Disease management programmes for older people with heart failure: crucial characteristics which improve post-discharge outcomes

Abstract

Aims Disease management programmes (DMPs) have evolved as an innovative clinical practice system to enhance the discharge outcomes of older people with heart failure. Yet, clinical trials which have examined their effectiveness have reported inconsistent findings. This may be explained by variations in the design of DMPs. The aim is to identify the characteristics of DMPs which are crucial to reducing hospital readmission and/or mortality of older people with heart failure.

Methods and results A systematic computerized search was conducted to identify randomized controlled trials of the last 10 years, which examined the effects of DMPs on hospital readmission and mortality of older people with heart failure. The identified DMPs were classified as effective and ineffective, according to statistically significant changes in discharge outcomes. Twenty-one trials were identified, 11 (52.4%) of which reported DMPs improving the discharge outcomes of older people with heart failure. The results indicate that an effective DMP should be multi-faceted and consists of an in-hospital phase of care, intensive patient education, self-care supportive strategy, optimization of medical regimen, and ongoing surveillance and management of clinical deterioration. Cardiac nurse and cardiologist should be actively involved and a flexible approach should be adopted to deliver the follow-up care.

Conclusion This study defines precisely the characteristics of the care team and the organization content and delivery method of the DMP which are crucial to enhance the discharge outcomes of older people with heart failure.

Introduction

Heart failure is a major public health problem: its rising incidence and prevalence render it almost a global epidemic.^1–5 Although advanced pharmacological intervention has improved the prognosis of heart failure,⁶ hospital readmission and mortality rates, particularly for hospitalized elderly patients, remain high.⁷ Acknowledging that such poor discharge outcomes are related to numerous reversible factors, including a lack of understanding of heart failure, inadequate medical prescription, poor treatment compliance, inadequate follow-up care, and inadequate communication between health-care providers,^8,9 disease management programmes (DMPs) have evolved to enhance the care of elderly patients. The European Society of Cardiology (ESC) guidelines recommend the key characteristics of a heart failure DMP as (1) using a team approach, (2) providing in-hospital and out-hospital care, (3) including discharge planning, (4) using education and counselling strategies, which focus on promoting self-care and teaching behavioural strategies, (5) optimizing medical therapy, (6) prescribing flexible diuretic regimen, (7) directing close attentions to clinical deterioration, (8) providing vigilant follow-up, and (9) enhancing access to health care.¹⁰

International clinical guidelines recommend DMPs as a model of best practice to enhance discharge outcomes of heart failure patients,¹¹ but studies examining their effects on hospital readmission and mortality report inconsistent findings.¹² These might be related to the wide variation in the designs and characteristics of DMPs, such as the structure of care team, the components of programme, and the method of delivery, all factors that might hinder the clinical application of this evidence-based intervention to contemporary heart failure. Although some review papers have examined the influence of the design of DMPs on their effectiveness, the analysis was crude and superficial and unable to delineate characteristics crucial to their success.

One review¹³ identified intensive patient education, close follow-up activities, and drug titration as essential features of an effective DMP, but without systematically comparing and contrasting the designs and content of the reviewed DMPs. Indeed, this review only focused on seven randomized clinical trials (RCTs) which adopted a multi-disciplinary approach to programme delivery. Another review conducted a more systematic analysis¹⁴ and classified the heart failure DMPs into effective and ineffective ones, according to statistically significant changes in their outcome measures. Although similar findings were reported, the inclusion of both RCTs and non-RCTs might cause bias, as non-RCTs of DMPs for heart failure patients were found to report an intervention effect twice as great as RCTs.⁸ Two meta-analyses addressed this limitation by only including RCTs,^8,15 but the comparison only focused on the influence of methods of programme delivery and duration of follow-up on the discharge outcomes. No attempt was made to compare the components of effective and ineffective DMPs. The purpose of this review is to identify the characteristics of DMPs which are crucial to reducing hospital readmission and/or mortality of hospitalized elderly heart failure patients.

