In reply:
We thank Dr Bach and colleagues for all of their comments, and we want to take this opportunity to reemphasize a few points of our publication. Evidence-based medicine is often contrasted to the real world of clinical practice. However, in our opinion, they are not contradictory but complementary.
From a methodological point of view, randomized controlled trials are the most rigorous way of determining whether a cause-effect relation exists between a treatment (drugs or medical devices) and a clinical outcome.1 It is the best way to avoid the risk of bias. The use of placebo groups is still an unresolved debate.2 Even if proven-therapy trials can be thought ethically preferable to placebo-controlled trials, the reality is more complex.
Robust published studies on treatments used in current practice regularly reflect the lack of demonstration of their effectiveness. The use of adaptive servo ventilation based on limited data is a good example.3–5 It was only after the SERVE-HF study was conducted that the cardiovascular risk in treated patients was discovered.6 The mortality rate in the intervention group was statistically higher than in the control group (10%/y vs 7.5%/y). These results have surprised the medical community and led to a safety alert. Although the SERVE-HF study was also criticized, it had the merit of formally highlighting the lack of evidence-based knowledge in this field.
Another example is chest physiotherapy, which was commonly used for bronchiolitis in infants. The available evidence now demonstrate that it is not effective. Its risk/benefit seems even unfavorable. A Cochrane review of 9 studies, carried out on 891 infants hospitalized for bronchiolitis, demonstrated no difference between children treated with or without physiotherapy,7 neither in terms of clinical outcome nor considering blood oxygenation, breathing frequency, duration of illness, and duration of hospitalization. Many adverse effects of physiotherapy, such as vomiting, pain, and even rib fractures (one fracture per 1,000 treated infants) are highlighted in these 9 studies. The 2016 update confirms the need for clinical evidence to adapt clinical practice.8
We pointed out in our discussion that designing a randomized study with a treated group versus an untreated group is sometimes difficult to consider due to ethical considerations.9 However, different experimental designs are available when conventional trials cannot be applied to the clinical development of medical devices.10 Aware of the difficulties of conducting randomized controlled studies, we decided to also consider observational studies to collect information on safety data in our systematic review.
The work of Bach et al11 in 2011, dealing with continuous noninvasive ventilatory support in subjects with Duchenne muscular dystrophy, is clearly interesting in that it provides descriptive information on the impact of noninvasive ventilation associated with mechanical in-exsufflation on patient survival. However, we wish to emphasize that the aim of our work was to evaluate the current level of evidence available for the use of mechanical in-exsufflation devices for airway clearance in patients with neuromuscular diseases.9 The purpose of this evidence-based evaluation is to provide the basis for health-care professionals' working group discussions to make public health recommendations.
To conclude, we believe that patients, caregivers, and health-care professionals need evidence-based data to be able to benefit from factual information so that they can take or deliver treatments in an informed manner.
Footnotes
The authors have disclosed no conflicts of interest.
- Copyright © 2017 by Daedalus Enterprises