Acute bronchiolitis (AB) is an important disease in paediatrics on account on its impact on the health of patients,1 the sharp increase it generates in the demand for services on account of its seasonal patterns at every level of care (primary, emergency and inpatient care)2 and the economic, social and family impact it has past the acute episode.3 Acute bronchiolitis provides a good example of the variability of clinical practice in paediatrics, with numerous studies demonstrating it at the primary care level,4 in emergency departments4,5 and in the management of patients staying in hospital wards4,6 or admitted to paediatric intensive care units.7,8 This variability sometimes entails the overuse of diagnostic and therapeutic resources.9 Clinical practice guidelines (CPGs) are developed with the aim of guiding physicians in making decisions about the management of patients in specific clinical situations.10 Adhering to their recommendations in clinical practice can help reduce the use of interventions that are not backed by high-quality evidence,10 reduce the variability of clinical practice and standardise the care delivered by clinicians, improving patient safety and quality of care.11 It can also result in lower costs for health care systems by increasing their efficiency.3 Studies in the literature show that variability in clinical practice and the excessive use of resources in the management of AB persist despite the publication of CPGs, which highlights a lack of adherence to their recommendations on the part of health professionals.12–15 The aim of our study was to assess whether the implementation of a bronchiolitis CPG using a structured approach to the dissemination of information succeeded in changing the practices of health professionals to improve adherence to the recommendations of the CPG, with the corresponding decrease in the use of diagnostic and therapeutic resources.

During the entire period under study, from October 2006 through May 2017, we initially included a total of 277 patients admitted during the 11 epidemic seasons. Of this total, we excluded 18 because they did not meet the inclusion criteria, so the final sample comprised 259 patients. Of these patients, 98 corresponded to the pre-intervention period (4 epidemic seasons, 2006–2010) and 131 to the post-intervention period (6 epidemic seasons, 2011–2017) (Fig. 1). The median age of the patients in the sample was 2.5 months (IQR, 1.5–5), with a predominance of the male sex (56.8%). As can be seen in Table 1, patients admitted for AB in the post-intervention period were younger compared to previous years. The mean length of stay of patients in the pre-intervention period was of 4.9 days (SD, 2.4), which was longer compared to the mean stay of the entire post-intervention group, of 4.3 days (SD, 2.1), although this difference was not statistically significant (P=.09). Table 1 presents these and other demographic characteristics of the sample. We found no differences between the periods in the frequency of readmission: there were 2 readmissions in the pre-intervention period and another 2 in the post-intervention period.

Figure 1.

Flowchart of the selection of patients and their distribution in the pre- and post-intervention groups. GA, gestational age.

(0.12MB).

When we compared the percentage of patients in whom specific resources were used in the pre-intervention and transition periods, we found that the only resource in which there had been a reduction was the chest radiograph, as can be seen in Fig. 2. However, when we compared the pre-intervention group with the post-intervention group, we found reductions in the use of every resource targeted in the intervention. This trend was observed both in the comparison of the two periods overall (Fig. 3) and in the comparison of the use of resources year by year after the intervention (Table 2). Thus, we see a shift from the near-routine performance of a chest X-ray in 73.5% of infants admitted with AB to performance of X-rays in only 3.7%, 7.7% and 4% of these patients in years 4, 5 and 6 after the 2-step educational intervention. We found a similar pattern in the use of other diagnostic resources, including the CBC, measurement of CRP levels and blood culture. However, we did not find any reduction in the use of other tests, such as blood gases and detection of RSV in nasal secretions. Last of all, we ought to highlight that the use of some resources started increasing again several years after the 2-step educational intervention, specifically in the 2015–16 seasons, that is, 5 years after the implementation of the CPG.

Figure 2.

Percentage of use of diagnostic resources in the pre-intervention period compared to the transition period.

CBC, complete blood count; CRP, C-reactive protein; RSV, respiratory syncytial virus.

(0.11MB).

Figure 3.

Percentage of use of diagnostic resources in the pre-intervention period compared to the post-intervention period.

CBC, complete blood count; CRP, C-reactive protein; RSV, respiratory syncytial virus.

(0.11MB).