Methods

Search of the literature

The review included studies which evaluated the effects of DMPs on hospital readmission and mortality of elderly heart failure patients. DMP was operationally defined as a programme that used multiple interventions in a systematic manner to manage heart failure across different health-care delivery systems.^16,17 In view of the rapid advances in the medical management of heart failure and the associated changes in the health needs of the patients,⁷ the review only focused on studies published in the last 10 years (i.e. 1995–2005). A search of MEDLINE (1995–2005 Week 4), EMBASE (1995–2005 Week 10) and the Cochrane Controlled Trial Registry (CCTR) for English-language articles was undertaken. As numerous terms have been used to name DMPs for heart failure, the words ‘cardiac failure or heart failure’ combined with ‘readmission and rehospitalization’ were searched as textural terms and Medical Subject Headings (MeSH). This search strategy has been recommended as a more sensitive method to identify the greatest number of studies.¹² The reference lists of the reviewed articles were also examined.

On the basis of assumption that non-RCTs would introduce greater source of bias to the findings,¹² this review only included studies that randomized the sample and recruited a control group. The studies also needed to (1) involve patients with heart failure, but not mixed samples; (2) include hospital readmission and/or mortality as outcome variables; (3) have sample of a mean age ≥60 years; and (4) provide a detailed description of the DMP. Exclusion criteria included studies which (1) evaluated a single-component intervention; (2) recruited too few subjects to ensure the study power; and (3) reported part of the results of another main publication.

Selection of studies and extraction of data

The literature search identified 2619 articles in MEDLINE (n=1047), EMBASE (n=1045), and CCTR (n=515) and another 12 from secondary references. The titles, which indicated that articles did not pertain to heart failure patients, or were reviews, case reports, editorials, letters, or conference proceedings, were excluded. For the remaining 284 potentially relevant articles, a review of their abstracts further excluded 127 studies as unrelated topics, 45 as reviews or discussion papers, and 13 as focusing on other disease groups. The full text of 99 articles were retrieved, of which 74 were excluded, as they adopted a non-RCT design (n=51), randomized the settings (n=2), had too few subjects (n=2), evaluated a single component intervention (n=14), recruited sample of mean age <60 (n=1) or dwelled in the community (n=1), and reported part or all of the findings of other identified studies (n=3). Thus, this review included 25 papers.¹ However, as four papers presented longitudinal follow-up findings of three other included studies,^29,30,37,39 the review included only 21 RCTs on DMPs.

In order to identify, from the reviewed RCTs, the crucial characteristics of the effective DMPs for heart failure patients, information about the designs of the reviewed DMPs was categorized according to the key characteristics as recommended by the ESC for DMP for heart failure patients. The statistical changes in the outcome variables, which consistently evaluated among the studies, including number of hospital readmissions, mortality rate, or event rate (i.e. combined endpoints of readmission or death), were used to classify the DMPs as effective or ineffective. The designs of the effective and ineffective DMPs were then compared, in order to identify those characteristics crucial to the effective delivery of such programme.

Results

Overview of the reviewed RCTs

The 21 RCTs were published between 1995–2004 and originated in eight countries (Australia, Canada, Ireland, Netherlands, Spain, Sweden, UK, USA). The sample size varied considerably among the studies (n=84–462), with a mean of 212 (SD=107.0), and the number of subjects in the study arms were similar. Table 1 summarizes the baseline characteristics of samples of the reviewed RCTs. The mean age of the samples was 73.3 (SD=4.8), and 76.1% of the RCTs (n=16) recruited<50% of female subjects. Most of the RCTs (80.9%, n=17) reporting information about LVEF: the mean LVEF of the samples of 10 RCTs was 35.9±4.30, and in the other seven RCTs, there was an average of 61.9±17.6% of patients presenting with an LVEF <40%. The NYHA class of the subjects was reported in 16 RCTs, and an average of 61.0±20.5% of subjects were categorized as NYHA Class III or above, indicating that they suffered from more advanced heart failure. Polypharmacy, including anginotensin-converting enzyme inhibitors (15–98%), beta-blockers (19–53.8%), digitalis (8–71%), and diuretics (25–100%), was commonly prescribed. Comorbidities, in particular, hypertension, diabetes mellitus, and coronary artery disease, were prevalent. The demographic and clinical backgrounds of the subjects in interventional arms of the RCTs were well matched.

