The impact of adenotonsillectomy on healthcare utilization in children with obstructive sleep apnea: Experience of an integrated medical model.

ABSTRACT

BACKGROUND AND OBJECTIVES: Studies suggest that children with obstructive sleep apnea (OSA) have increased healthcare utilization patterns in comparison to matched controls. However, the effect of adenotonsillectomy (AT) on utilization patterns in these children is poorly understood. Additionally, no previous studies have compared the effect of AT on healthcare utilization patterns across different OSA severity groups. The aim of this retrospective cohort study is to assess the effects of surgical treatment on the level of healthcare utilization among children with OSA at a large integrated multicenter healthcare system. METHODS: Retrospective analysis was performed of children aged 3-12 diagnosed with OSA via an attended polysomnogram (PSG) between December 2016 and February 2019. Demographic variables including age (at time of PSG), body mass index (BMI), race, and ethnicity were obtained. Variables for healthcare utilization were assessed for 12 months prior to PSG, and for 12 months after PSG (or after AT, delayed for the first 30 days after surgery to account for surgery-related visits). Healthcare utilization variables assessed included the total number of outpatient visits, inpatient, and emergency department (ED) visits, visits involving diagnostic codes associated with upper respiratory infection (URI), otitis media (OM), and allergic rhinitis (AR), prescription data involving intranasal steroids or leukotriene receptor antagonists (LTRA), and communication data such as secure message load and specialty referrals. Repeated measure linear difference-in-difference (D-I-D) models were used to assess the causal impact of AT on healthcare utilization outcomes. Sensitivity analyses were performed using modeling with a Poisson distribution and as an unadjusted model, with statistical significance set to p < 0.05. RESULTS: Analysis elicited 577 children identified with OSA. Of these, 336 (58.2 %) underwent observation while 241 (41.8 %) underwent AT. The mean age was 6.4 years, with a slight male predominance (60.5 %). Analysis of baseline healthcare utilization patterns revealed that the treatment group had a significantly higher number of baseline inpatient/ED visits and OM visits in comparison to the observation group, but no differences in regards to baseline outpatient visits, or in visits involving URI or AR. Analysis of the entire OSA cohort via D-I-D modelling showed a significantly larger reduction in outpatient visits, secure messages, specialty referrals, and the use of intranasal steroid and LTRA in the treatment group compared to the observation group. Stratification of children based on OSA severity showed that the significant differences in healthcare utilization attributed to surgical treatment were primarily driven by the severe OSA group. Children with severe OSA who underwent AT showed significant reductions in most variables including outpatient visits, inpatient/ED visits, and OM visits. Alternatively, the only significant reductions in healthcare utilization among children with mild OSA treated with AT were in AR visits, intranasal steroid use, and LTRA use. Pattern changes among children with moderate OSA compared similarly to those with mild OSA. CONCLUSION: To the authors' knowledge this study represents the largest available study assessing the impact of AT on healthcare utilization in children with OSA that also considers the effect of OSA severity on utilization patterns. AT appears to decrease healthcare utilization patterns, particularly in children with severe OSA. Alternatively, children with mild or moderate OSA treated with AT had only modest reductions in healthcare utilization patterns.