Telemonitoring for Heart Failure: The Only Feasible Option for Good Universal Care?

This editorial refers to ‘A randomized trial of home telemonitoring in a typical elderly heart failure population in North West London: results of the Home-HF study’† by Owais Dar et al., on page 319 and ‘Home telemonitoring in heart failure patients: the HHH study (Home or Hospital in Heart failure)’‡ by Andrea Mortara et al., on page 312 Telecare will become the preferred method for managing most long-term medical conditions and this revolution may well be as seismic for health care as the industrial revolution was to Western economies 200 years ago. When and how this revolution will occur is not yet clear. The change will be driven by the increasing numbers of patients with long-term conditions because of the triad of: earlier diagnosis, longer survival after diagnosis and population ageing. This increasing prevalence of chronic illness will occur in the context of fewer people of working age who may be reluctant to pay the higher taxes required in order to fulfil all of society's needs. Increasing expectations from care, increasingly complex but effective treatment, improved sensors that detect more problems and ubiquitous, secure, familiar means of telecommunications, will create new incentives and pressures to deliver more expert care. Reaping the benefits of the telecare revolution is not going to be easy. Telecare is not only an immature technology, but a whole new set of medical disciplines. No one who has conducted a trial so far has ever received more than theoretical training and so telecare information is being used in the same conventional way that it would be in a hospital clinic. This is not good use of potentially smart technology. Training programmes for health professionals are only now being set-up. Monitoring alone probably achieves little. Monitoring should be integrated with expert decision-support systems and diagnostic alerts that identify both the nature and optimal response to a problem rather than just its mere existence.1 The sheer density of information provided by telecare changes its nature and ensures that health professionals unsupported by computer technology are ill-equipped to make best use of it. Many sensors are not really appropriate for telecare and merely mimic the bad habits of routine clinical practice. Further work is required to establish the optimal sensors for inclusion in telecare systems. Fragmented health-care organizations that are not interested in unified services are a further impediment to innovation. A recent systematic review identified seven randomized trials of telemonitoring that included 807 patients, and suggested that telemonitoring reduced mortality substantially (hazard ratio 0.620; 95% confidence intervals 0.45–0.85; P = 0.003).2 However, hospitalization rather than mortality was the primary target of these studies, but no reduction in hospitalization was observed. The lack of effect on hospitalization could simply reflect a lack of benefit. However, timely hospitalization may have been the reason for the reduction in mortality and keeping patients alive means they have a chance of being hospitalized again. In this sense, hospitalization is a good outcome rather than evidence of disease progression and casts doubt on the wisdom of hospitalization as an endpoint in clinical trials. Telemonitoring may also generate false alerts leading to inappropriate hospitalization.1 This might be reduced by training, the development of robust clinical pathways integrated with existing services and the development of smart decision-support software. However, sceptics may consider that the reduction in mortality merely reflects the play of chance or bias in what are inevitably unblinded studies. Clearly, more data from randomized controlled trials are required, but are we really ready and able to conduct them? This issue of the European Journal of Heart Failure reports two well-conducted randomized controlled trials of telemonitoring for heart failure.3,4 Both were neutral and sceptics might even suggest that they showed some evidence of harm from telemonitoring. In HOME-HF,3 182 patients with a recent hospitalization for heart failure were randomly assigned to daily telemonitoring of symptoms, weight, blood pressure, heart rate and blood oxygen saturation or to a control group that received a package of intensive, conventional, expert care. Patient characteristics differed from previous studies. A large proportion had new-onset heart failure (44%) and a normal left ventricular ejection fraction (40%). The 91 patients in the control group were seen, on average, each month by a specialist and once every two weeks by some health-care professional (1136 visits) during the study's six-month follow-up. In addition, patients in the control group had access to advice by telephone during usual working hours. Differences in treatment uptake were not reported but seem unlikely given the intensity of follow-up. Compliance with telemonitoring was high but generated an alert only about once every 3 months. Only four patients (4.4%) died in the control group compared with 14 patients (15.4%) in the telemonitoring group. There was no difference in hospitalization rate. The outcome in the telemonitoring group is entirely consistent with that in TEN-HMS,5 where both intensive nurse support and telemonitoring were associated with a 6-month mortality of 18%, compared with a mortality in patients assigned to usual care (i.e. discharged with advice to the primary care physician and a 3-month review by a specialist) of 26%. The mortality in the control group of HOME-HF was not significantly lower than for telemonitoring and is probably a chance result. The HHH trial4 randomly assigned patients with a hospitalization for heart failure in the previous year and a left ventricular ejection fraction ≤40% already receiving optimal care (up to 90% on ACE-inhibitors and up to 87% on beta-blockers) either to conventional care (n = 160) or to telemonitoring (n = 301). Patients assigned to telemonitoring reported data on symptoms, weight, heart rate, respiratory rate, systolic blood pressure, and unspecified blood tests, but only on a weekly basis. Mortality was again extraordinarily low in both groups, attesting to the power of intensive high-quality care. The HHH study also showed, as in other trials, that well-managed patients with heart failure do not spend much time in hospital (approximately 1% of follow-up). Patients assigned to telemonitoring did slightly worse, which might have been because of a higher proportion of sicker patients being assigned to telecare in Poland. The study also suggested that sleep apnoea might be a useful marker of mortality risk. Data from TEN-HMS suggests that both strategies used in each trial should reduce mortality by a similar amount.5 If both strategies are equal then efficiency and cost become important. Paying for staff rather than hospital buildings is by far the largest component of health-care costs in Europe. If telemonitoring services can provide similar benefits with fewer staff, then it will become the preferred management option. The health economic analysis of HOME-HF suggests that this has not yet been achieved, but as telemonitoring services mature and become smarter this balance will change. So what are the take-home messages? The value of telemonitoring is not yet conclusively proved. Larger trials with better technologies, which address patient populations with unmet needs, are required. DOT-HF was one such trial using an implanted telemonitoring technology, but has been stopped for administrative reasons.6 A major problem with the current trials is that the needs of the patients in the control group may have already been met by expensive interventions. Most patients with heart failure receive much poorer care than the patients included in any control group of any telemonitoring study conducted so far. In the EuroHeart Failure Survey, on average, each hospital identified 16 patients with heart failure per week and of those who survived to discharge, mortality was 7% at 12 weeks.7,8 Annual mortality for contemporary patients discharged from hospital after worsening heart failure in the real world is between 25% and 40%.9–12 Many patients do not get the basic investigations and treatment that they should. The point of telehealth is that it will enable many more patients to enjoy the standard of care offered to only a selected few at the moment. The secret weapon of telemonitoring, as yet not deployed, is that it can actively engage the largest health-care workforce in the world; that is the patients and their carers. However, there are many issues to address before a definitive outcome trial should be attempted. European Union-funded initiatives such as MyHeart (http://www.hitech-projects.com/euprojects/myheart/home.html) and HeartCycle (http://heartcycle.med.auth.gr/press.php) will do much needed ground-work. Hopefully, within the next decade, telemonitoring will become a part of routine management for most patients with serious cardiovascular disease. Conflict of interest: J.C. was Principal Investigator in the TEN-HMS study and is a consultant for MyHeart and a partner in HeartCycle which are jointly funded by EU grants and Philips. K.G. is partly funded through HeartCycle and has been involved in analysis of the TEN-HMS study.