Overuse of unnecessary medical services is pervasive in medicine accounting for $210 billion in a 2010 report by the Institute of Medicine.1 Viral bronchiolitis is a significant health burden with an estimated 132,000 to 173,000 associated hospitalizations due to respiratory syncytial virus (RSV) alone.2 The cost of bronchiolitis-associated outpatient encounters and hospitalizations contributes to hundreds of millions of dollars annually in health care costs.3 Bronchiolitis represents the third most common admission diagnosis at our institution.
Available Knowledge and Rationale
The clinical presentation of bronchiolitis overlaps that of other respiratory illnesses, which explains some of the diagnostic and therapeutic interventions that providers have used in the past.4 Evidence has failed to support the routine use of therapies such as bronchodilators,5 steroids,6 and hypertonic saline7 for bronchiolitis in reducing its morbidity or the need for hospitalization. Diagnostic procedures such as chest radiography (CXR) and blood cultures have shown limited utility because of the low incidence of coexistent bacterial infections such as pneumonia or bacteremia in patients with bronchiolitis.8–10 Diagnostic and therapeutic interventions that lack scientific evidence constitute waste and/or overtreatment and are estimated to represent one-third of all health care–related costs in the United States.11 Four of the 5 Choosing Wisely recommendations of The Society for Hospital Medicine–Pediatrics Section address management practices applicable to patients with bronchiolitis.12,13 In 2014, the American Academy of Pediatrics published its Bronchiolitis Guideline advising against the routine use of bronchodilators, CXR, and viral and laboratory testing in patients with bronchiolitis.14 In that same year, Children's Healthcare of Atlanta (CHOA) initiated a multidisciplinary quality improvement (QI) project (involving attending physicians, nurses, respiratory therapists, and residents) to improve the management of bronchiolitis patients across urgent care (UC) and emergency department (ED) service areas where our bronchiolitis patients receive care.
Our project's global aim was to reduce the use of diagnostic and therapeutic interventions demonstrated not to impact the clinical course and outcomes of patients with bronchiolitis by the end of the project's first bronchiolitis intervention season in 2015. In addition, we sought to sustain and improve upon the first season's gains during the second and third bronchiolitis seasons of this project. We also aimed to study the impact of these QI interventions on the cost of care for patients with bronchiolitis.
This QI project was conducted in the UC and ED service areas of CHOA, a 575-bed health care system comprising 2 tertiary hospitals. Our 5 UC centers (located around the metropolitan Atlanta area) and 2 EDs had 125,076 and 175,536 visits, respectively, in 2016. One hundred seventy-one physicians (84 in UC and 87 in the ED) participated in this project; these physicians also supervised pediatric and emergency medicine residents from the Emory and Morehouse School of Medicine. This QI project was not considered human subjects research, and per institutional policy, approval from the institutional review board was not required.
This QI project was implemented system-wide in September 2014. The data for analysis were categorized by bronchiolitis season, which extends from September to April, when more than 90% of bronchiolitis cases are seen in our system. Baseline data include bronchiolitis seasons (September–April) from 3 consecutive years (2011–2014). Postimplementation data include bronchiolitis seasons beginning September 2014 through April 2017.
Patients aged 1 to 18 months with a final primary International Classification of Diseases, Ninth Revision diagnosis of bronchiolitis (466.1, 466.11, and 466.19) were included for analysis (corresponding International Classification of Diseases, Tenth Revision diagnoses used as of October 1, 2015). For the bronchodilator use measure, we included the younger subset of patients 1 to 12 months of age because of the higher prevalence of asthma in our community in patients older than 12 months, in whom a trial of albuterol may be helpful in some situations (such as history of wheezing, atopy or strong family history of asthma). We included both patients discharged home as well as those admitted to our hospital's general pediatric wards in our analysis. We excluded patients with previously diagnosed cardiopulmonary and chronic conditions (information available by request) to select for a relatively healthy population to which our bronchiolitis guideline was applicable. In addition, patients with a diagnosis of pneumonia or asthma were also excluded.
