Objective: To examine the frequency of hospital admissions before and after gastrostomy insertion in children with severe intellectual disability. Study design: We conducted a retrospective cohort study using linked health administrative and disability data from Western Australia (WA) and New South Wales (NSW). Children born between 1983 and 2009 in WA and 2002 and 2010 in NSW who had a gastrostomy insertion performed (n = 673 [WA, n = 325; NSW, n = 348]) by the end of 2014 (WA) and 2015 (NSW) were included. Conditional Poisson regression models were used to evaluate the age-adjusted effect of gastrostomy insertion on acute hospitalizations for all-cause, acute lower respiratory tract infections (LRTI), and epilepsy admissions. Results: The incidence of all-cause hospitalizations declined at 5 years after procedure (WA cohort 1983-2009: incidence rate ratio, 0.70 [95% CI, 0.60-0.80]; WA and NSW cohort 2002-2010: incidence rate ratio, 0.63 [95% CI, 0.45-0.86]). Admissions for acute LRTI increased in the WA cohort and remained similar in the combined cohort. Admissions for epilepsy decreased 4 years after gastrostomy in the WA cohort and were generally lower in the combined cohort. Fundoplication seemed to decrease the relative incidence of acute LRTI admissions in the combined cohort. Conclusions: Gastrostomy was associated with health benefits including reduced all-cause and epilepsy hospitalizations, but was not protective against acute LRTI. These decreases in hospitalizations may reflect improved delivery of nutrition and medications.