Table 2 summarizes the characteristics of DMPs. Concerning the care team, the provider designations were explicit in most RCTs, except one study which did not describe the professional background of the case manager.²⁸ Although some DPMs (n=7, 33.3%) adopted multi-disciplinary care^23,31,35 and case management strategies,^19,28,33,38 the others were mainly conducted by nurses, physicians, and/or pharmacists. An in-hospital phase of care was incorporated into half of the reviewed DMPs (n=12). Five DMPs provided patients with structured discharge planning. Patient education related to management of heart failure was almost a unifying component of care of all the DMPs. More than half of the DMPs (n=14) encouraged self-care management by prescribing self-monitor activities (e.g. body weight), providing device (e.g. weighting scale) and/or adherence aid (e.g. pill sorter) to enable the patients to do so. Exercise and psychosocial counselling were least adopted. About half of the reviewed DMPs (n=12, 57.1%) optimized the patients' medication regimen according to clinical guidelines or advice of cardiologists and/ or prescribed patients with flexible diuretic regime (n=7, 33.3%). Fifteen DMPs (71.4%) directed close attention to patients' clinical deterioration by arranging prompt medical consultation or implementing strategies to reduce the symptom manifestations. Post-discharge interventions were mainly conducted through home visit, HF clinic visit, and/or telephone contact. Among the reviewed DMPs, majority of them (n=15, 71.4%) also provided elderly heart failure patients with telephone hotlines to access health-care professionals.

Table 3 summarizes the effects of the DMPs on hospital readmission, mortality, and event rate, as well as on functional status, quality of life, and health-care cost. On the basis of statistical changes on the readmission and/or mortality outcomes, 12 DMPs (57.1%) of the programmes were regarded as effective.^{18,19,24–27,31,33,35,36,38,42} The demographic and clinical characteristics of the samples of the effective and ineffective programmes, including age [t(19)=−0.01, P=0.97], LVEF [t(8)=2.01, P=0.35], NYHA [t(14)=−0.01, P=0.93], use of angiotensin-converting enzyme inhibitor (ACE-I) [t(10)=0.0103, P=0.92], and beta-blockers [t(11)=1.93, P=0.08], were comparable. Effective programmes (n=11) significantly reduced the number of hospital readmissions by 29–85% (mean±SD=44.15±14.36%; RR: 0.56±0.14). Four of these programmes significantly reduced the mortality rate by 28–78% (mean±SD=57.6±21.9%; RR: 0.42±0.22), and eight of them significantly reduced the combined event rate by 38.0±17.3% (RR: 0.61±0.17). Effective DMPs showed more promising effects in improving the quality of life and functional status, and seven out of eight effective programmes (87.5%), which undertook cost analysis, were cost saving.

Comparison between effective and ineffective DMPs

Overall design

Table 4 shows the comparison of the overall designs of the effective DMPs with the ineffective ones. The results indicate that more than half of the effective DMPs (63.6%, n=7) conformed to the ECS guidelines and had majority of the recommended characteristics. These programmes were conducted by cardiac specialists including cardiac nurse and/or cardiologists. They were comprehensive and concomitantly included education, exercise, psychosocial counselling, self-care supportive strategy, optimization of medication regimen, and medical referrals for clinical deterioration. These programmes also adopted both in-person and telephone approaches to deliver follow-up care and provided patients with telephone hotlines to access health-care professionals. The ineffective DMPs were less likely to follow the ESC guidelines. Although some of them did not adequately incorporate the recommended interventions in the programmes,^23,24,34 the others used single approach (i.e. in-person or telephone) to deliver follow-up care.^{20,21,28,32,42} In particular, majority of them (n=8, 88.9%) did not provide the patients with exercise and psychosocial counselling.

Care team

Table 5 outlines the more detailed comparisons of the characteristics of the effective and the ineffective DMPs. Concerning the care team, five effective DMPs (41.7%) adopted either multi-disciplinary approaches (16.7%, n=2) or case management strategies (25.0%, n=3) to enhance the care of heart failure patients. Although the former approach involved active collaboration of various health-care providers such as a specialist nurse, cardiologist, social worker, and dietitian, the later approach relied on a case manager for making critical linkages to multi-disciplinary health-care team. These approaches, however, were only adopted in 22.2% (n=2) of the ineffective DMPs.

