendations. Nevertheless, with the exception of the ICD discussions, the recommendations made to them were simple and noncontroversial, giving us reason to believe other practitioners who are less familiar with the CTR team would adopt such recommendations at similar rates. Third, the study leveraged the VA’s robust electronic health record system and large patient database, allowing us to perform chart review-based assessments and make therapeutic recommendations relatively easily. Without such an integrated charting and communication system, this project would have been much more difficult, if not impossible. Fourth, the CTR process was time-intensive and required dedicated, experienced personnel to implement. The Portland VA is an academic institution, and as such it was possible to recruit students to assist with a quality improvement process such as the CTR. Additionally, the employees involved in the process utilized protected time designated for scholarship and teaching activities. Not all facilities have such resources to implement this intervention. However, anyone somewhat familiar with medical terminology could be trained relatively easily to use the chart abstraction tool we developed to gather the data necessary to make recommendations regarding therapy. Thus the key to success in other institutions would likely be the commitment of a heart-failure—trained physician to review the data and make the recommendations. In addition, the 45 minutes of time spent on each patient in our study is not insignificant and has to be balanced against the potential gains from such an investment. In our setting these fell into 3 categories: the quality of heart failure care was improved for almost half of eligible patients with minimal impact on the workload of our very busy PCPs; the 45-minute review process avoided the necessity of a 60-minute new-patient appointment in the CHF clinic (a scarce resource in our hospital); and the patients did not need to travel for such an appointment and thus avoided substantial cost and inconvenience. Finally, as our study was performed in the VA population, most of the patients included in the study were male. As such, our findings may be less generalizable to other non-VA healthcare systems.


To date, few studies have identified effective strategies for optimizing utilization of evidence-based therapies for HF patients in an ambulatory care setting. One observational study in outpatient cardiology care of HF showed that strategies such as chart reviews and a system of reminders and educational materials can facilitate implementation of evidence-based care.16 Another study demonstrated that a simple intervention based on electronic medical records, directed to providers, improved the rates of referral for ICD implantation in patients with reduced ejection fraction.17 Our CTR process provided PCPs with simple and direct recommendations that were evidence-based. We found that a simple chart review strategy using a standardized chart extraction tool along with CTR recommendations could be effectively employed to improve the management of these patients.

Patient-centered medical homes are an emerging modality in healthcare, but because this model is primarily built around a robust primary care core, it can increase the workload for PCPs. We believe the CTR process can support PCPs by leveraging the experience of specialty care providers. The structure and concept of our CTR is flexible; the basic mechanism of the process can be used by other healthcare institutions and modified to meet their needs and structures. As these new collaborative modalities of care evolve, employing evaluation processes such as the CTR may help to ensure high-quality patient care and efficient work flow processes for clinicians. Future studies aimed at identifying best practices for optimizing medication management in HF are needed to compare CTR with other systems that facilitate communication with PCPs (such as periodic electronic feedback to PCPs, care management programs aimed at high-risk patients, and PCP education practicums).Author Affiliations: From Oregon State University/Oregon Health & Science University College of Pharmacy, Portland, OR (HS, JCM, SJN, AH); Portland Veterans Affairs Medical Center, OR (GCL).

Source of Funding: None.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (HS, GCL); acquisition of data (HS, GCL, SJN, AH); analysis and interpretation of data (HS, JCM, GCL, SJN, AH); drafting of the manuscript (HS, JCM, SJN); critical revision of the manuscript for important intellectual content (HS, JCM, GCL, AH); statistical analysis (JCM); administrative, technical, or logistic support (HS, JCM, SJN); and supervision (HS).

Address correspondence to: Harleen Singh, PharmD, Oregon State University/Oregon Health & Science University College of Pharmacy, 3303 SW Bond Ave, CH12C, Portland, OR 97239. E-mail: singhh@ohsu.edu.REFERENCES

1. Hunt SA, Abraham WT, Chin MH, et al. 2009 focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines: developed in collaboration with the International Society for Heart and Lung Transplantation. Circulation. 2009;119(x):e391-e479.

2. Fonarow GC, Yancy CW, Hernandez AF, Peterson ED, Spertus JA, Heidenreich PA. Potential impact of optimal implementation of evidence-based heart failure therapies on mortality. Am Heart J. 2011;161(6):1024-1030.

3. Fonarow GC. How well are chronic heart failure patients being managed? Rev Cardiovasc Med. 2006;(7 suppl) 1:S3-S11.

4. Dobscha SK, Corson K, Perrin NA, et al. Collaborative care for chronic pain in primary care: a cluster randomized trial. JAMA. 2009;301(12):1242-1252.

5. Callahan CM, Boustani MA, Unverzagt FW, et al. Effectiveness of collaborative care for older adults with Alzheimer disease in primary care: a randomized controlled trial. JAMA. 2006;295(18):2148-2157.

6. Fihn SD, Bucher JB, McDonell M, et al. Collaborative care intervention for stable ischemic heart disease. Arch Intern Med. 2011;171(16):1471-1479.

7. Chang-Quan H, Bi-Rong D, Zhen-Chan L, Yuan Z, Yu-Sheng P, Qing-Xiu L. Collaborative care interventions for depression in the elderly: a systematic review of randomized controlled trials. J Investig Med. 2009; 57(2):446-455.

8. Pape GA, Hunt JS, Butler KL, et al. Team-based care approach to cholesterol management in diabetes mellitus: two-year cluster randomized controlled trial. Arch Intern Med. 2011;171(16):1480-1486.

9. Lee DS, Stukel TA, Austin PC, et al. Improved outcomes with early collaborative care of ambulatory heart failure patients discharged from the emergency department. Circulation. 2010;122(18):1806-1814.

10. Gheorghiade M, Albert NM, Curtis AB, et al. Medication dosing in outpatients with heart failure after implementation of a practice-based performance improvement intervention: findings from IMPROVE HF. Congest Heart Failure. 2012;18(1):9-17.

11. Heywood JT, Fonarow GC, Yancy CW, et al. Comparison of medical therapy dosing in outpatients cared for in cardiology practices with heart failure and reduced ejection fraction with and without device therapy: report from IMPROVE HF. Circ Heart Fail. 2010;3(5):596-605.

12. Albert NM, Yancy CW, Liang L, et al. Use of aldosterone antagonists in heart failure. JAMA. 2009;302(15):1658-1665.

13. Fonarow GC, Yancy CW, Albert NM, et al. Heart failure care in the outpatient cardiology practice setting: findings from IMPROVE HF. Circ Heart Fail. 2008;1(2):98-106.

14. Steinman MA, Dimaano L, Peterson CA, et al. Reasons for not prescribing guideline-recommended medications to adults with heart failure. Med Care. 2013;51(10):901-907.

15 Al-Khatib SM, Sanders GD, Carlson M, et al. Preventing tomorrow’s sudden cardiac death today: dissemination of effective therapies for sudden cardiac death prevention. Am Heart J. 2008;156(4):613-622.

16. Fonarow GC, Albert NM, Curtis AB, et al. Improving evidence-based care for heart failure in outpatient cardiology practices: primary results of the Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF). Circulation. 2010;122(6):585-596.

17. Gupta A, Gholami P, Turakhia MP, et al. Clinical reminders to providers of patients with reduced left ventricular ejection fraction increase defibrillator referral: a randomized trial. Circ Heart Fail. 2013;7(1):140-145.