Method: We conducted a retrospective cross-sectional study of all Medicaid eligible children in Colorado (CO), New York (NY), North Carolina (NC) and Washington State (WA) over a 3 year period from January 1, 2006 to December 31, 2008. Using ICD-9 codes we identified all admissions with a G-tube or a PEG tube procedure. We calculated an incidence density rate per 10,000 person-years with 95% confidence intervals, and also examined the frequency with which a gastric fundoplication procedure was combined with the gastrostomy.
Result: There were 1,983 gastrostomy procedures during the study period in the four states; 64% were G-tubes. Many children received the procedure early in life, with 35.8% of G-tubes and 16.2% of PEGs occurring under 3 months of age. For all gastrostomy procedures the annual incidence density rate per 10,000 person-years varied between states: CO: 2.7 (2.3, 3.0), NY: 3.4 (3.1, 3.6), NC 2.2 (2.0, 2.4) and WA: 4.5 (4.0, 5.0). For G-tube insertion the rates were: CO: 2.1 (1.8, 2.4), NY: 1.8 (1.6, 2.0), NC: 1.7 (1.5, 1.8), and WA: 2.5 (2.2, 2.9). PEG insertion rates were as follows: CO: 1.1 (0.8, 1.5), NY: 2.7 (2.4, 2.9), NC: 1.2 (1.0, 1.4), and WA: 3.1 (2.6, 3.6). State-to-state variation was most pronounced in the frequency with which gastric fundoplication was combined with a G-tube: CO: 60.2%, NY: 31.3%, NC: 49.6%, and WA: 36.1% (p<.0001).
Conclusion: The substantial variation in rates suggests that surgeon or provider preference may be driving decisions of whether to perform a gastrostomy feeding procedure, or the type of procedure chosen. The question of why certain states have higher rates is likely complex, but may relate to local medical opinion, training of pediatric providers and perhaps to the perceived efficacy of gastric fundoplication.