For the other DMPs, which did not involve multiple health-care providers in the team, they were commonly delivered by nurses and physicians. Among these DMPs, the effective programmes were more likely to be conducted by cardiac nurses (effective programme: 58.3%; ineffective programme: 33.3%) and involved more active participation of cardiologists. The roles of the cardiologists included, in various combinations, simplifying the pre-discharge medication regime (n=4), participating in care planning (n=1), titrating medications (n=2), providing follow-up care in the clinic (n=3), and developing a comprehensive treatment plan of medications, exercise, and diet (n=1). In the ineffective DMPs, the cardiologists provided more indirect care, which included supervising nurses to deliver care (n=1), developing algorithms for drug titration (n=1), and responding to patients' initiated telephone consultation (n=1). Indeed, Table 5 showed that effective DMPs were more likely to involve both the cardiac nurse and the cardiologist to deliver the programme. Having the primary care physician as a member of the care team (n=2) also appeared to result in more promising discharge outcomes.

The programme intervention

Concerning the programme intervention, Table 5 indicates that the effective DMPs (n=6, 50%) provided a wider range of in-hospital care, despite the ineffective DMPs (n=7, 77.8%) were more likely to incorporate an in-hospital phase of care. In addition to the pre-discharge education as provided by the ineffective DMPs, the effective DMPs^{31,33,35,37,42} also arranged in-hospital counselling by different allied health-care professionals, utilized structured medical input to optimize the patients' clinical condition and medication regimen, and conducted risk assessment for poor discharge outcomes. Discharge planning, in contrast, was seldom incorporated in the DMPs, whether effective and ineffective.

As for patient education and self-care supportive strategies, Table 5 indicates that there was not much difference in their designs and usage between the effective and the ineffective DMPs. The ineffective DMPs were only a bit more likely to prescribe self-monitoring activity to the patients and to use written cues to enhance self-care management. Concerning exercise and psychosocial counselling, even though they were the least adopted among interventions of the reviewed DMPs, 71.4 and 100% of the programmes, which, respectively, incorporated psychosocial care and exercise training, were effective in improving discharge outcomes. Psychosocial care was mainly provided by nurses, who might liaise with social workers, to bolster patients' social support or to arrange additional community resources. Details of exercise training, however, were not outlined in the reviewed RCTs.

Table 5 also indicates that effective DMPs (n=10, 83.3%) were more likely than the ineffective DMPs (n=5, 55.6%) to optimize the medication regimen and to prescribe patients with flexible diuretic regime. Effective DMPs were more likely to rely on nurses or cardiologists for drug titration (n=6, 50%). Although three ineffective DMPs (33.3%) incorporated drug titration in the programmes, a telephone approach was used to conduct this intervention. The accuracy of the corresponding clinical assessment was, hence, in question.^21,41 As for directing attention to clinical deterioration, the effective DMPs (n=9, 75%) were almost twice more likely than the ineffective DMPs (n=4, 44.4%) to arrange medical consultations or offer prompt interventions to patients who were found to demonstrate early signs and symptoms of clinical deterioration.

Methods of delivering follow-up care

Table 5 indicates that the effective and ineffective DMPs adopted different methods to deliver follow-up care. Although 91.7% of the effective DMPs adopted multiple methods, including clinic/home visit, telephone follow-up, and/or telephone hotlines, only half of the ineffective DMPs delivered follow-up care with both in-person and telephone approaches. As for the use of each individual method, the effective DMPs were twice more likely than the ineffective ones to incorporate with home visits. Indeed, although the effective DMPs, which incorporated with home visits, recruited elderly heart failure patients in NYHA Class II and Class III,^19,26,35,38 the ineffective counterparts applied this kind of follow-up care to patients in higher NYHA classes (i.e. NYHA Classes III and IV).²⁴ Home visit appeared to be a more effective follow-up method for patients with moderate, but not severe heart failure. In contrast, McDonald et al.³¹ demonstrated the favourable effect of clinic visit in patients of NYHA Class IV. Although Ekman et al.²² reported contradictory findings, it should be noted that McDonald et al.³¹ provided patients with more intensive care such as medical review by cardiologist and infusion of parenteral diuretics in the clinic. Table 5 indicates that clinic visits were also used in patients with moderate heart failure among the reviewed DMPs, whether effective^25,27 or ineffective.^20,34 These effective programmes, however, would arrange home visit for the elderly heart failure patients who were unable to attend the clinic.