Clinical practice guidelines are among the most important tools developed to improve the quality and efficiency of health care, translating rigorous scientific evidence into measures that have clinical impact.20 Our study demonstrated that the dissemination of knowledge in health care institutions through a structured approach using direct, repeated and multimodal interventions addressed to physicians, with presentation of indicators of the quality of the care currently delivered and emphasising the specific aspects where modifications were being pursued, can achieve a sustained reduction in the use of diagnostic resources that are not recommended by CPGs,17–19 such as chest radiography, CBC, CRP measurement and blood culture, without a concomitant reduction in the use of resources not targeted in the intervention, such as venous blood gases and detection of RSV in nasal secretion samples. In addition to this reduction in the use of diagnostic resources, we also observed a decrease in length of stay in our patients, although the study design did not allow us to establish causality for this association. This reduction in the use of diagnostic resources and the associated decrease in the length of stay did not result in an increased readmission rate during the post-intervention period.

An interesting finding in our study was that while an initial intervention was implemented in October 2010, we did not start observing the desired effects until the intervention was repeated and feedback provided a year later, in October 2011. The effects did not only exceed those observed in other research projects that assessed the impact of the publication of CPGs without further intervention,14,15,21,22 but were also superior to those obtained with other structured interventions with dissemination of the evidence included in CPGs through a multimodal approach and with specific goals. Such is the case of the study by Mittal et al., where CPG implementation achieved a reduction in the use of chest X-rays from 59.7% to 39%, a significant difference, although smaller than the one observed in our study,23 or the study of Ralston et al., which found a decrease in the use of this resource of 36% compared to the pre-intervention period.24

Some of the reasons that physicians do not adhere to the recommendations of CPGs include lack of awareness of their existence, lack of familiarity with their use in actual practice, the excessive length and complexity of the CPG documents, lack of time to devote to the individual reading and summarising of their contents, disagreement with some of their recommendations, negative expectations regarding the clinical outcomes of their application, internal and external institutional barriers that impede their implementation on a large scale and difficulty overcoming the inertia of habit and continuing to deliver care using the approach learned during medical education and practiced for several years.25–27

On the other hand, to our knowledge there are no reports in the literature of the waning of the beneficial effects obtained with the implementation of CPGs. This may be because other studies in the literature did not have long follow-up periods, unlike our study. Our hypothesis is that this type of intervention requires regular reinforcements, and after the 2-step intervention we have described, there were no others. Also, there were no substantial changes in the medical staff of the clinical departments that participated in the study that would explain the increase in the use of some of the diagnostic resources that followed the initial reduction. However, further research focused on this phenomenon would be necessary to obtain a plausible explanation.

Some of the strengths of our study are that it had a prospective design for the majority of its duration, which allowed the assessment of outcomes at regular intervals and therefore implementation of corrective measures where needed, and that it analysed 11 epidemic seasons, with prospective follow-up in 8 of them, which we consider a long-term follow-up.

We believe that among the measures implemented in pursuit of our objectives, the one that may have had the greatest impact could be having made the physicians in the emergency setting the main target of our intervention, as the ordering of the initial diagnostic tests has always been their responsibility. We believe that having designed an educational intervention particularly targeted to them, providing them with a summary of the management protocol, and informing them of the outcomes of the assessment of the use of resources before and after the first intervention (October 2010) as positive feedback for their clinical practice were crucial to the success of the project.

Although this does not constitute a strength of our study, we would like to highlight the considerable difference in the number of chest radiographs performed before and after the 2-step intervention in the absence of negative consequences such as increases in misdiagnosis or readmission, which increased patient safety in our infant population by reducing their exposure to ionising radiation.

However, one limitation of our study was its small sample size, which was due to the small size of the paediatric population in the catchment area of our hospital.

Another aspect to consider is that the success of our project may have been due, at least in part, to the small size of the health care teams that manage AB in our hospital, both in the emergency department and in the paediatrics ward. This small size favours the development of closer interprofessional relationships and facilitates the attendance of a larger proportion of physicians to educational activities. For this reason, it may not be possible to generalise the results of our study to large-scale hospitals with much larger staffs. Nevertheless, and considering the favourable outcomes of our intervention, we propose that similar interventions be implemented in hospitals and primary care centres in Spain in order to assess their effectiveness in improving efficiency in the management of acute bronchiolitis at every level of care.

The dissemination of the scientific evidence published in CPGs by a structured, multimodal, and direct approach addressed to physicians, setting specific goals and repeatedly providing positive feedback, can achieve a reduction in the use of diagnostic resources in the management of AB that have been specifically targeted in the intervention, such as chest X-rays, CBCs, CPR tests and blood culture without an associated increase in the readmission rate in small and mid-sized hospitals. These interventions can increase the efficiency of health care facilities, which will in turn be able to offer safer and higher-quality care to their patients.

The authors have no conflicts of interest to declare.