Author + information
- John G.F. Cleland, MD⁎ ()
- ↵⁎Department of Cardiology, University of Hull, Castle Hill Hospital, Castle Road, Kingston-upon-Hull, HU16 5JQ, United Kingdom
We thank Dr. Chan and colleagues for their compliments about TEN-HMS (Trans-European Network–Home-Care Management System) and its primary end point (1). We agree that time-to-first-event analyses of composite outcomes are difficult to interpret and may be clinically misleading. Time-to-first event analysis is a powerful statistical tool, but it does appear that the “servant has become the master” in the design and analysis of many contemporary clinical trials. Patients with serious, symptomatic conditions have similar goals to most other people; they want to live a long and happy life. The emphasis on well-being and longevity will vary depending on the severity of symptoms and the expectations of the individual. Accordingly, clinical trials should try to capture this information, and composite outcomes are the preferred health measure because neither well-being nor longevity alone describes what most patients want.
Time-to-event analysis is clearly appropriate for mortality but not for most other outcomes in trials of heart failure; this is because it ignores all events subsequent to the first and ignores the potential of many patients to make a good recovery from what may be a minor incident. The reason for the continued popularity of this test is its track record, which has had the unfortunate side effect of suppressing the development of alternatives. There is no doubt that alternatives are required, but these new and more clinically sensible tools also need to be subjected to testing and critical analysis in terms of their internal logic, relevance, and statistical power.
It is clear that the measure of days-alive and out of hospital is not going to be a robust end point for most clinical trials because of the skewed distribution and large variation among individuals. Also, hospitalization is generally being used as a surrogate measure for well-being, and the validity of this assumption has not been adequately tested. Outcomes such as “Patient-Journey” or QALY (quality adjusted life-years) are easily measured and describe effects both on well-being and longevity (2). The distribution of derived values is more normally distributed than days-alive and out of hospital, and therefore statistical testing is more straightforward and powerful. Very simple quality-of-life or symptom-assessment tools should be used as patients may tire of filling in complex questionnaires repetitively. The method suggested by Dr. Chan and colleagues and variations on that theme used by others is certainly an attractive, alternative method for assessing the effects of therapy (3). However, we believe that this approach is better suited to medium-sized studies investigating the mechanism of drug effect. Such studies provide valuable information that will help decide whether it is worth proceeding to a definitive outcome study and whether its primary end point should be all-cause mortality, “Patient-Journey,” or QALY.
We also thank Drs. Koehler and Anker for their comments. We would have liked to take the credit for being the first study of telemedicine to show a reduction in mortality, but, in fact, we were only the second study to do so (4). Also, the DIAL (ranDomised trial of telephone Intervention in chronic heArt faiLure) study showed a reduction in hospitalizations for heart failure, although not mortality, in a very simple telemedicine project (5). The TEN-HMS trial shows one of the largest absolute differences in mortality for any intervention in heart failure. Accordingly, there is already more evidence of a mortality benefit from telemedicine in heart failure than for many other interventions. It is hoped that the next round of guidelines will recognize the importance of organization of care, which is greatly assisted by telemedicine, for patients with heart failure.
Hospitalization makes up just a small fraction of the patient experience of heart failure. Evolutions and revolutions in care are likely to increase expenditures on out-of-hospital care disproportionately. We believe that the importance of hospital care for patients with heart failure will diminish in the future owing to improvements in treatment, the delivery of care, and the development of hospices for the terminally ill.
Drs. Koehler and Anker rightly point out that telecare is still in its infancy and is only now approaching the threshold of mainstream medicine. Much more work is required on integrating health professionals into this system of care. We are confident that a telemedicine strategy can offer much greater benefit than we could demonstrate in TEN-HMS. We had no learning period to identify appropriate triggers and responses to the new information being acquired. The analyses they request are only now being done; the techniques required are complex to apply. We also agree that remote telemonitoring by people who have little knowledge of the patient and who are not in a position to offer practical help if needed is unlikely to be optimal. Telemedicine that integrates care at the local and regional level is most likely to meet with success.
- American College of Cardiology Foundation
- Cleland J.G.,
- Louis A.A.,
- Rigby A.S.,
- Janssens U.,
- Balk A.H.
- GESICA Investigators