Comments on “Head and Neck Reconstruction: Does Surgical Specialty Affect Complication Rates?”Funding None.
12 September 2019
29 September 2019
13 November 2019 (online)
We appreciate and acknowledge the concerns voiced by Bur et al regarding our recently published article “Head and Neck Reconstruction: Does Surgical Specialty Affect Complication Rates?” First and foremost, we want to stress that patient safety is paramount in all that we do and that the aim of this publication was to determine if it was possible that the specialty affected outcomes for head and neck reconstruction in the American College of Surgeons National Surgical Quality Improvement Program (American College of Surgeons (ACS)-NSQIP) database. The reason for this is to possibly identify differences that could be resolved with increased collaboration of clinical teams and basic science researchers.
Our hypothesis was that surgical specialty would not affect outcomes. We are of the opinion that it is the execution of a procedure that matters, not an individual's specialty that garners excellent and safe delivery of care. We believe that there are exceptional microsurgeons of many diverse backgrounds (orthopedics, urology, otolaryngology, and plastic surgery) all performing innovative surgery and that if we want to continue to improve the care delivered to our patients, we need to collaborate more and share ideas readily.
We agree with the concerns by Bur et al that the “surgeon specialty (SURG-SPEC)” variable contained in NSQIP only records the surgeon of record for a primary procedure as reflected in NSQIP. However, there are misconceptions regarding cases including multiple surgeons that lead to nonuniform reporting of this variable. If additional surgical timeouts are performed for each surgeon, then the case will be recorded as a separate primary procedure. However, if cases are concurrent and no additional timeout is performed, then only the surgeon, performing the designated primary procedure, will have his/her specialty recorded.
We conducted this study with sound methodology. We did include current procedural terminology (CPT) code 15757 in the supplemental table as the first code of included CPT codes, included in the study. To exclude cases of free-tissue transfer outside of the head and neck, we identified all head and neck cancer cases by using both the International Classification of Disease 9th and 10th (ICD-9 and ICD-10) codes to identify cases. After this, we identified cases where a free tissue transfer was performed. Concurrent cases report only one of the involved surgeons as the surgeon specialty for a case involving two or more surgical teams, thus we excluded them to prevent errors in attributing a free flap to a given specialty.
We choose not to study the site of the defect reconstructed to enable the largest possible sample size. Furthermore, comparing the defects based upon wound classification is meaningless, as all defects involving the upper and lower aerodigestive tract and orbit are not sterile body cavities but rather clean-contaminated operative fields.
We too agree that large databases are powerful tools that must be used responsibly to generate meaningful research. With the help of our data scientist, we conducted a sound statistical analysis of the data. The post hoc corrections are controversial amongst the statistical community. To simplify reporting, all variables with a p < 0.05 had small p-values that were under 0.001 and if a Bonferroni's correction was performed, the corrected α would have fallen to 0.0488 resulting no changes to our conclusions. It should be noted, that our data clearly reflected a trend to a statistically significant higher unplanned return to the operating room in cases completed by otolaryngology and a nonsignificant trend to higher readmission rates for patients treated by otolaryngologists. It is incumbent upon medical researchers to understand statistical analysis and properly design an analysis because as many as 78% surgical publications contain significant statistical errors. 
When examining the differences in patients' baseline characteristics, we noted differences that were significant. However, after we examined the distribution of preoperative functional class, and American Society of Anesthesiology class (ASA) scores, the groups were found to be distributed similarly with no statistically significant difference between the specialties leading us to conclude that these differences were not clinically applicable or significant. If these preoperative characteristics were indeed clinically relevant and one group of surgeons was indeed operating on “sicker” patients, it would be reflected in the average ASA score or functional status of patients operated on.
We thank Bur et al for taking the time to reply to our publication and agree that greater collaboration in head and neck reconstruction is needed. The aim of this publication is to understand if there is a difference in outcomes by specialty so that changes can be made thereby improving the quality of care received by patients. This is not meant to determine which specialty is the best at performing head and neck reconstruction. Increased collaboration will allow for development of standard defect classifications, reconstructive techniques, and improved execution of head and neck reconstruction independent of the operating surgeons training pathway. In the end, the only thing that matters is safe and effective execution of the reconstructive endeavor.
- 1 Drinane JJ, Drinane J, Nair L, Patel A. Head and neck reconstruction: does surgical specialty affect complication rates?. J Reconstr Microsurg 2019; 35 (07) 516-521
- 2 American College of Surgeons-National Surgical Quality Improvement Program. User guide for the 2017 American College of Surgeons National Surgical Quality Improvement Program participant use data file. Available at: https://www.facs.org/-/media/files/quality-programs/nsqip/nsqip_puf_userguide_2017.ashx . Accessed October 23, 2019
- 3 Levine MR, Allen R. , eds. Manual of Oculoplastic Surgery. Switzerland: Springer; 2018
- 4 Perneger TV. What's wrong with Bonferroni adjustments. BMJ 1998; 316 (7139): 1236-1238
- 5 Thiese MS, Ronna B, Robbins RB. Misuse of statistics in surgical literature. J Thorac Dis 2016; 8 (08) E726-E730