hosppeds
December 2018, VOLUME8 /ISSUE 12

Discharging Asthma Patients on 3-Hour β-Agonist Treatments: A Quality Improvement Project

  1. Huay-ying Lo, MDa,
  2. Amanda Messer, MDb,
  3. Jennifer Loveless, MPHc,
  4. Esther Sampayo, MD, MPHd,
  5. Robert H. Moore, MDd,
  6. Elizabeth A. Camp, PhDe,
  7. Charles G. Macias, MD, MPHc,e and
  8. Ricardo Quinonez, MDa
  1. aPediatric Hospital Medicine,
  2. cEvidence-Based Outcomes Center,
  3. dSection of Pulmonary Medicine, and
  4. eSection of Emergency Medicine, Texas Children’s Hospital and Baylor College of Medicine, Houston, Texas; and
  5. bPediatric Hospital Medicine, Louisiana State University Health Sciences Center and Children’s Hospital, New Orleans, Louisiana
  1. Address correspondence to Huay-ying Lo, MD, Division of Pediatric Hospital Medicine, Texas Children’s Hospital, 17600 Interstate 45 South, Suite WL 460, The Woodlands, TX 77384-4170. E-mail: hxlo{at}texaschildrens.org
  1. Drs Lo and Messer conceptualized and designed the study, oversaw improvement activities, coordinated and supervised data collection and analysis, and drafted the initial manuscript; Ms Loveless conceptualized and designed the pilot study and oversaw improvement activities; Drs Sampayo, Moore, and Macias assisted with improvement activities and review of the manuscript; Dr Camp performed statistical analysis; Dr Quinonez participated in the design of the study and review of the manuscript; and all authors critically reviewed the manuscript and approved the final manuscript as submitted.

Abstract

OBJECTIVES: Asthma exacerbations are a leading cause of hospitalization among children. Despite the existence of hospital protocols and national guidelines, little guidance is available regarding appropriate short-acting β-agonist (SABA) frequency discharge criteria. Our aim was to reduce the median length of stay (LOS) for children hospitalized with asthma exacerbations by 4 hours by changing the discharge requirement SABA frequency.

METHODS: Multiple plan-do-study-act cycles based on findings in our key driver diagram were used to decrease LOS. Our primary intervention was reducing the SABA administration frequency discharge requirement from every 4 hours to every 3 hours. After a feasibility pilot, this change was implemented throughout the hospital. Our intervention bundle included updating our evidence-based guidelines, electronic health record order sets and note templates, house-wide education, and a new process for respiratory therapists to notify physicians of discharge readiness. Our primary metric was LOS, with 3-, 7-, and 14-day same-cause emergency department (ED) revisits and hospital readmissions as balancing metrics. Statistical process control charts and nonparametric testing were performed for data analysis.

RESULTS: Median hospital LOS was significantly lower in the postintervention period compared with the preintervention period (30.18 vs 36.14 hours respectively; P < .001). Statistical process control charts indicated special cause variation was achieved. No significant differences were observed in rates of ED revisits or hospital readmissions.

CONCLUSIONS: Reducing the discharge requirement of SABA frequency from every 4 hours to every 3 hours resulted in a reduction in LOS, with no increase in ED recidivism or hospital readmission rates.

In 2010, pediatric asthma-related admissions accounted for >136 000 hospital inpatient discharges in the United States.1,2 Rising health care costs have prompted initiatives to improve discharge timeliness without compromising patient safety. Multiple global initiatives have been used to advocate for the use of structured protocols to guide asthma management.35 However, the 2007 National Heart, Lung, and Blood Institute asthma guideline does not specify an optimal short-acting β-agonist (SABA) administration frequency that signifies discharge readiness.6

Although the use of clinical pathways for pediatric inpatient asthma admissions has been shown to reduce length of stay (LOS), and many children’s hospitals have evidence-based asthma guidelines, SABA-related discharge criteria vary across children’s hospitals in the United States.713 Previous studies have revealed that using quality improvement (QI) techniques to standardize discharge criteria improves discharge efficiency and reduces LOS.14