Study endpoints

In order to determine whether the effectiveness of the DMP for elderly heart failure patients would be affected by the time of its evaluation, the endpoints for evaluating effective and ineffective DMPs were compared. Among the 21 RCTs, the DMPs were evaluated during a period ranging from 90 days to 18 months. One study extended the longitudinal evaluation of the DMP to a median endpoint of 4.2 years.³⁸ By excluding this outlier, the mean time for programme evaluation was 9.6 months (SD=4.6 months). A greater number of effective programmes (75%, n=9), when compared with ineffective programmes (33.3%, n=3), evaluated the intervention effect at a time beyond this mean study endpoint. This result indicates that the DMP for elderly heart failure patients was more likely to take time to be effective. Indeed, the reviewed RCTs, which performed a longitudinal outcome evaluation, also indicated that DMPs were only effective to curtail mortality rates or combined event rates at 1 year, but not the 3-month endpoint.^36,38,40

Discussion

Heart failure is a disease associated with a very high hospital readmission and mortality rate. This review indicates that DMPs are effective in ameliorating these poor discharge outcomes in half of the RCTs published in the latest 10 years. The results of cost-effectiveness for several DMPs were encouraging, with cost savings from reduced hospital readmissions greatly outweighing expenditure incurred by the programmes. In order to facilitate the effective clinical application of this health-care intervention, the review deconstructed the DMPs, as reported in RCTs, into their constituent components for comparison. The results complement and extend previous work^12–15 and define more precisely the special features of the effective DMPs. This review indicates that the DMPs which entail the characteristics as recommended by the ESC guidelines are more effective in reducing hospital readmission and mortality rate among elderly people with heart failure.

Concerning the crucial characteristics of effective DMP, the results conform to the work of others,¹⁴ which suggested that a multi-disciplinary care team with an effective structure, and using a wider range of expertise, was more effective to address the complex transitional needs and enhances the discharge outcomes of elderly heart failure patients. The joint effort of cardiac nurse and cardiologist was critical to success of the programme.^43,44 Although the competency of cardiac nurse in responding to the health needs of heart failure patients, handling clinical deterioration, and monitoring treatment effectiveness have been reported,⁴⁵ this review highlighted the beneficial effect of extending the roles of cardiologists from giving professional guidance to delivering more direct care. Their active participation is important, as cardiologist was reported to be more likely to execute standard treatment protocols^46,47 and prescribe evidence-based drug regimes.⁴⁸ Involving a primary care physician in the care team is also beneficial, as they can supplement specialist care with more comprehensive treatment for the multiple comorbidities that present in elderly heart failure patients.^49,50

As for the programme content, the results reinforce previous work and indicate that intensive education, self-care supportive strategies, and optimization of drug therapy are the core interventions of DMPs for elderly heart failure patients.^13,14 This review also extends such findings and highlights several areas about the programme content that merit closer attention.

Firstly, the results of this study suggest that an in-hospital phase of care should be built into the DMP. The care should focus on optimizing the patients' clinical status and medication regimen and on developing a post-discharge plan to address the individualized risk factors of poor discharge outcomes. As high prevalence of clinical instability among elderly heart failure patients in the first few days and weeks after hospitalization has been documented,⁵¹ this phase of care is important to ensure that the patients would be stable and be better prepared for hospital discharge. Secondly, this review identifies exercise training and psychosocial care as essential elements of effective DMPs. There is strong evidence to suggest the beneficial effects of exercise training in improving exercise capacity, muscular function, neurohormonal balance, symptom control, psychological feeling, health perception, and quality of life of heart failure patients.^52–59 Psychosocial care, which includes ameliorating negative emotions and bolstering patient's social support and community resources, also plays an important role in buffering the heightened post-discharge stress of these patients. Under-utilization of these interventions among the reviewed DMPs, therefore, deserves prompt attention. Special effort should be directed at incorporating well-researched home-based exercise training such as walking and resistance training^60,61 and psychological interventions such as progressive muscle relaxation⁶² and guided imagery⁶³ in the DMP. Thirdly, this review indicates that among the various methods of optimizing the medication regimen for the elderly heart failure patients, drug titration by cardiac nurse with cardiologist back-up was the most promising to promote a better discharge outcome. Reviewing and optimizing the drug regimen before hospital discharge would be inadequate, as the unstable clinical condition of elderly heart failure patients in the initial period of hospital discharge⁵¹ implies their heightened needs for serial assessments and medication adjustment. Relying on cardiac nurse to advise primary care physician on drug titration was less effective, as a time gap might exist between the nurses conducting the patients' assessment and the primary care physicians revising the treatment in the scheduled clinic visit. Finally, arranging prompt medical referrals for any clinical deterioration is also a vital component of DMP for elderly heart failure patients, as prompt treatment can be administered to prevent the more reversible early heart failure decompensation from exacerbating.

