Sustaining Long-Term Asthma Outcomes at a Community and Tertiary Care Pediatric Hospital Free

The tertiary care hospital site is a 490-bed free-standing children’s hospital affiliated with a quarternary care academic medical center in a large metropolitan area with 24 399 admissions in 2019. The community hospital site is a 72-bed free-standing children’s hospital affiliated with the quarternary care academic medical center located in a suburban area with 7363 admissions in 2019. ACS at both sites were staffed primarily by university affiliated pediatric hospital medicine faculty, with no significant staffing changes during the CPG implementation at both sites. The tertiary care hospital was staffed using both a traditional teaching service model with residents and students and a solo hospitalist-only service staffing model since 2002. The community hospital was staffed by a solo-hospitalist staffing model since 2008. Continuous albuterol treatments were used in the PICU only at the community hospital site versus in both the PICU and ACS at the tertiary care hospital.

A retrospective analysis of hospitalizations for patients ages 4 to 17 years old with primary discharge diagnosis of asthma with principal International Classification of Diseases-9th and 10th edition category codes of 493.×, J44.×, and J45.×, was done between January 1, 2009 to February 28, 2020 at 2 hospitals – a community hospital and a tertiary care hospital. Pre-CPG implementation data were collected from January 2009 to November 2012. Data were collected for the 2 hospitals from the Pediatric Health Information System (PHIS, Children’s Hospital Association, Lenexa, KS), an administrative and billing database for children’s hospitals and their associated community hospital site.

From 2002 to 2009 the institution used an asthma clinical score based on the Pediatric Asthma Severity Score (PASS) system (Fig 1A).¹⁶  In June 2009, computerized physician order entry was introduced, and in the following month electronic documentation was introduced. An asthma CPG specific order was introduced to allow RT-driven weaning of nebulized albuterol in July 2009 with a plan for transition to albuterol MDI treatments once the patient was spaced to every 3-hours of nebulized albuterol treatments (Fig 1B).

In 2012, asthma CPG were developed through a multidisciplinary collaboration between nurses and physicians from pediatric hospital medicine, PICU, pulmonology, allergy and immunology, emergency department, community physicians, RTs, and additional ancillary staff. The CPG included early albuterol MDI initiation through a RT-driven BD weaning protocol that allowed RTs to wean per a mutually agreeable evidence and consensus-based protocol using asthma score (Fig 1, C and D). MDI treatments were started once the patient was appropriate for spacing to every 1-hour treatment based on score and evaluation. This was a substantial change compared with the 2009 practice. The CPG also clearly defined recommended discharge criteria of spacing to every 4 hours albuterol treatments. In pre-CPG implementation there was variability in spacing of albuterol treatments to a range of 4 to 6 hours before discharge. The CPG were introduced in both the tertiary care hospital and the community hospital at the same time and in a similar manner. Although no changes were made to the BD protocol or asthma clinical practice guidelines in 2018, additional education was distributed to medical staff regarding the asthma clinical practice guidelines at the community hospital as part of the implementation of the Pathways for Improving Inpatient Pediatric Asthma Care¹⁴  project in January 2018.

A new BD protocol order set was developed for the EHR with a prechecked asthma CPG order set that incorporated RT-driven BD weaning guidelines using the Asthma score rather than PASS score (Fig 1, C and D). The previous nebulized albuterol treatments were replaced with albuterol MDI treatments. The new order set went live in March 2013 (Fig 1B).

Principal International Classification of Disease diagnosis codes from the 9th and 10th editions included were 493.×, J44.×, and J45.×. Exclusion criteria included direct admission after transfer from an outside hospital, PICU stay, previous asthma admission within the past 30 days, complex chronic condition,¹⁷  mortality, or the presence of specific secondary diagnosis code (Fig 2). Secondary diagnosis codes that would alter the standard course for an asthma patient, such as pneumonia, pneumothorax, surgical complications, paralysis, bronchiectasis, suicide attempt, cystic fibrosis, sickle cell anemia, or poisoning were also excluded (Supplemental Fig 4).

Patient demographics, including age (4–9 years, 10–14 years, 15–17 years), sex, race, payor, disposition, median household income, and child opportunity index were collected. Child opportunity index is a measure of relative opportunity between neighborhoods in the same metropolitan area determined by the quality of the resources and conditions that are important for child development.¹⁸ 

The primary outcome measure was LOS in hours. LOS was calculated from the time of admission to the facility to ACS discharge time. Other outcome measures were ED revisit within 7 and 14 days, and 7- and 14-day readmissions to the hospital.

