Heart failure (HF) is associated with high mortality and morbidity, readmission rates and costs.1 Costs related to HF account for 1–2% of all healthcare expenditure, mainly the result of recurrent hospital admissions.2–4 Despite recent advances in medical and device therapy, patients with HF still suffer from repeated hospitalizations due to the combination of progression of the disease, poor adherence to diet and medical therapy, occurrence of comorbidities and limited support.5,6 Thus, this clinical entity remains a major medical and epidemiological problem which carries a heavy economic burden.6 New strategies to detect early decompensation and prevent HF-related hospitalizations and hence reduce health care costs are needed.

Multidisciplinary HF management programs and HF clinics, considered ‘usual care’ in several European countries, have been successful in reducing all-cause hospitalization rates.7,8 However, because of geographic barriers, socioeconomic constraints and other obstacles, only a relatively small proportion of HF patients have access to such programs. Interventions have therefore evolved to better monitor HF patients at home.9

Telemonitoring is a novel tool to improve patient care and adherence which encompasses the use of recent communication technologies to monitor simple clinical variables which are transmitted to the health care provider. Its goal is to detect early signs of heart failure decompensation, providing an opportunity for intervention before the patient requires hospitalization (Figure 1).10–12

Figure 1.

Clinical cascade of heart failure decompensation. The main purpose of telemonitoring is to interrupt this cascade, preventing hospitalization.

(0,11MB).

Several non-invasive telemonitoring strategies have been proposed, using regularly scheduled structured telephone interviews or more sophisticated systems, such as electronic transfer of physiological data with remote access control via external, wearable or implantable devices. They have been assessed in retrospective and prospective clinical studies, with conflicting results. This paper sets out to provide a critical review of the current evidence on telemonitoring in HF.

Various observational studies on HF telemonitoring have called attention to its potential benefit. Two major meta-analyses aimed to assess the overall effect of HF telemonitoring on prognosis.

In 2009, Klersy et al.13 reviewed 96 articles, comparing multidisciplinary HF approaches by either usual care or remote patient monitoring. The cumulative incidence of events in the usual care approach (in-person visit) and in remote monitoring strategies (telephone or technology-assisted monitoring approaches) was compared. A total of 6258 patients were included in randomized controlled trials (RCTs) and 2354 patients in cohort studies, with a median follow-up of 6 and 12 months, respectively. In RCTs telemonitoring was associated with a significant reduction in mortality compared to usual care (relative risk [RR]: 0.83, p=0.006), total hospitalizations (RR: 0.93, p=0.030) and hospitalizations for HF (RR: 0.71, p=0.001). The combined endpoint of death or first hospitalization showed similar results (RR: 0.86, p=0.001). In cohort studies, telemonitoring was also associated with a significantly lower number of deaths (random-effects RR: 0.53, p=0.001) and hospitalizations (random-effects RR: 0.52, p=0.001). Hence, according to this meta-analysis, remote monitoring significantly reduced the risk of death and hospitalization for any cause in both RCTs and even more markedly in cohort studies.

In 2011, Inglis et al.,14 updating a study by Clark in 2007,15 published an extensive meta-analysis of RCTs on structured telephone support or telemonitoring compared to standard practice for patients with HF. They included five abstracts and 25 studies, of which 16 evaluated structured telephone support (5613 participants), 11 assessed telemonitoring (2710 participants) and two tested both interventions. Telemonitoring reduced all-cause mortality (RR: 0.66, p<0.0001) with structured telephone support demonstrating a non-significant positive effect (RR: 0.88, p=0.08). Both structured telephone support (RR: 0.77, p<0.0001) and telemonitoring (RR: 0.79, p=0.008) reduced HF-related hospitalizations. In several studies both interventions improved quality of life and reduced healthcare costs. Improvements in therapeutic adherence, patient education and self-care, and NYHA functional class were also observed. This meta-analysis thus showed that both structured telephone support and telemonitoring are effective in reducing all-cause mortality and HF-related hospitalizations in patients with HF. Additionally, they improve quality of life, reduce costs, and enhance evidence-based prescription.

