To empirically evaluate the impact of physician volume on medical expenditures with very low birth weight infants who were initally haspitalized. Method: NHI claim record data from 1191 cases of very low birth weight (500g-1499g) infants were obtained from the NHI research database. The regression model was employed to evaluate the relationship between physician volume and medical expenditures. Results: After adjusting for hospital characteristics, patient characteristics and outcome, we found a negative correlation between physician volume and medical expenditures. Conclusion: Medical expenditures may be lower for very low birth weight infant when cared for by a high volume physician, but the effect was limited.