The CHOA Bronchiolitis taskforce consisted of clinicians from outpatient service areas involved in the care of patients with bronchiolitis (UC and ED physicians, respiratory therapists, nurses, and QI process improvement specialists). This team developed area-specific SMART (Specific, Measurable, Achievable, Relevant, Time Bound) aims. Key driver diagrams were created to determine primary drivers influencing management of bronchiolitis patients; these were used to guide system interventions to achieve area-specific aims as well as to address barriers to improvement. Figure 1 shows an example of the key driver diagram focused around the reduction of albuterol for patients with bronchiolitis.
Our interventions centered on education, integration of guideline recommendations around clinician workflows and periodic feedback to sustain momentum and encourage continuous improvement:
- Guideline development and implementation: During the first half of 2014, we updated our existing bronchiolitis guideline to reflect the anticipated 2014 American Academy of Pediatrics Bronchiolitis Guideline. Our updated guideline, clinical evidence, and order set were integrated into the electronic medical record (EMR) in September 2014 to allow for access during patient care. The EMR changes included easy linkage to the bronchiolitis guideline from within the order set as well as removal of all bronchodilators from the available medications within the bronchiolitis order set (although physicians could still order bronchodilators, they would have to go outside the order set, making the undesired action more difficult to do).
- Education: Members of the Bronchiolitis taskforce conducted educational sessions describing the aims of the QI project several months before implementation of our new bronchiolitis guideline. These sessions were conducted with daytime and evening clinical staff and included a review of the current literature and evidence together with the proposed guideline changes and a discussion of the process and outcome metrics goals. They were led by physician, nursing, and respiratory therapy clinical leaders in their respective areas and also targeted pediatric residents. During these sessions, frontline providers identified potential barriers to the success of the project. Emergency department physicians highlighted the need to educate midlevel ED providers who sometimes initiate care before a physician evaluating a patient. Providers voiced that parent expectations of care (influenced by community pediatricians) were a primary driver of resource utilization in our UC and ED. We therefore also included education targeted toward community pediatricians via a bimonthly newsletter where we shared the published clinical evidence and the planned UC and ED management guidelines, together with the project's timeline and outcomes of interest.
- Performance measurement and feedback: We tracked bronchiolitis clinical performance metrics through a data registry (Qlikview)15 dashboard updated on a daily basis. This allowed for access to real-time clinical performance data in a format structured around the project's SMART aims so we could implement rapid improvement cycles to modify processes to help change provider behavior and rapidly impact outcomes. In addition, representatives from both the ED and UC provided their teams with quarterly feedback regarding their progress toward the project's aims. This feedback was provided both at the individual and group levels. The first data feedback occurred 2 months after the project's launch and then every 1 to 3 months.
- In 2015, to augment physician engagement in this QI project, we instituted an American Board of Pediatrics–approved Maintenance of Certification (MOC) Part 4 project on the management of bronchiolitis. As part of the project, 72 participating ED physicians and 22 UC physicians reviewed individual and group performance for the project's measures; this helped to increase awareness of current performance relative to the goals.
- A final component of our intervention was inclusion of this project's QI aims in the quality goals of UC and ED service areas, representing a small financial incentive for physicians effective only during the calendar year 2015
System and Service Area Specific Measures
Urgent care and ED service areas shared the SMART aim of reducing the percentage of bronchiolitis patients 1 to 12 months of age treated with albuterol to less than 42% (UC) and 32% (ED) by the end of the QI project's first bronchiolitis season (October 2014–April 2015). These targets represented attainable goals of reducing albuterol usage by 15% to 20% from baseline. We defined albuterol usage as any (≥1) doses of this bronchodilator. This definition is more stringent than some other studies where albuterol usage was defined as administration of more than one bronchodilator dose allowing for a “trial of albuterol.” For UC, 2 additional SMART aims were established related to reducing CXR utilization and RSV viral testing. For the ED, one additional SMART aim was developed to reduce CXR use. Our EDs could only order a multiplex viral and bacterial panel and not an RSV test alone, making it challenging to establish the clinical intent for such testing; for this reason, we did not target viral testing as an ED measure. We manually reviewed 50 patient medical records to ensure the validity and accuracy of our electronic data capture and the appropriate application of inclusion and exclusion criteria. As balancing measures, we reviewed admission rates from the ED and also admission rates after return visits to the ED within 48 hours of discharge to compare preimplementation and postimplementation rates to assess for any unintended impact our interventions may have had.