Although supporting the use of telemonitoring, the above meta-analyses present several intrinsic methodological limitations.

The first is inherent to most meta-analyses and is due to publication bias: only large studies or small studies with positive results, as opposed to those with negative results, tend to be published and therefore included in the analysis.

An additional important limitation refers to the potential lack of quality of the studies included, some of which are small and single-center.

Moreover, there is also heterogeneity in the studies included regarding selection criteria, design, methodology, telemonitoring technology and follow-up periods. For example, with regard to the structural telephone support group, the type of professional providing the support and the number of telephone contacts varied. The control groups also varied widely and in some studies the description of this group was too brief or vague to allow replication or comparison with other trials. Additionally, the inclusion and exclusion criteria varied across trials, for example regarding NYHA class or clinical setting (post-hospital discharge or community settings). This variability leads to a heterogeneous overall population, presenting important limitations to the overall analysis.

Finally, the majority of the included studies were performed in a previous era of HF treatment: recent advances in pharmacological and device therapy, which would now be offered in the standard of care arm, might have changed the overall results; furthermore, the current standard of care of heart failure treatment includes disease management programs and multidisciplinary care, which in previous studies were implemented in the remote monitoring arm only.

Given the above, can it be assumed that the studies included in these meta-analyses are sufficiently homogeneous to give credibility to the results? And if so, to which patients are the findings applicable? And if telemonitoring and structured telephone support are beneficial, what level of technology is appropriate?

Large RCTs have tried to respond to these questions and their main results will be described in the next section.

Among trials conducted in a more contemporary setting, two will be reviewed in view of their size and importance: Tele-HF and TIM-HF.

The Telemonitoring to Improve Heart Failure Outcomes (Tele-HF) trial16 was undertaken to determine the prognostic effect of telephone-based automated symptom and self-reported weight monitoring compared with usual care in patients recently hospitalized for heart failure. A total of 1653 patients were enrolled from 33 cardiology centers; 826 were randomly assigned to undergo telemonitoring and 827 to receive usual care. Clinicians were instructed to treat their patients according to guidelines and all patients received educational materials, even in the usual care arm.

The results regarding the primary endpoint of all-cause readmission or death within 6 months were similar for the telemonitoring and standard-of-care groups (52.3% vs. 51.5%, respectively, p=0.75). Secondary endpoint results were also similar for the telemonitoring and standard-of-care groups (death: 11.1% vs. 11.4%, respectively, p=0.88; readmission rate: 49.3% vs. 47.4%, p=0.45; readmission for heart failure: 27.5% vs. 27.0%, p=0.81). The time to event for the composite endpoint of readmission or death from any cause was not significantly different between the two groups. Also, there were no significant differences in subgroup analyses, suggesting that no demographic characteristic was likely to predict benefit.

The Telemedical Interventional Monitoring in Heart Failure (TIM-HF) trial17 was designed to determine whether physician-led remote telemedical management compared with usual care would result in reduced mortality in HF. A total of 710 optimally treated, stable, ambulatory patients in NYHA class II or III, left ventricular ejection fraction (LVEF) of ≤35% and a history of HF decompensation within the previous two years or with LVEF ≤25% were randomized to usual care (356) or daily telemonitoring (354 patients whose daily electrocardiogram, blood pressure and body weight measurement were sent to telemedical centers with 24-hour physician availability), with a follow-up of around two years. There was no difference in the primary endpoint of total mortality (hazard ratio [HR]: 0.97; p=0.87) or in the secondary endpoint composite of cardiovascular mortality or hospitalization due to HF (HR: 0.89; p=0.44). Potential benefit was suggested in the subgroup analysis for those with a prior heart failure hospitalization and LVEF of 25% or higher. Other secondary endpoints included cardiovascular mortality, all-cause and cause-specific hospitalizations (all time to first event), days lost due to heart failure hospitalization or cardiovascular death (in % of follow-up time), and changes in quality of life and NYHA class, with similar results between the arms.