All statistical analyses were stratified by the hospital site. Categorical variables were summarized with frequencies and percentages, whereas continuous variables were summarized with geometric means and standard deviations. Comparisons between pre-CPG LOS, ED revisit, and hospital readmission at 7 and 14 days and post-CPG LOS, ED revisit and hospital readmission at 7 and 14 days were made using χ² or Wilcoxon Rank-Sum tests as appropriate. Segmented regression analyses with interrupted time series (ITS) were used to assess the impact of the CPG on quarterly outcomes while adjusting for age, race, and severity using the H-RISK¹⁹  score. The ITS provided rates of change (ie, slopes) in each measure for the pre- and post-CPG periods and assessed if there was a significant change in slopes. The ITS was also used to assess if a level change at the time of the CPG’s introduction occurred.²⁰  Quality control charts were used to assess changes in the variability (ie, SD) of LOS after the introduction of the CPG. We used standard s-charts with Western Electric Rules to detect significant changes in the post-CPG period from the pre-CPG period. All statistical analyses were performed using SAS v.9.4 (SAS Institute, Cary, NC) and P values <.05 were considered statistically significant. Institutional Review Board exempt status was approved.

There were 9699 hospitalizations for asthma that met the inclusion criteria over the 11-year study period with about 80% of the children being hospitalized at the tertiary care hospital (Table 1). Compared with the tertiary hospital, the community hospital had a higher percentage of private insurance, higher median household income, and higher child opportunity index.

A trend toward decrease in LOS was noted at both hospitals during the time period of the study (Table 2). The LOS was lower by 6.6 hours (P < .001) at the tertiary care hospital and 6.7 hours (P < .001) at the community hospital during the time period after CPG implementation. There was no significant difference in 7- and 14- day readmission rate or 7- and 14- day ED revisits post-CPG implementation at both hospitals.

Outcome measures were further analyzed with interrupted time series to account for secular trends over time (Fig 3). The LOS was reported as a geometric mean to reduce the influence of outliers at the tertiary care hospital (Fig 3A) and the community hospital (Fig 3C). The LOS showed a statistically significant downward trend at both hospitals before the official CPG implementation in November 2012 as evidenced by the negative slope of −0.952 (P < .001) at the tertiary care hospital (Fig 3A) and the negative slope of –1.178 (P < .001) at the community hospital (Fig 3C). After official CPG implementation, the slope of the curves flattened in the post model to a slope of 0.047 at the tertiary care hospital and a slope of −0.042 at the community hospital, indicating no further significant decline (Fig 3, A and C). However, the statistical process control chart demonstrated that variation in the LOS narrowed and showed improvement after the implementation of the protocol in 2012 (Fig 3, B and D).

The reduction in LOS at each site and the decreased variation in LOS at each site was sustainable for 7 years (30 quarters) after CPG implementation (Fig 3). A statistically significant decrease in variability of LOS compared with the preperiod (before November 2012) according to the Western Electric rules was noted for March 2014 to December 2015 (quarters 21–28, Fig 3B) at the Tertiary Care Hospital. A significant decrease in variability of LOS was also noted for June 2018 to February 2020 (quarters 38–45, Fig 3D) at the community hospital.

Implementation of asthma CPG with a revised RT-driven BD weaning protocol and EHR order entry use at both tertiary care and community hospital setting was associated with a reduction in the variation in LOS and maintenance of a decreased length of stay for 7 years after CPG implementation without significant changes in 7 and 14-day readmission rate or ED revisits. These findings have positive implications for patients, families, clinicians, and hospital administrators.

Long term adherence to CPG is challenging in a busy ACS. We believe that the key driver for sustained reduction in variation of LOS is streamlining of practice. A multidisciplinary collaborative was used to develop asthma CPG and order sets. CPG were created through consensus with all key stakeholders involved in the development, which led to common goals, mutual trust, and respect. Team building and engagement led to ownership of management between all stakeholders. Data sharing and celebrating early success improved team performance. Moreover, institutional support for using CPG metrics as financial incentives for RTs and physicians encouraged further CPG compliance.

Variation in LOS was also reduced by shifting autonomy through RT-driven weaning protocol and an order-set. Asthma exacerbations are a high-volume condition that can be managed using objective score-based assessments.⁷  Delay in physician assessment in a busy inpatient care setting can negatively impact LOS. Collaborating with RTs likely led to improved assessment time by permitting RT autonomy. The transition from using the PASS score (Fig 1A) to the Asthma score (Fig 1C) allowed for the removal of some of the subjectivity of the scoring system. Through physician and RT collaboration, we developed an environment of mutual trust and teamwork. Incorporating the CPG in an order set at both sites provided both physician and RT with a framework for weaning MDI treatments. The community site benefited from the use of the order set and education resources designed at the tertiary care site. These required resources, including RT time, education, and several EHR builds to incorporate CPG recommendations, that were supported by the hospital.

