Viral bronchiolitis is one of the most common reasons for infant hospitalization in the United States. Despite the morbidity attributable to the disease, there are no generally recommended therapeutic strategies for bronchiolitis other than supportive care.1 The pull of the therapeutic vacuum surrounding this disease has been noted in the pages of this journal for at least 50 years, with Wright and Beem2 writing in 1965 that “energies should not be frittered away by the annoyance of unnecessary or futile medications and procedures” for the child with bronchiolitis.
Pairwise meta-analyses, the pinnacle of the traditional evidence pyramid, have been controversial in settling the question of which, if any, therapies are effective for bronchiolitis. For example, multiple meta-analyses examining the benefit of nebulized hypertonic saline (HS) on hospital length of stay (LOS) have revealed disparate conclusions.3 Brooks et al4 suggested that the beneficial effect of HS on LOS noted in previous meta-analyses was likely due to heterogeneity and imbalance of baseline patient characteristics.5 Using newer statistical techniques, such as trial sequential analysis, researchers have attempted to resolve these controversies.3
In this issue of Pediatrics, Elliott et al6 present the results of a rigorous network meta-analysis (NMA) of therapeutics for bronchiolitis. Unlike pairwise meta-analyses, which are used to compare one intervention with another, NMAs can be used to compare multiple therapeutics against each other by using a network of direct and indirect comparisons. Unique to NMAs, an indirect comparison is a comparison between the effects of therapeutics that have not been studied against each other directly. For example, if treatment B has a theoretical effect size of 4 compared with treatment A in one trial, and a separate trial reveals an effect size of 2 of C compared with A, the effect size of treatment B over C would be 2 in an indirect comparison. This comparison is contingent on the assumption of transitivity, that every condition is similar between the trials save for the intervention being compared itself.7 NMAs have the advantages of combining all available evidence into a single effect estimate and the ability to rank therapeutic options by effectiveness, even if they have not been compared head-to-head. However, they involve considerable statistical complexity and are limited by the underlying assumption of transitivity.
Elliott et al6 performed 3 separate NMAs analyzing the effect of all potential published bronchiolitis therapeutics on 3 outcomes: hospital admission rate on day 1, hospital admission rate within 7 days, and total hospital LOS. They found, with low confidence, that nebulized epinephrine and nebulized HS plus salbutamol reduced admission on day 1 compared with a placebo, no treatment decreased admission by day 7, and nebulized HS and HS plus epinephrine reduced LOS in hospitalized children. However, when only trials with low risk of bias were evaluated, many of these effects disappeared, with no treatment superior to placebo for the admission outcomes, whereas HS and HS plus epinephrine continued to have efficacy in reducing LOS.
What, then, are the take-aways? First, that even the most robust and statistically complex meta-analyses are sensitive to the quality of the underlying studies. Second, that even these complex methods are subject to the same challenges with heterogeneity as noted by Brooks et al,4 namely the inclusion of outlier studies with vastly heterogenous discharge criteria (eg, 12 hours without any respiratory symptoms in Luo et al8 ), for the LOS outcome.
We are at a precipice: do we jump farther down the cliff of continued research into these treatments, or is it time to focus on strategies to de-implement their widespread use? The authors conclude that further studies are needed; however, at this stage in the bronchiolitis journey, we must be vigilant against more of the same. If further studies are to be performed, several criteria must be met: (1) only therapeutics with any potential of efficacy should be evaluated (this NMA suggests that epinephrine, HS, or HS plus salbutamol may be effective; the time for evaluation of salbutamol or corticosteroids alone has passed); (2) studies must be sufficiently powered to evaluate clinically appropriate outcomes; (3) studies must be rigorously designed to have low risk of bias; (4) disease and outcome definitions must be valid, reproducible, and applicable to most children; and (5) results must be generalizable to most infants with bronchiolitis. In our opinion, further time and resources should not be used for trials that do not meet these criteria.
In the meantime, we must curb the rampant use of therapies repeatedly revealed to be ineffective. Team engagement, clear clinical practice guidelines, and securing institutional resources, namely information technology, have been cited as key factors in successful deimplementation.9,10 Multifaceted deimplementation packages, including clinician and family education, audit and feedback, and clinical decision support have had success.11,12 Although frustrating for clinicians and parents alike, the mantra for most children with bronchiolitis should continue to be “don’t just do something, stand there.”
