We thank Dr. Pandit for his letter. It is interesting that Medicare anesthesia claims, chart abstraction in the United States,1and computerized records from operating rooms in the United Kingdom all produce similar procedure durations, consistent with the validity of all three measures.

Dr. Pandit observes that procedure times at his center are similar to the median times for 183 hospitals in Pennsylvania—this one United Kingdom center would fit near the center of the distribution in Pennsylvania. Although this observation is interesting, it is not logically inconsistent with variation in procedure time between hospitals in both Pennsylvania and the United Kingdom. To see variation between hospitals, one must study several hospitals.

Dr. Pandit asks whether differences between Pennsylvania hospital procedure times are different for different procedures. Indeed they are. We looked at the five most common procedures in our data base (total knee replacement, open reduction of fractured femur with internal fixation, laparoscopic cholecystectomy, total hip replacement, and partial hip replacement) for the 10 largest hospitals in Pennsylvania, adding interaction effects to the model in our article.2Of the 5 × 10 = 50 possible interactions, 15 were significant using the Bonferroni correction (P < 0.05/50 = 0.001), so that at least for common procedures at large hospitals, procedure duration differs in different ways at different hospitals. This is consistent with the notion that procedure duration at a hospital is a function of the individual physicians, nurses, and care team members practicing within hospitals and possibly performing different procedures with different skills. For example, future research exploring physician characteristics and style of practice inside and across hospitals may help to explain the differences in procedure times we observed by the race and income of the Medicare patient.2 

Using our algorithm, Medicare claims can now provide investigators with an opportunity to study a vast number of questions involving patients, surgeons, anesthesiologists, hospitals, and even health systems. Although procedure times have generally been reported from single-institution studies, having times now available from literally 40 million procedures yearly, dating back to 1994 and extending to the present, should allow many new investigations that will greatly benefit our understanding of surgical care and healthcare delivery.

*The University of Pennsylvania School of Medicine and The Wharton School, and The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania. silber@email.chop.edu†The Wharton School, The University of Pennsylvania, Philadelphia, Pennsylvania.

Silber JH, Rosenbaum PR, Zhang X, Even-Shoshan O: Estimating anesthesia and surgical times from Medicare anesthesia claims. Anesthesiology 2007; 106:346–55
Silber JH, Rosenbaum PR, Zhang X, Even-Shoshan O: Influence of patient and hospital characteristics on anesthesia time in Medicare patients undergoing general and orthopedic surgery. Anesthesiology 2007; 106:356–64