S. C. Pitt1, M. A. Nehs2, N. L. Cho2, D. T. Ruan2, F. D. Moore2, A. A. Gawande2 1University Of Wisconsin,Department Of Surgery,Madison, WI, USA 2Brigham And Women’s Hospital,Department Of Surgery,Boston, MA, USA
Introduction: The impact of resident involvement on surgical outcomes is controversial since much of the data are contradictory. Furthermore, the role of resident participation specifically in thyroid surgery is not well understood. Therefore, we sought to determine whether teaching hospitals (THs) had disparate outcomes in a large, population level cohort of thyroidectomy patients.
Methods: We analyzed 9,798 patients who underwent thyroidectomy in the 2011 California State Inpatient (SID) and Ambulatory Surgery Databases (SASD). To assess TH status, the SID and SASD were linked to the California Hospitals Utilization data. Surgical procedures and outcome variables were defined by CPT and ICD-9 codes. Outcomes were analyzed based on TH status using ANOVA, as well as univariate and multivariate logistic regression methods.
Results: The minority of the 9,798 thyroidectomies were performed at THs (20.4%). Those patients treated at THs were similar to those at non-THs, but more likely to have thyroid cancer (42.1% vs. 31.7%), a total thyroidectomy (53.0% vs. 46.7%) or a lateral neck dissection (11.8% vs. 5.0%, p<0.0001 for all). Despite these differences, the overall complication and 30-day readmission rates were similar at TH and non-THs (9.1% vs. 8.9%, p=0.76, and 3.6% vs. 3.8%, p=0.65, respectively). However, when analyzing specific complications, recurrent laryngeal nerve (RLN) injuries were significantly more common at THs (1.9% vs. 0.9%, p<0.0001), while hypoparathyroidism (1.0% vs. 0.7%, p=0.20) and reoperative hematoma (0.7% vs. 0.8%, p=0.55) were comparable. In addition, the volume status of THs did not affect the RLN injury rate when examined as low (<50 cases/yr), medium (50-130 cases/yr), and high (>130 cases/yr) volume hospitals (p=0.15).
Assessment of factors, other than TH status, that were associated with RLN injuries revealed that nerve injuries patients were similar in gender, but significantly (p<0.001 for all) older (61.0 ± 16.4 vs. 53.3 ± 14.7 yrs), non-White (0.8% vs. 2.5%), not privately insured (38.4% vs. 62.7%), more likely to have >1 chronic medical condition (99.0% vs. 74.5%), thyroid cancer (6.7% vs. 1.7%), hypothyroidism (2.5% vs. 1.0%), and undergo total thyroidectomy (1.6% vs. 0.7%), or lateral neck dissection (4.8% vs. 0.9%). Multiple logistic regression revealed that TH status was independently associated with RLN injury (OR 2.43, 95% CI 1.54-3.83) when controlling for age, race, insurance type, comorbidities, thyroid cancer diagnosis, and procedure type.
Conclusion: While overall complication rates are not impacted when thyroidectomy is performed at a TH versus a non-TH, RLN injury rates are twice as high at TH. Whether this difference is due to resident involvement or is a reflection of referral bias and case complexity deserves further investigation.