INTRODUCTION: The timing of surgical intervention in the management of spontaneous pneumothorax remains controversial. The aim of this multicenter review was to compare management strategies and outcomes in children with spontaneous pneumothorax. METHODS: We retrospectively reviewed patients 10-19 years old in the Pediatric Health Information System admitted for spontaneous pneumothorax from 2010 to 2014. Three treatment groups were identified based on initial hospital management-no intervention, initial chest tube placement, and operation; and outcomes were compared. RESULTS: A total of 1040 patients were included. The majority were male (82.1%) and White (71.1%). The mean age at first encounter was 15.7 +/- 1.7 years. Initial treatment included no intervention in 336 (32.3%), chest tube in 497 (47.8%), and video-assisted thoracoscopic surgery (VATS) in 207 (19.9%). Ultimately, 417 (40.1%) patients underwent VATS during the initial admission and 559 (53.8%) during the initial admission or a subsequent encounter. Aggregate length of stay (LOS) was highest for those treated initially with chest tube alone (P < .001). For patients managed initially with chest tube, the probability of requiring surgery increased with each day of hospitalization. Initial operation was associated with a decreased risk of readmission (OR 0.67, 95% CI 0.50-0.90). Estimated adjusted hospital costs, aggregated across all encounters, were highest for chest tube alone (P < .001). CONCLUSION: Early VATS is associated with decreased hospital LOS, charges, and readmissions. For those managed initially with chest tube alone, the likelihood of requiring operation increases with each day hospitalized, and early conversion to operative management should be considered in patients with persistent pneumothorax or air leak.