Because of a change in our billing system in 2014, to ensure valid cost comparisons across the 6 bronchiolitis seasons in the study period, we based variable cost savings calculations using average costs in place during the preimplementation period and applied those to the reduction in resources observed during the postimplementation period. Because of the lag in availability of financial data, cost analysis includes the first 2 postimplementation seasons.
We used statistical process control methods to analyze the impact of our project on the system and area-specific quality measures. Attribute data (monthly and quarterly resource use measures) were analyzed with Shewhart p-charts. We chose 2 SDs from the centerline as the upper and lower control limits for our statistical process control charts and used established special cause variation definitions to detect changes in our system.16
We also compared preintervention and postintervention measures, including balancing measures, for which we used χ2 tests. Analysis was done in SAS EG 7.1 (SAS Institute, Inc, Cary, NC). All statistical tests were performed as 1-sided tests with a 0.05 level of significance.
There were 3271 preimplementation and 3901 postimplementation visits in UC and 5319 preimplementation and 6169 postimplementation visits in the ED. Table 1 shows the key patient sociodemographic characteristics during the preimplementation and postimplementation periods. Patient visits for bronchiolitis in our EDs and UCs increased across the 2 periods we studied, as also did the annual ED and UC total volumes.
Albuterol usage, a key measure in both ED and UC, decreased significantly throughout the study period (Table 2, Fig. 2). Targets for UC and ED CXR utilization and UC viral testing were also achieved (Table 2; Figs. 3, 4), although the opportunity for improvement in these metrics was more challenging owing to relatively low preimplementation rates. In all metrics across both service areas, improvements noted in the first postimplementation season were sustained or even enhanced in the second and third bronchiolitis seasons after this QI project's implementation. Rates of ED admission as well as admission after return to the ED within 48 hours of discharge remained unchanged at 2% before and after project implementation.
Direct cost of care in UC and ED decreased by $57,127 and $225,907, respectively, during the first 2 postimplementation bronchiolitis seasons. In both service areas, the greatest impact occurred through a reduction in the cost of supplies and administration of inhalational treatments ($20,915 in UC and $157,837 in the ED). We also observed reductions in both settings for costs of CXR, viral testing, and facility.
This QI project resulted in significant improvements in evidence-based management for patients with bronchiolitis (reduced albuterol, CXR, and viral testing) in the EDs and UCs of our health care system. This was associated with a reduction in the cost of care. All of these achievements exceeded initial SMART aim targets and continued to improve for 3 bronchiolitis seasons without negatively impacting rates of inpatient admission or return after discharge from the ED.
Our study is similar to a recently published study by Tyler et al17 in design, measures and interventions. However, we exceeded all of our goals and sustained improvements for 3 bronchiolitis seasons. Our study differed from past studies on improvements in care of bronchiolitis patients in that we included patients who received any albuterol, whereas some previous studies have not counted a first “trial” dose of albuterol. Despite this, the reduction we observed in bronchodilator use was comparable with or greater than that in previous ED studies.18–21 Furthermore, the reduction in CXR utilization we observed was greater than that observed in studies previously reported.19
The success of this QI project was based on the engagement of a multidisciplinary team involving all providers (physicians, nurses, respiratory therapists) caring for bronchiolitis patients and representing all outpatient service areas (ED and UC) throughout the health care system where bronchiolitis patients are seen. Although the project was a system initiative, it was driven by leaders on the frontlines of care. Respiratory therapists played an important role in this project and ensured clinical documentation of any benefit of bronchodilator therapy or lack thereof, thus limiting multiple concurrent interventions of ineffective therapies.