Limited data exist to support discharging patients on every–3-hour (Q3H) SABA treatments. The few studies that have been used to address this are international studies with small sample sizes.15,16 Similar to other institutions, when our hospital created evidence-based clinical care guidelines for asthma, patients were required to tolerate every–4-hour (Q4H) SABA treatments before discharge.913 Because this practice was based on expert consensus, our objective was to reduce LOS for asthma patients by testing the feasibility and implementation of changing our SABA discharge readiness criteria from Q4H to Q3H treatments. We hoped that changing the SABA frequency discharge requirement would decrease median hospital LOS by 4 hours, without increasing emergency department (ED) revisit or hospital readmission rates and that with our results, we would add to the body of literature regarding SABA discharge criteria and inform the development of future clinical practice guidelines.

Methods

This was a QI study in which we used the method for improvement17 and was part of a larger institutional effort to improve care for pediatric asthma patients.

Context

This study was conducted at a 650-bed freestanding, quaternary pediatric hospital in the southwestern United States with ∼114 000 ED visits and 32 000 hospital admissions per year. Patients at the medical center campus hospital (main campus) and a smaller, community-based satellite hospital (west campus) were included. At both sites, the majority of asthma patients are admitted to dedicated inpatient pediatric hospital medicine (PHM) teams.

Interventions

Our institution has a multidisciplinary asthma outcomes improvement team composed of physicians (PHM, critical care, pulmonary), respiratory therapists (RTs), nurses (ED, acute care), pharmacists, outcomes analysts, and support staff whose focus is to improve care for asthma patients. This team created a key driver diagram to identify potential barriers to changing our SABA discharge frequency criteria from Q4H to Q3H (Fig 1), and a bundle of interventions was implemented on the basis of their findings.

FIGURE 1

Key driver diagram for reduction in LOS.

Pilot

Given the lack of robust evidence supporting Q3H discharge criteria, a feasibility pilot was conducted on 1 acute care unit at our main campus from May 15 to July 24, 2013. Written and verbal information on the initiative was disseminated to attending physicians, residents, RTs, and nurses on the pilot unit by members of the outcomes improvement team. During the pilot, a decrease in time from administration of the first Q3H SABA to discharge (17.7 hours preintervention compared with 11.5 hours postintervention; P = .01) was observed. Follow-up phone calls to families yielded no revisits to outside facilities, and there was no increase in internal ED revisits or hospital readmissions.

Plan-Do-Study-Act Cycle 1

On the basis of the data from our feasibility pilot, the decision was made to implement the following new discharge criteria for asthma patients throughout the hospital: stable on room air, receipt of 2 doses of SABA spaced to Q3H, and clinical respiratory score (CRS) of <3. The CRS is an internally validated institutional score that represents respiratory acuity. Score components include respiratory rate, auscultation, accessory muscle use, mental status, room air oxygen saturation, and color, with each measure scored 0 to 2 with a maximum possible score of 12.13

This change in discharge criteria was updated within our institution’s evidence-based clinical guideline and order sets within the electronic health record (EHR) Epic (Epic Systems Corporation, Verona, WI). The updated guideline was distributed hospital-wide via e-mail. A conditional discharge order was added to the asthma order set within the EHR describing the new discharge criteria.

Plan-Do-Study-Act Cycle 2

Our second plan-do-study-act (PDSA) cycle was focused on house-wide education of physicians, RTs, and nursing staff regarding the new discharge criteria. Badge cards detailing the changes were distributed to residents, and information about the initiative was added to PHM resident orientation. Asthma outcomes improvement team members disseminated the new guideline and discharge criteria to their respective sections via in-person staff meetings.

PDSA Cycle 3

Our third PDSA cycle focused on the creation and dissemination of a specific asthma history and physical (H&P) template within our EHR. This H&P template was designed to be user friendly and incorporated asthma-related history, assessment, and plan elements intended to promote physician use. The new asthma discharge criteria were delineated in the H&P template as a reminder to physicians.