Thus, despite the previous claims of success, in large RCTs a telemonitoring strategy failed to provide benefit over usual care. Although the reported results should prompt a critical reappraisal of telemonitoring, two main questions need to be addressed. Should these findings be considered as definitive evidence against telemanagement for heart failure? Are the assessed features adequate warnings of decompensation?

First, regarding Tele-HF, it is important to note that adherence was a significant issue, with 14% of those randomized to telemonitoring never using the system and only 55% using it at least three times per week by the end of the study. Furthermore, given the large amount of generated data, physicians’ adherence to the system might also be a concern. Data were transmitted directly to the attending physicians, who were required to document the existence of variances from the basal status and report their responses to these variances. As the authors state, given the lack of systematic recording of these data, it is difficult to assess the extent to which the variances were promptly reviewed and purposeful decisions were made. It can thus be assumed that if monitoring of patients were coupled with more systematic reporting and detailed and prompt interventions, it still might show some benefit. Also, the intervention in this study consisted only of data collection. Newer technologies capable of supporting patient education and self-care by using daily, real-time monitoring of physiological data, direct patient feedback and coaching, and a high level of patient–clinician interaction, might achieve positive results.

To summarize, the study's negative findings may be due to low patient and physician adherence to the intervention and/or to inadequate intervention.

In the case of TIM-HF, the results may have been less influenced by patient compliance (81% of the patients were compliant to the daily transfer of data to telemedical centers) or physician compliance (two telemedical centers provided physician-led telemedical support 24 hours a day, 7 days a week). However, TIM-HF presented lack of power to detect clinically relevant differences between the groups, as evidenced by the wide 95% confidence intervals. Given this lack of power, the authors state that the results found do not rule out the potential role of telemonitoring as an addition to the management of HF, but emphasize the need to identify the HF population that could benefit from using this intervention.

Despite notable advances in the understanding of the pathophysiology and treatment of heart failure, it still carries an enormous clinical and economical burden. This is due in great part to hospitalizations.

Detecting decompensation before it leads to hospitalization appears a promising strategy, and so, conceptually, telemonitoring should be a valuable tool to improve outcomes in this population.

In fact, several observational studies and small RCTs have supported this hypothesis. However, the results from major RCTs did not support the benefit of telemonitoring over usual care in HF patients.

Ambulatory hemodynamic monitoring via implanted devices would theoretically offer more accurate and robust data to identify patients at greater risk of decompensation. However, the few published studies still lack consistency in demonstrating their benefit, which must also be weighed against safety concerns due to the invasive nature of the procedure.

Thus, at this point, it is difficult to draw firm conclusions regarding the clinical efficacy of telemanagement (Table 1). Despite the growing interest in telemonitoring in the cardiology community, many questions remain unanswered. Which patients benefit most, and how often? Which parameters should be monitored? How could these parameters be monitored more efficiently? How should responses of health care professionals to the data obtained from monitoring be managed?

These are the million-dollar questions. Future research should focus on these aspects in order to find the most ‘active ingredients’ which can make telemonitoring work.

Perhaps the biggest revolution in this area is the development of tools that analyze the data automatically and provide advice to both patients and health professionals in making care decisions. These can give patients more control in managing their problems and much more personalized health-care. Mobile phone-based remote monitoring systems are relatively inexpensive and convenient tools to improve HF home management. Mobile phones are now widely available, inexpensive and portable, enabling patients to be monitored anywhere. Initial studies have shown the potential of this approach in HF home management,28 but further studies are needed.

This review underscores the need for careful assessment of telemonitoring as a disease management system before its widespread adoption. Until the ideal tools are found, dedicated monitoring for HF may be a practical adjunct in selected centers and patients, additional to usual care, but should not replace it as a standard of care so long as the evidence remains conflicting, insufficient and heterogeneous.

The authors have no conflicts of interest to declare.