As noted in other studies, buy-in from all team members caring for a given patient with bronchiolitis was a prime factor in the success of this project.22 In addition to interpersonal factors that led to the acceptance of evidence-based care, structural issues such as implementing changes within our EMR and access to real-time performance data also helped in the project's success. Although the rollout of our updated bronchiolitis guideline with its associated order sets was an important intervention in our QI project, guidelines alone (specifically bronchiolitis guidelines with similar recommendations to ours) have not always shown significant reductions in test ordering or use of ineffective treatments.23 Rather than relying solely on guideline implementation, in this project, we engaged frontline clinicians to enable culture change and used the project's leaders as catalysts to drive improvement.
Other factors contributing to success of this project included providing physicians with MOC Part 4 credit, a nonmonetary benefit that was valuable as a meaningful board recertification activity. Access to real-time clinical performance data made it possible to modify processes and change provider behavior to rapidly impact outcomes. For example, during the first bronchiolitis season, we identified an outlier UC center with a significantly higher than average rate of RSV testing and albuterol usage. A review of that UC center's data led to the identification of a single individual physician's practice pattern, which allowed us to intervene and provide targeted education to that physician. This resulted in the outlier UC center rapidly achieving performance goals during the second season.
One of our interventions was a small quality financial incentive for groups meeting certain specific aims of this QI project. Some evidence suggests that the removal of financial incentives can result in a significant erosion of performance gains24 and that overreliance on them can “crowd out” intrinsic good professional behaviors.25 However, if used judiciously in a setting to augment QI efforts over which physicians have control of the outcomes, financial incentives may have a positive impact for some individuals. In our project, the overall size of the incentive provided was relatively small (<$2500 per physician for the first year only). This incentive may have had an impact on physician behavior (including an anticipatory influence). However, we believe that its contribution was relatively small when compared with the many other interventions in this project such as education, EMR changes, MOC project and performance feedback (which we provide regularly to ED physicians in our system).26 This is supported by the sustained performance improvement in the second and third postimplementation bronchiolitis seasons when the financial incentive was not provided.
There is tension between current reimbursement models and QI efforts because delivering high-quality care while reducing resource utilization can have a negative financial impact on health systems. In Georgia, there are no financial incentives for lowering costs by reducing utilization and testing. As a result, health systems like ours must balance the impact of delivering care more efficiently with the decision and timing to enter value based contracts.
Some limitations of this QI project include reproducibility in smaller community hospitals with varying degrees of availability of EMR data and limited access to a dedicated QI staff. However, the work by Ralston et al20 on reducing unnecessary bronchiolitis care in community hospitals shows that adherence to evidence-based practices around bronchiolitis can be achieved with collaboration between large children's hospitals and community hospitals. It is also possible that variation in the severity of different bronchiolitis seasons could have impacted our results, although having 3 preimplementation seasons as baseline and 3 postimplementation seasons makes this less likely. In addition, our bronchiolitis admission rates and rates of admission within 48-hour of discharge from ED return rates remained unchanged despite lower rates of both diagnostic and therapeutic interventions. A limitation of our cost methodology is that the cost calculations are based only on variable costs and not fixed costs, which may be the major determinant of long-term cost savings for an institution.
Our multidisciplinary project using QI methodology and tools demonstrated significant and sustained improvements in evidence-based management of bronchiolitis patients across an entire hospital system's outpatient settings. The reduction in unnecessary diagnostic and therapeutic interventions resulted in improved health care value to patients by reducing resource use while improving the quality of care.
1. Institute of Medicine (US) Roundtable on Evidence-Based Medicine, Yong PL, Saunders RS, Olsen LA, eds. In: The Healthcare Imperative: Lowering Costs and Improving Outcomes: Workshop Series Summary
. Washington, DC: National Academies Press (US); 2010. Available at: http://www.ncbi.nlm.nih.gov/books/NBK53920/
. Accessed September 12, 2018.
