The plot thickens... It seem that "my" Hermite polynomials are the physicist's version and have exp(-x^2) in the definition so the procedure wouldn't work that trivially. But nevertheless, I managed to evaluate the integral using recurrence relations. Thanks to the both of you!