PDSA Cycle 4

Our final PDSA cycle involved a new process in which the RT called the bedside nurse or physician when patients had received their second Q3H treatment, to inform providers of discharge readiness. Although we observed significant uptake of the new discharge criteria with our previous interventions, our team encountered a lack of timely notification to providers when these criteria were met. Our institution’s SABA-weaning protocol is almost entirely RT driven and does not require physician reassessment during each phase of the weaning process. Thus, physicians were often unaware when patients had reached their second Q3H treatment and were ready for discharge.

Over time, as care teams became more accustomed to the Q3H discharge standard, this process naturally evolved so that RTs now typically call after the first Q3H treatment to inform providers of anticipated discharge readiness.

Study of the Interventions

Automated tracking of data relevant to our outcomes for asthma patients, including LOS, readmissions, and demographic data, was outputted via an enterprise data warehouse (EDW) (Health Catalyst, Salt Lake City, UT) that incorporates near-time data from the Epic EHR.

To track outcomes, an asthma patient cohort was developed. The following inclusion criteria were used: age ≥2 years, administration of ≥3 β-agonists and ≥1 systemic oral or intravenous steroid during the hospitalization, and an International Classification of Diseases, Ninth Revision or International Classification of Diseases, 10th Revision code for asthma or wheezing as 1 of the top 4 discharge diagnosis codes. Exclusion criteria included patients with significant comorbidities including airway abnormalities, sickle cell disease, renal disease, solid-organ transplants, cancer, interstitial lung disease, cystic fibrosis, ventilator dependence, and cardiac disease. We also excluded patients with a diagnosis of acute respiratory infection, such as tuberculosis, pneumonia, or bronchiolitis.

For data analysis, the preintervention period was defined as October 2011 through April 2013 and the postintervention period as October 2013 through April 2015. The 5-month interval between the pre- and postintervention periods (May through September 2015) encompasses the intervention period and was excluded from data analysis. This also allowed for parallel months between the pre- and postintervention groups, to account for seasonal changes in asthma admission rates.

Measures

Our primary outcome measure was hospital LOS, which was defined as the time from ED triage to the time a discharge order was placed. ED revisits and hospital readmissions for a related diagnosis were tracked as balancing metrics. Readmissions were tracked at 3, 7, and 14 days from discharge.

For LOS data, manual chart review was performed on a random sampling of 10 charts to confirm the above inclusion criteria and to ensure accurate timestamps were obtained via the EDW application. For hospital readmissions, 2 investigators (H.L. and A.M.) independently reviewed all readmission charts, and patients who were readmitted for unrelated diagnoses were excluded.

Statistical Analysis

For all descriptive comparisons, the Pearson χ2 test was used to detect statistically significant differences for categorical variables; Fisher’s exact test was used if cell values were <5. The Mann–Whitney U test was used to find significant differences between continuous variables (cofactors and outcome), given nonnormal data distribution. Statistical significance was defined as a P value <.05. All analyses were conducted by using Statistical Package for the Social Sciences (SPSS) version 23 software (IBM SPSS Statistics, IBM Corporation, Armonk, NY).

A statistical process control (SPC) chart was created to track LOS and identify instances of special cause variation that would indicate a likely change to the system. SPC charts were created by using Minitab 17 Statistical Software (Minitab Inc, State College, PA).

Ethical Considerations

This study was approved by the hospital’s institutional review board.

Results

Population

Of 3261 pediatric asthma patients admitted during the study period, 2909 met our study inclusion criteria (1530 [52.6%] in the preintervention period and 1379 [47.4%] in the postintervention period). When comparing demographic factors between pre- and postintervention populations, the only significant difference was for the first admitted inpatient department, with a higher percentage of patients in the postintervention group initially admitted to the ICU (P < .001) (Table 1).

TABLE 1

Comparison of Patient Demographics for Eligible Patients (N = 2909)

LOS

Median hospital LOS in hours was significantly lower in the postintervention timeframe compared with the preintervention timeframe (30.18 vs 36.14 hours; P < .001).

An Xbar-S SPC chart was created to track monthly average LOS throughout the study period (Fig 2). Special cause variation was achieved during the postintervention period on the basis of SPC chart rules. The average pre- and postintervention LOSs were 47.4 and 39.9 hours, respectively.

