349 - Identifying Disparities in Bronchiolitis Quality Metrics

Publication Number: 4342.349

Theresa McAllister, Medical College of Wisconsin, Wauwatosa, WI, United States; Jinger Haynes, Medical College of Wisconsin, Milwaukee, WI, United States; Kaitlyn M. McQuistion, Medical College of Wisconsin, Oconomowoc, WI, United States

Background: With growing efforts to integrate quality improvement (QI) and health equity, stratifying quality metrics by patient characteristics to identify health disparities has been recognized as a critical early step in the design and implementation of equitable interventions. This QI project leveraged the work of an established interdisciplinary bronchiolitis QI committee to perform a stratified analysis and identify targets for ongoing improvement.

Objective: To identify any baseline disparities in bronchiolitis quality metrics (albuterol use, length of stay (LOS), 7-day all-cause readmission) based on sociodemographic variables (race/ethnicity, language for care, payer, Child Opportunity Index (COI)).

Design/Methods: All patients 0-2 years old without significant pulmonary comorbidity admitted to our free-standing children's hospital with bronchiolitis between 1/2016-6/2025 were included. Statistical process control (SPC) charts for each quality metric were stratified by sociodemographic variables using rational subgrouping. Multivariable linear and logistic regression models were then used to estimate associations between the sociodemographic variables and each quality metric, adjusting for clinical characteristics.

Results: The analysis included 7,156 patients. Overall albuterol use decreased from 29% to 19.6%, with reductions observed across all sociodemographic groups (Figure 1). Baseline disparities for Black and publicly insured patients persisted over time. Overall average LOS decreased from 68 to 42.5 hours and decreased across all sociodemographic groups (Figure 2). At baseline, LOS was longer for LOE patients; however, this disparity narrowed over time. 7-day readmission averaged 2.5% and trended up in the last year. Figure 3 summarizes the results of the regression analysis. The odds of albuterol use were higher among Black patients (OR: 1.3, 95% CI: 1.1, 1.6) and lower among those with private insurance (OR: 0.8, 95% CI: 0.7, 1.0). LOS was longer for LOE patients (β = +7.1, 95% CI: 1.7, 12.5) and shorter for patients who were self-pay (β = -10.1, 95% CI: -16.7, -3.5). The odds of 7-day readmission were lower among Black (OR: 0.6, 95% CI: 0.4, 1.0), LOE (OR: 0.4, 95% CI: 0.2, 1.0), and privately insured patients (OR: 0.6, 95% CI: 0.4, 0.9).

Conclusion(s): SPC charts with rational subgrouping and multivariable regression are both valuable tools to analyze quality metrics for sociodemographic disparities. Findings will guide our bronchiolitis QI team in the implementation of targeted interventions to promote equitable bronchiolitis care.

Figure 1. Decrease in Inpatient Albuterol Use between 2016-2025
Panel A - SPC p chart demonstrating the overall decrease in % of patients admitted with albuterol administered at some time during their stay in the context of the interventions listed. Panel B - SPC p chart with rational subgrouping by race/ethnicity demonstrating baseline disparity for black patients that was maintained despite improvement across all groups over time. Panel C - SPC p chart with rational subgrouping by language demonstrating similar decrease in albuterol use for both English and language other than English (LOE) groups. Panel D - SPC p chart with rational subgrouping by payer demonstrating baseline disparity for patients with public insurance that was maintained despite improvement across both groups over time.

Figure 2. Decrease in Hospital Length of Stay between 2016-2025
Panel A - SPC Xbar chart demonstrating the overall decrease in length of stay (LOS) in the context of the interventions listed. Panel B - SPC Xbar chart with rational subgrouping by race/ethnicity demonstrating similar decrease in LOS across groups. Panel C - SPC Xbar chart with rational subgrouping by language demonstrating baseline disparity with higher LOS for the language other than English (LOE) group, with that disparity decreasing over time as improvement was seen across all groups. Panel D - SPC Xbar chart with rational subgrouping by payer demonstrating improvement across both groups over time with a possible disparity emerging with higher LOS for patients with public insurance.

Figure 3. Adjusted Associations Between Sociodemographic Variables and Each Quality Metric
Results of the multivariable logistic and linear regressions for each quality metric are shown as forest plots in panels A-C. Each model simultaneously controlled for age, year of admission, use of high flow nasal cannula support, ICU level care, race/ethnicity, language for care, payer, and COI. Panel A - The odds of albuterol use were higher among black patients compared to white patients (OR: 1.33, 95% CI: 1.08, 1.63); and lower among those with private insurance compared to those with public insurance (OR: 0.82, 95% CI: 0.69, 0.98). Panel B - Length of stay was longer for LOE patients compared to English-speaking patients (β = +7.11, 95% CI: 1.67, 12.54), and shorter for patients who were self-pay compared to those with public insurance (β = -10.09, 95% CI: -16.65, -3.53). Panel C - The odds of 7-day readmission were lower among black (OR: 0.59, 95% CI: 0.35, 0.98), LOE (OR: 0.42, 95% CI: 0.18, 1.00), and privately insured patients (OR: 0.60, 95% CI: 0.41, 0.87).