Although previous work reported no substantial influence of the method of delivering follow-up care on the effectiveness of DMPs,^12,15 this review indicates that methods should be determined by the clinical condition of elderly heart failure patients. Clinic visit would be more appropriate for those who demonstrate risk factors of developing clinical instability, as it provides them with accesses to cardiologists and a channel for receiving advanced treatment without hospital admission. As for the comparatively more stable elderly heart failure patients, in whom effective self-care practice plays a more important role in controlling heart failure decompensation, this review suggests that home visit would be more beneficial. This can be explained by the fact that visiting the patients at home allows nurses to identify home environmental factors that hinder effective self-care and transfer the health information into the patients' real-life practices. The patients can also have increased readiness to learn in the less stressful home environment.^33,45,64 Clinic visit would also be an effective follow-up method for this group of patients, provided home visit is arranged whenever the patients have difficulty to access the clinics. As for the telephone approach, although this review does not suggest its beneficial effect when it is used as the sole method of delivering follow-up care, combining this method with an in-person method appears to be crucial to enhance the success of the DMPs. The cost-saving telephone approach is a necessary complement to the more expensive home visit/clinic visit in keeping track of early clinical deterioration and problems in self-care management.

DMP is an approach that has been widely adopted to manage various chronic diseases such as hypertension, diabetes mellitus, and asthma. Previous studies summarized their programme characteristics and effects.^17,65 The key difference between these DMPs and the reviewed DMPs for heart failure is that the latter programmes have seldom directed interventions to the providers as the DMPs for other chronic diseases did. The provider interventions focused on using education, feedback, and reminders to enhance the adherence of health-care providers to clinical guidelines or standard protocols. They are, in fact, of great value in enhancing the discharge outcomes of elderly heart failure patients, as 10–21% of heart failure decompensation was found to be related to the providers' factors such as inability to translate clinical evidence to practice and inadequacy of continuing education of health-care providers.⁶⁶ A lack of provider interventions among the reviewed DMPs for elderly heart failure patients might contribute to their lesser effectiveness in improving the disease control, when compared with the DMPs for the other chronic disease.^17,65

Another difference between the DMPs for other chronic diseases and those for heart failure lies in the context of application. Although the former DMPs were adopted to enhance the care in both hospital and community settings, the DMPs for heart failure patients, which were published in the last two decades, were mainly applied to patients at hospital or shortly after discharge.⁶⁷ Greater application of DMPs to this group of highly selected patients might be related to the intention of researchers to direct these programmes to those who would derive most benefit.⁶⁸ Nevertheless, as the disease trajectory of heart failure is characterized by a series of peaks and valleys, a continuum of care reflective of periods of acute exacerbation and stabilization has been described as the best to enhance the health outcomes.⁶⁹ The underlying goal of DMPs for heart failure patients should also be directed at shifting the care from hospital to primary care setting. DMP has been used successfully to maintain the phase of stabilization of patients with hypertension,⁷⁰ diabetes mellitus⁷¹ and hyperlipidemia.⁷² The use of DMP should be extended to maintain the health of a broader range of community-dwelling heart failure patients.

