Field substitution of nonresponders can maintain sample size and structure without altering survey estimates-the experience of the Italian behavioral risk factors surveillance system (PASSI).