2. Stockman LJ, Curns AT, Anderson LJ, et al. Respiratory syncytial virus-associated hospitalizations among infants and young children in the United States, 1997–2006. Pediatr Infect Dis J
3. Paramore LC, Ciuryla V, Ciesla G, et al. Economic impact of respiratory syncytial virus-related illness in the US: an analysis of national databases. Pharmacoeconomics
4. Florin TA, Byczkowski T, Ruddy RM, et al. Variation in the management of infants hospitalized for bronchiolitis
persists after the 2006 American Academy of Pediatrics bronchiolitis
guidelines. J Pediatr
5. Gadomski AM, Brower M. Bronchodilators for bronchiolitis
. Cochrane Database Syst Rev
6. Fernandes RM, Bialy LM, Vandermeer B, et al. Glucocorticoids for acute viral bronchiolitis
in infants and young children. Cochrane Database Syst Rev
7. Brooks CG, Harrison WN, Ralston SL. Association between hypertonic saline and hospital length of stay in acute viral bronchiolitis
: a reanalysis of 2 meta-analyses. JAMA Pediatr
8. Schuh S, Lalani A, Allen U, et al. Evaluation of the utility of radiography in acute bronchiolitis
. J Pediatr
9. Titus MO, Wright SW. Prevalence of serious bacterial infections in febrile infants with respiratory syncytial virus infection. Pediatrics
10. Chao JH, Lin RC, Marneni S, et al. Predictors of airspace disease on chest x-ray in emergency department patients with clinical bronchiolitis
: a systematic review and meta-analysis. Acad Emerg Med
11. Berwick D, Hackbarth A. Eliminating waste in the U.S. healthcare. JAMA
13. Quinonez RA, Garber MD, Schroeder AR, et al. Choosing wisely in pediatric hospital medicine: five opportunities for improved healthcare value. J Hosp Med
14. Ralston SL, Lieberthal AS, Meissner HC, et al. Clinical practice guideline: the diagnosis, management, and prevention of bronchiolitis
15. Qlikview®, Version <11.2SR10>
. Lund, Sweden: QlikTech International AB; 1993–2006.
16. Nelson L. The Shewhart chart-test for special causes. J Qual Technol
17. Tyler A, Krack P, Bakel LA, et al. Interventions to reduce over-utilized tests and treatments in bronchiolitis
18. Mussman GM, Lossius M, Wasif F, et al. Multisite emergency department inpatient collaborative to reduce unnecessary bronchiolitis
. 2018;141:pii: e20170830.
19. Kotagal UR, Robbins JM, Kini NM, et al. Impact of a bronchiolitis
guideline: a multisite demonstration project. Chest
20. Ralston SL, Garber MD, Rice-Conboy E, et al. A Multicenter collaborative to reduce unnecessary care in inpatient bronchiolitis
21. Akenroye AT, Baskin MN, Samnaliev M, et al. Impact of a bronchiolitis
guideline on ED resource use and cost: a segmented time-series analysis. Pediatrics
22. Ralston SL, Atwood EC, Garber MD, et al. What works to reduce unnecessary care for bronchiolitis
? a qualitative analysis of a national collaborative. Acad Pediatr
23. Mittal V, Hall M, Morse R, et al. Impact of inpatient bronchiolitis
clinical practice guideline implementation on testing and treatment. J Pediatr
24. Petersen LA, Simpson K, Pietz K, et al. Effects of individual physician-level and practice-level financial incentives on hypertension care: a randomized trial. JAMA
25. Frey B. Not Just for the Money; An Economic Theory of Personal Motivation
. Cheltenham, UK: Edward Elgar Publishing; 1997.
26. Jain S, Frank G, McCormick K, et al. Impact of physician scorecards on emergency department resource use, quality, and efficiency. Pediatrics
Keywords:Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
bronchiolitis; quality improvement; value care; urgent care