A heart failure DMP, which reflects the concept of ‘continuum of care’, has recently been conducted in United States.⁷³ The programme starts at a phase which focuses on stabilizing the clinical condition of heart failure patients after hospital discharge and then transit to a more longitudinal phase, which gives emphasis to maintaining the health of stabilized heart failure patients, and focusing on empowerment, optimization of care, continuous monitoring, and prompt resolution of clinical deterioration. In that longitudinal phase, heart failure nurse specialists and primary care physicians work closely together to deliver the care in a community-based heart failure clinic and by proactive telephone care. They provide intensive education, manage the comorbidities, and continuously monitor the clinical condition and therapeutic regimen for the patients. Family caregivers are involved, as they are regarded as the silent arm of the health-care system. The schedule of care is determined by the perceived risk of destabilization of the individual heart failure patients. Medical review by cardiologist is scheduled, but on a less frequent basis. The patients are discharged to the final ‘telemanagement’ phase when they demonstrate satisfactory functional status, competent self-care, and no evidence of congestion. This phase provides patients with access to a 24-h telephone call center, through which interventions including education and symptom surveillance are delivered. Whenever the clinical condition of the heart failure patients is found to be deteriorating, they would be moved back to the earlier phases of care, at which interventions such as referrals to cardiologists, diagnostic studies, and complex treatment such as parenteral infusion of diuretics would be arranged at an outpatient setting. The programme, hence, acts as a gatekeeper to prevent avoidable hospital admission. Although the effectiveness of this programme has not yet been published, a nurse-led DMP, which had similar design as its longitudinal phase of care, was found to reduce hospital readmission by 81% among a group of community-dwelling heart failure patients in United Kingdom.⁷⁴ In that nurse-led DMP, the specialist nurse also optimized the medication for the patients by getting reference to a protocol or seeking advice from the cardiologists. These two programmes provided good insights into the method of extending the application of DMP to a boarder spectrum of elderly heart failure patients.

Limitations

Several limitations of this study should be considered. First, this study identified the crucial characteristics of successful DMPs by comparing the content of the effective programmes with that of ineffective ones. There is a possibility that the programme effectiveness might be affected by variations in factors, other than the designs of DMPs, among the reviewed studies. Particular concern was given to the use of heart failure related medication such as ACE-I and beta-blocker. Although the prescriptions of such medications in the effective and ineffective DMPs were comparable with each other, considerable variations of this clinical factor among the 21 reviewed studies were noted. Their pharmacological effects on morbidity and mortality¹⁰ might interact with effects of the DMPs to influence the study endpoints. Indeed, the equivalence of the use of these medications between the study arms was also not examined in nine (42.9%) of the reviewed studies. The observed benefit of some effective DMP might be related to the higher use of ACE-Is in the experimental arm.¹⁹ Another factor that would influence the effectiveness of DMPs was the geographic variation among the studies. This review included studies from eight countries. Previous studies have reported the regional differences in patient demographics, heart failure aetiology and severity, specialist availability, length of hospital stay, physician's practice patterns such as ACE-I prescription and LVEF documentation, and access to health care, among these countries.^75–77 These geographic variations might imply that a DMP may be successful in one region but not in the other. The geographic factor, hence, competed with the characteristics of the DMPs to account for the changes in the study endpoints. Indeed, the effects of a DMP on the outcome variables would also be influenced by the quality of the control intervention. A lack of information describing the control interventions among majority of the reviewed studies does not permit meaningful comparisons to be made with regard to this aspect.

The other limitation is that meta-analysis was not undertaken to collapse the data of the reviewed studies and to identify the statistically justified relationship between process variables and outcomes of the DMPs. Instead, systematic comparison was used to identify the crucial characteristics of an effective DMP. Despite the intention of using this less stringent method was to allow the precise information about the characteristics of DMPs to be incorporated in the analysis and, hence, derived more qualitative findings to inform the real-life clinical practice, the bias arose from inadequate statistical control should be considered. Indeed, analysing the data without using meta-analytic technique did not allow the relative importance of the process variables of DMP for elderly heart failure patients to be identified. Future studies should conduct moderator analysis to address this limitation. The findings of this review, nevertheless, provided a platform for defining the process variables for undergoing the subsequent quantitative analysis. Finally, publication bias should also be considered, as studies demonstrating statistically significant benefits may be more likely to be published than those with non-significant or negative findings.⁶⁵

Conclusion

By systematically comparing the effective DMPs with ineffective ones, this review identified characteristics which are crucial to enhancing the discharge outcomes of elderly heart failure patients. The DMP should be multi-faceted and consists of an in-hospital phase of care, intensive patient education, exercise and psychosocial counselling, self-care supportive strategy, optimization of medical regimen, and ongoing surveillance and management of clinical deterioration. Cardiac nurse and cardiologist should be actively involved and a flexible approach should be adopted to deliver follow-up care. The beneficial effects of DMP also imply the need to extending its use in a boarder scope of community-dwelling heart failure patients. Health-care professionals are encouraged to apply these findings in their practice, with consideration given to the geographic factors and nature of their health-care system.

Conflict of interest: none declared.

References