Variation in LOS was further reduced by an early transition from nebulized treatments to MDIs and standardizing a 4-hour gap in albuterol treatments as an appropriate discharge criteria to reduce practice variation. This was a culture and practice change and was adopted as a long-term institutional plan through education of all stakeholders. This change allowed for early demonstration of MDI delivery and MDI teaching by RT for patients and families, which likely improved comfort with discharge. Organizational commitment to resources to support these changes was critical in sustaining long-term outcomes at both sites. Schefft et al²¹  similarly found that clear discharge criteria for asthma resulted in decreased variability in LOS, hypothesizing that increased communication between providers and nurses during development of discharge criteria improved the significant practice variation.

Secular trends in the ITS did not demonstrate significant CPG impact on LOS. There was a trend toward decreased LOS before CPG implementation with a sustained decreased LOS over the 7 years after CPG implementation. There was LOS reduction preintervention likely because of improved communication through EHR and institutional practice change before the study period in 2012 (Fig 2), including EHR introduction in 2009 and increased acceptance of RT led BD weaning first introduced in 2002. Currently, there are no national benchmarks for asthma LOS and achievable benchmarks of care for asthma have focused primarily on resource utilization.²²  Therefore, it is possible that current LOS might be within the range for asthma hospitalization for pediatric tertiary and community site and might not be reduced any further.

When compared with the community site, the mean LOS at the tertiary site was longer. We believe that this was likely because of the management of patients requiring continuous albuterol treatments on ACS at the tertiary hospital. Children requiring continuous albuterol therapy were managed in the PICU at the community site and were excluded. Other factors that may have contributed to the lower LOS at the community hospital included the higher average child opportunity index, increased median household income, and higher percentage with private insurance as compared with the tertiary hospital. Kenyon et al demonstrated a higher risk of asthma rehospitalization for children with public insurance, compared with private.²³ 

Sustaining CPG implementation long term is important for patients, families, clinicians, and administrators. A systematic review of 4219 CPG implementations found that mean time between end of active implementation and sustainability of a CPG implementation was 2.6 years and that adherence to CPG is reduced to 50% by 1 year.²²  Similarly there was a significant decline in CPG adherence after 1 year for the PIPA national pediatric asthma QI collaborative.²⁴  Nkoy et al demonstrated that a tertiary care hospital was able to sustain 90% of asthma outcomes for 5 years and related community hospitals maintained 80% to 90% of outcomes for >2 years.²⁵  To our knowledge, no previous studies have shown reduction in variation of LOS over a long-term period. Sustaining CPG long term is important as it implies streamlining of evidence-based practice, thus patients get more consistent care, are less likely to stay in the hospital, may improve quality of life for patients and their families. Decreased variability in the asthma LOS implies decreased variability in the care that the patient received. Decreased variability and maintenance of lower LOS for hospitalized asthma patients over 7 years has a positive impact on school attendance, parental work absenteeism, hospital capacity, and cost of care.

Future asthma CPG implementations may evaluate additional factors. For example, strict oxygen saturations requirement may be able to be removed from the asthma score, and discharge of patients could be considered before spacing the patient to every 4-hour albuterol treatments. Lo et al found that changing the discharge criteria to every 3 hours albuterol treatments rather than every 4 hours did not increase the 14-day readmission rate.²⁶ 

Our study has several limitations. Our data represents only 1 hospital system and both hospitals in this study are free-standing children’s hospitals. PHIS is an administrative database of 43 pediatric hospitals and study findings might not include the entire breadth of hospitals caring for children. The study did not identify baseline asthma severity or acuity of initial asthma exacerbation. The community site opened in 2009, limiting the data that could be collected before the confounding factors of introduction of computerized order entry in June 2009 and introduction of electronic documentation in July 2009. We were unable to measure impact of reduction in LOS variability on school attendance, parental leave, hospital capacity and cost of care.

Asthma CPG implementation led to a significant reduction in the variability of LOS at both a tertiary care and the community pediatric hospital without change in 7-day and 14-day readmission and ED visit rate. The trend of decrease in LOS that started before CPG implementation was sustained for 7 years after CPG implementation. Sustaining decreased asthma LOS long-term can improve parent and child quality of life and have hospital capacity benefits. Thus, those implementing asthma CPG should consider using EHR for order sets, early MDI use, and collaborating with RTs to develop BD weaning protocols embedded within the EHR.