Background and Significance: Croup affects more than 1.4 million children under 6 years of age each year in the United States. It is the most common cause of airway obstruction in children, and, in severe cases, leads to respiratory failure and arrest. Overall, 1.5-6% of patients with croup are hospitalized at an estimated cost of $56 million annually in the US. Marked variation exists across free-standing children’s hospitals in the use of tests and treatments not routinely indicated in the care of croup such as chest and neck radiographs and viral testing. The hospital of admission rather than average patient case mix predicts this outstanding variation. Systemic corticosteroids are the mainstay of treatment in croup and there is strong evidence to support a single dose of dexamethasone in all patients with croup. However, these data are limited in scope as the research on corticosteroids in croup to date has focused on outpatient and emergency room management of mild to moderate croup and compared a single dose of dexamethasone to placebo to prevent outcomes such as return visits and hospital admission. However, hospitalized patients represent a distinct patient population often with associated morbidity with 9% of admitted patients requiring intensive care services. Surprisingly, despite the large number of children hospitalized for croup and the significant burden on health care resources associated with this common childhood illness there are essentially no data on the optimal management of inpatients with croup. Moreover, no studies compare the impact of single dose versus multiple dose dexamethasone therapy in patients with croup. These gaps in knowledge are critical to address in order to understand the optimal treatment for croup. Addressing these gaps forms the basis of my long term research goal which to decrease the frequency of poor outcomes (e.g. intubation, readmission and death) and reduce unnecessary health care utilization among children hospitalized with croup. The overall objective of this application is to examine the comparative effectiveness of one versus multiple doses of dexamethasone in otherwise healthy children hospitalized with croup and to understand how provider medical decision making about testing and treatment in croup influences care. Our central hypothesis is that inpatients with croup may benefit from multiple doses of dexamethasone compared to a single dose, as is currently recommended. The data derived from AIM 1 will begin to elucidate how, after matching on patient severity within hospital, the number of dexamethasone doses relates to patient outcomes. Concurrently, AIM 2 will identify factors that influence provider medical decision making to inform future implementation studies. The following Specific Aims are proposed.
Study Aims: Aim 1) Determine the comparative effectiveness of multiple versus single dose dexamethasone within 36 hours to reduce undesired outcomes among generally healthy children admitted with croup. In this retrospective study of data from a national, multi-institutional dataset called PEDSnet, we will employ propensity score methods (full matching and weighting) within hospital to control for confounding by severity to detect differences in outcomes among hospitalized patients treated with < 2 and >2 doses of dexamethasone within 36 hours of initial presentation to medical care (inclusive of both in and outpatient settings). We will compare frequency of undesired outcomes (such as escalation of care to the ICU, intubation, return to ED after discharge, need for additional steroids after discharge, readmission, and frequency of measurable dexamethasone adverse side-effects) between groups. H1: When exposure groups are propensity-matched (or -weighted) within hospital on covariates, patients who are eventually hospitalized and receive >2 doses of steroids within 36 hours of initial presentation to medical care will have better outcomes compared to those who only receive 1 dose. Aim 2) Understand clinician perspectives and factors influencing clinical decision making around the care of patients hospitalized with croup. In this prospective qualitative study, we will conduct semi-structured interviews with individual emergency medicine and hospitalist clinicians to identify factors guiding work-up and treatment decisions.
Study Design: This will be a retrospective analysis of data from a national, multi-institutional dataset called PEDSnet which includes outpatient and inpatient data for over 5 million children from 8 large children’s hospitals. The analysis will employ propensity score full (optimal) matching methods within hospital to detect differences in outcomes among hospitalized patients treated with < 2 and >2 doses of dexamethasone within 36 hours of initial presentation to medical care (inclusive of both in and outpatient settings)34,35. Matching within hospital will be applied due to the expected high variation between hospitals in the treatment of croup. The propensity score of each patient will be defined as the probability to receive < 2 (versus > 2) doses of corticosteroids conditional on pre-treatment covariates. This will allow comparison of the frequency of undesired outcomes (such as escalation of care to the ICU, intubation, return to ED after discharge, need for additional steroids after discharge, readmission, and adverse side-effects of dexamethasone) between groups.