Opinions expressed in these commentaries are those of the authors and not necessarily those of the American Academy of Pediatrics or its Committees.
FUNDING: No external funding.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2020-040816.
- HS
hypertonic saline
- LOS
length of stay
- NMA
network meta-analysis
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments
RE: We need robust evidence to support deimplementation in bronchiolitis
Over the last 60 years clinician's management of bronchiolitis has been littered with rapidly escalating use of interventions and therapeutics with poor evidence base. Recently the exponential up-take of high-flow nasal (HFN) therapy ahead of the evidence wave provides caution against widespread use of interventions without high-quality evidence. Early observational evidence (mainly using historical controls) showed an association between HFN therapy and reduced invasive ventilation.(2) Yet subsequently a RCT in intensive care favoured nasal continuous positive airway pressure vs. HFN therapy, and RCTs in paediatric wards vs. standard oxygen therapy failed to show any benefit in terms of ICU admission.(3) These RCTs resign HFN therapy to rescue therapy for hypoxemia failing to respond to standard oxygen therapy, yet all too often pediatric wards are full of infants with bronchiolitis being managed with HFN therapy, even when hypoxemia is absent.
Just as we need robust RCTs to evaluate therapeutics, we need well designed theory-informed intervention packages and robust RCTs to evaluate deimplementation strategies. Recently we developed such an intervention package following a qualitative study to understand barriers and enablers to evidence-based bronchiolitis management.(4) The resulting intervention package addressed barriers and enablers identified using behaviour change techniques most likely to alter practice. The deimplementation package included: 1)Clinical leads in both emergency departments and inpatient paediatric units, 2)Stakeholder meetings to get hospital buy-in, 3)Train-the-trainer workshop for clinical leads, 4)Educational material for clinical leads to deliver to clinicians in their workplace, 5)Additional educational materials, and 6)Monthly audit and feedback cycles. In our multi-center cluster RCT (26 hospitals; n=3,727) we successfully deimplemented five recommendations from the Australasian Bronchiolitis Guideline which had high-quality evidence of no benefit: chest x-ray, albuterol, glucocorticoids, antibiotics and epinephrine (adjusted risk difference, 14.1%; 95%CI, 6.5%-21.7%; P<.001).(5) There was strong evidence for improvement in intervention infants in decreasing use of albuterol and chest x-ray. All intervention activity occurred during clinical leads' nonclinical time and existing local educational programs, thus our pragmatic design is likely to be generalizable to other settings.
Improving the treatment of infants with bronchiolitis is important as practice variation continues to be an international problem. We believe we are now at the precipice to jump further into upscaling deimplementation of non-evidence-based care of infants with bronchiolitis, by following the mantra "don't just do something, stand there." Let's follow the mantra with high-quality evidence supporting these deimplementation strategies.
1. Elliott SA, Gaudet LA, Fernandes RM, Vandermeer B, Freedman SB, Johnson DW, et al. Comparative Efficacy of Bronchiolitis Interventions in Acute Care: A Network Meta-analysis. Pediatrics. 2021;147(5).
2. Schlapbach LJ, Straney L, Gelbart B, Alexander J, Franklin D, Beca J, et al. Burden of disease and change in practice in critically ill infants with bronchiolitis. Eur Respir J. 2017;49(6).
3. O'Brien S, Craig S, Babl FE, Borland ML, Oakley E, Dalziel SR, et al. 'Rational use of high-flow therapy in infants with bronchiolitis. What do the latest trials tell us?' A Paediatric Research in Emergency Departments International Collaborative perspective. J Paediatr Child Health. 2019;55(7):746-52.
4. Haskell L, Tavender EJ, Wilson C, Babl F, Sheridan N, Oakley E, et al. Understanding factors that contribute to variations in bronchiolitis management in acute care settings: a qualitative study in Australia and New Zealand using the Theoretical Domains Framework. BMC Pediatr. 2020;20(1):189.
5. Haskell L, Tavender EJ, Wilson CL, O'Brien S, Babl FE, Borland ML, et al. Effectiveness of Targeted Interventions on Treatment of Infants With Bronchiolitis: A Randomized Clinical Trial. JAMA Pediatr. 2021.