Our study suggests that nebulized HFNC is unequivocally cost-effective and dominant over the control strategy, achieving better outcomes at a lower cost in hospitalized patients with bronchiolitis. The potential magnitude of cost savings for the health system (US$ 72 per patient), especially in a middle-income country, is not negligible if we consider that in our country we have around 95,000 annual cases of children under 2 years old being hospitalized with acute respiratory infections (21). Our findings provide inputs for changes in clinical practice guidelines in similar settings, where efficiency in the allocation of health resources can be maximized.
Common concerns regarding the adoption of technologies relate to safety and effectiveness. Current evidence is conclusive regarding the safety of HFNC. In a recent systematic review of seven RTCs, no severe adverse events and no increased risk of air leak syndromes were reported (9). Only mild skin lesions were reported in under 3% of patients. In addition, parents and carers favored HFNC because of the ability to feed and overall comfort (9, 37). This aspect differs from CPAP, which is usually not well tolerated and interferes with patients’ normal care, and is associated with a higher risk of air leak syndromes (38). In terms of HFNC’s effectiveness, the evidence is still debatable, with studies for and against its use concerning CN or CPAP in children with moderate or severe bronchiolitis both existing. The HFNC has been demonstrated to have had a beneficial effect on treatment failure, but many authors criticize this effect for the absence of effects on other related outcomes. For example, only one clinical trial included in the previously mentioned systematic review (39) found an effect on LOS and the duration of oxygen therapy,
and in none of the trials included was there an effect on the probability of PICU transfer. It is clear that there was no impact in this trial on the risk of admission to PICU because all patients with CN failure went onto HFNC, while those with HFNC failure went directly to PICU in these studies. This therefore underestimated the impact on this outcome and affected the difference in hospital stay rates and the duration of oxygen therapy. Indeed, over 60% of patients with a failure of standard oxygen therapy that switched to HFNC successfully recuperated (23, 37).
There is clearly a need for more clinical trials with standardized oxygen weaning protocols, but this therapy currently constitutes an optimal resource prior to mechanical ventilation. While our study showed only a slight effect on utilities between the two strategies evaluated, due to the fact that HFNC does not directly impact the chance of complications, HFNC has a significant positive impact on the cost because it reduces the use of secondary resources, especially those arising from PICU admission. These findings arose from a previous economic study carried out in parallel during a clinical trial, demonstrating that the HFNC arm required fewer resources than the CN arm (37). This is particularly important in scenarios with more limited economic resources. In at least four observational studies carried out in developing countries, a significantly lower proportion of children were transferred to PICU before the initial HFNC in emergency settings (5, 7, 40, 41). Indeed, clinical trials published until now with larger sample sizes show that the magnitude of the effect of HFNC was higher in hospitals without on-site PICUs than in hospitals with on-site PICUs. This escalation of care occurred at a rate of 7% in the HFNC group compared to 28% in the standard therapy group, while in hospitals with on-site PICUs this endpoint occurred at 14% in the HFNC group and 20% in the standard therapy group (23). Without a doubt, and reinforced by our results, the use of HFNC in emergency settings is an efficient alternative for infants with moderate-severe bronchiolitis to optimize the cost in scenarios without on-site PICUs or with limited economical resources.
A very important aspect of our model is that it was robust to changing the model's utility and cost values. HFNC was always the most cost-effective strategy in all ranges of thresholds evaluated with a low population EVPI. This was consistent with the finding that, although our utilities were collected from other populations, our results did not change when exploring the change in the ICER in the range of values of each utility explored. The same happens with costs. Although the resources, frequencies of use, and costs were collected from tertiary centers in Rionegro and not from a national study of all hospitals in Colombia, modifications to their values in the sensitivity analysis also did not significantly change the ICER. These aspects give us confidence in relation to the ability to make decisions based on our results. As is always necessary in science, more studies are needed to replicate our results (42). Our study has some limitations. The cost data were collected retrospectively. Bronchiolitis treatment and the costs in question, including hospital prices, did not markedly change. Furthermore, our country has been characterized by having very low price variation in the last 10 years, especially in terms of health services (24). Additionally, we use utilities extracted from the literature and not estimated directly from our population. As was mentioned previously, the reliability and robustness of the results were evaluated using sensitivity analyses. However, an additional strength is the perspective of the society on which the economic analysis was focused, which allows a faster transfer of results to health policies.
In conclusion, HFNC in emergency settings was cost-effective for the hospital treatment of infants with moderate or severe bronchiolitis. Our study provides evidence that should be used by decision-makers to improve clinical practice guidelines and should be replicated to validate their results in other countries.