FIGURE 2

A, X-bar SPC chart of average LOS by month (pre- and postintervention). B, Chart of sample SD by month (pre- and postintervention). Tests were performed with unequal sample sizes. LCL, lower control limit; S, standard deviation; UCL, upper control limit; X, average LOS.

Sustainability

Continuous monitoring via the EDW revealed sustained rates of decreased median LOS for asthma patients for an additional 3 years beyond the initial study period (Fig 3).

FIGURE 3

Boxplot of median annual LOS (with median confidence interval bars) for hospitalized children with asthma.

Readmissions

After the exclusion of non–asthma-related readmissions, 46 patients across both study periods returned to the ED within 14 days of discharge (22 patients in the preintervention group and 24 in the postintervention group [P = .51]). Of the 46 patients who returned to the ED, 23 were readmitted.

Rates of ED revisits and hospital readmissions to the hospital were calculated at 3, 7, and 14 days postdischarge. χ2 testing for all 3 time points revealed no difference between pre- and postintervention groups (Table 2).

TABLE 2

Comparison of Readmissions at 3, 7, and 14 Days Pre- and Postintervention (N = 2909)

We also calculated the median time from discharge to ED revisit. The postintervention time of 86.4 hours was higher than the preintervention time of 56.5 hours; however, the difference was not significant in either Mann–Whitney U testing or simple linear regression. Because only 46 patients had 14-day ED revisits across both time periods, this analysis lacked statistical power to detect a significant difference. Post hoc calculations revealed <12% power to detect a difference between readmission rates pre- and postintervention.

When demographic and clinical factors were compared for readmitted versus nonreadmitted patients, it was noted that the initial CRS for nonreadmitted patients was actually higher, indicating a greater degree of respiratory distress, than readmitted patients (P < .01) (results not shown).

Discussion

Our data reveal that changing SABA frequency discharge criteria to Q3H instead of Q4H for asthma exacerbations decreased LOS without negatively impacting readmission or ED revisit rates. We successfully implemented this intervention in a large, quaternary academic center and community hospital and have sustained this reduction in LOS for 5 years after the initial intervention bundle. Compared with previous studies,15,16 our study includes a much larger sample size and revealed a statistically significant decrease in LOS.

Timely discharge of patients aligns with the Institute of Medicine’s 6 Domains of Health Care Quality; however, the benefit of efficiency must be balanced so that negative outcomes such as readmissions are not affected.18 Because children with asthma are already at a higher risk for subsequent hospitalization when compared with children without asthma,19 we tracked ED revisits and hospital readmissions to ensure that our efforts did not lead to undesired increases in these rates. A change to Q3H discharge criteria did not result in a difference in 3-, 7-, or 14-day readmission rates. We believe that 3-, 7-, and 14-day readmission rates capture a readmission that would have been impacted by changing the SABA frequency discharge criteria for the index admission. In fact, previous work at our institution suggests that these shorter-timeframe readmission rates are more likely to represent preventable readmissions than standard 30-day readmissions.20 However, as previously noted, our power to detect a difference in readmission rates was limited by the rarity of these events. On the basis of our study sample size, we had the power to detect a 2% increase in readmission rates, which was not observed.

With our results, we corroborate findings from previous studies suggesting that discharging asthma patients on Q3H SABA treatments is not associated with an increase in adverse patient outcomes. Authors of a randomized controlled trial of 63 pediatric patients in Australia in 1999 found that discharge on Q3H SABA treatments compared with Q4H treatments did not result in statistically significant differences in ED revisits, hospital readmissions, or mean hospital LOS.15 This study lacked statistical power because of small sample size, although the authors did approximate that a sample size of 500 may have revealed a significant 6-hour decrease in LOS. An additional Australian study in 2003 was a retrospective review of 419 pediatric asthma admissions.16 The investigators designated a theoretical time of discharge readiness on the basis of patients receiving at least 2 Q3H SABA doses and due for a third dose, no oxygen or intravenous fluid requirement, and a daytime discharge time. With their study, they concluded that this model to assess discharge readiness would have shortened LOS by an average of 5.5 hours and would not have increased ED or hospital readmissions. However, this study was limited by smaller sample size and was based on modeling rather than an actual intervention. Our study served to confirm some of this theoretical modeling by revealing a 6-hour reduction in LOS with the change in discharge criteria, without impacting revisit or readmission rates. To our knowledge, this represents the largest study to date on the impact of changing SABA frequency discharge criteria for children admitted with asthma.

Of note, there was a statistically significant higher number of patients initially admitted to the ICU in the postintervention compared with the preintervention period. However, we would expect that ICU admissions would generally prolong LOS. Because we demonstrated a decrease in LOS in the postintervention phase despite higher ICU admission rates, this may reflect additional effectiveness of our intervention.

There are several limitations to our study that are important to note. Despite having a large sample size, the infrequency of readmissions led to low statistical power in our ability to detect a difference in readmission rates. We hope our study will lead to multicenter collaborative studies that could be powered sufficiently to address this limitation and also will be used to examine 30-day readmission rates given that this is a national standard. Another limitation of our study is that our readmission rate only captures patients readmitted within our hospital system. However, our institution is the dominant children’s hospital in the region, with an extensive outpatient presence and an integrated EHR, which we believe should capture most revisits to the ED and hospital. Also, during our pilot study, follow-up phone calls were made to patients, and at least for that limited period, no patients were readmitted to a different hospital. Although we believe that discharging on Q3H SABA treatments is likely a safe practice, we did not have adequate numbers to definitively demonstrate this because of the rarity of safety events such as readmissions.

As is true for any QI study that uses a prepost design, our study cannot account for secular trends. Other QI interventions focused on improving care for hospitalized asthma patients occurred during the study period and may have also impacted LOS. Interventions included transition from prednisolone to dexamethasone as standard oral steroid therapy, efforts to decrease chest radiograph use, allowing usage of continuous albuterol on designated acute care floors, and efforts to improve influenza immunization rates among children hospitalized with asthma exacerbations. Additionally, there were other hospital-wide initiatives to improve discharge timeliness, such as the push and pull of patients from the ED and initiation of daily multidisciplinary discharge progression rounds, which were instituted during this time period and may have impacted LOS. However, as part of our continuous data monitoring efforts, we did not observe similar LOS decreases during this time period in diagnoses such as bronchiolitis, pneumonia, failure to thrive, or gastroenteritis. Further studies in which authors use more robust design and analysis such as randomized cluster designs are needed to account for secular trends and assess the impact of specific interventions aimed at improving the care of patients with asthma.

Despite these limitations, our institution has successfully adopted this change in SABA frequency discharge criteria for hospitalized asthma patients, with no increase in readmissions and a persistent trend of lower LOS.

Barriers

The most substantial barrier we faced involved changing a perceived existing standard of practice at our institution with regards to discharge SABA frequency criteria. The results of our feasibility pilot, in which we observed decreased LOS without increases in ED revisits or hospital readmissions, were instrumental in convincing providers to adopt this change. Our intervention bundle incorporated multiple methods to address these challenges, including modification of our evidence-based clinical guideline and creation of supporting EHR functions, such as conditional discharge orders within asthma order sets and the creation of an asthma H&P template that included the new discharge criteria. House-wide education was successful because it was led by members of an invested multidisciplinary team who disseminated the new guideline and discharge criteria among their peers. Another important barrier was our heavily RT-driven SABA-weaning protocol. Developing a process to enhance communication between providers, wherein RTs notified nurses and physicians of discharge readiness, was critical to our success in discharging patients after their second Q3H treatment. Finally, the ability to have real-time LOS and readmission data through our EDW application was essential to our success and ability to monitor for sustained improvement. Through the EDW, we have demonstrated a sustained decrease in LOS for hospitalized asthma patients for 5 years beyond the initial intervention period.

Conclusion

Changing discharge criteria for SABA administration frequency from Q4H to Q3H in hospitalized pediatric patients with acute asthma exacerbations resulted in decreased LOS and did not lead to increased asthma-related ED revisits or hospital readmissions.

Acknowledgments

We thank Terri Brown and other members of the Texas Children’s Hospital asthma outcomes improvement team for their QI efforts.

Footnotes

  • FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

  • FUNDING: No external funding.

  • POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

References