Refining the identification of a precise cut point for the association between annual surgeon total thyroidectomy volume and complications

Presentation: AHNS-006
Topic: Endocrine
Type: Oral
Date: Wednesday, May 1, 2019
Session: 1:10 PM - 1:55 PM Scientific Session 1 - Endocrine
Authors: Charles Metzer, MD1, Michaela Hull, MS2, John L Adams, PhD2
Institution(s): 1The Permanente Medical Group, 2Kaiser Permanente, Center for Effectiveness and Safety Research

Background: Higher surgeon volume for thyroid procedures is associated with improved clinical outcomes. Published definitions of high-volume surgeons vary widely. Previous estimates using regression modeling and generalized additive models (GAMs) may be inaccurate due to the clustering of cases within surgeons. In addition, little is known about the annual procedure volumes at which hypocalcemia, vocal cord paralysis (VCP), and hematoma start to decrease. The objective of this study was to control for physician-level effects and identify a precise cut point at which surgeon annual total thyroidectomy (TT) volume was associated with decreased 30-day complication rates for hypocalcemia, vocal cord paralysis, hematoma, and a composite measure that also included selected general complications of surgery.

Methods: We studied 10,546 TT procedures performed by 338 surgeons in the Northern and Southern California regions of Kaiser Permanente in 2008-2015. We used generalized additive mixed models (GAMMs), an extension of GAMs that allows for smoothing of volume-outcomes curves with splines and including random (physician-level) effects. Outcome measures were hypocalcemia, VCP, hematoma, and a composite outcome that included these complications and acute myocardial infarction, AMI, chyle fistula, neck swelling, surgical site infection, seroma, stridor, and UTI. Modeling was adjusted for gender, pregnancy, health care use in the year before surgery, region, surgery year, and selected comorbidities (thyroid cancer, dyspnea, cardiovascular disease, DxCG risk score, Charlson comorbidity index, and history of acute myocardial infarction, clotting disorder, dialysis, seizures, or stroke).

Results: The overall rate of the composite outcome was 13.4%, which began to decrease at 17.0 (95% CI, 10.5 - 23.6) TTs per year. 281 surgeons performing ≤ 16 TTs per year completed 5501 (52%) procedures with a 15.3%. complication rate. 61 surgeons performing ≥ 17 TTs per year completed 5045 (48%) procedures with a complication rate of 11.8%. Of 1438 complications, 841 (58.5%) followed TTs by lower volume surgeons. The overall rate of hypocalcemia was 6.0%, and it began to decrease at 17.8 (95% CI, 13.5 -22.8) TTs per year. 282 surgeons performing ≤ 17 TTs per year completed 5722 (54%) procedures with a hypocalcemia rate of 7.3%. 57 surgeons performing ≥ 18 TTs per year completed 4824 surgeries with a hypocalcemia rate of 4.4%. Of 632 instances of postoperative hypocalcemia, 418 (68.4%) followed TTs by lower volume surgeons. Rates of VCP and of hematoma both decreased at approximately 20 TTs per year but neither reached statistical significance.

Conclusions: Rates of a composite measure of 30-day complications began to decrease to a statistically significant degree when surgeons performed ≥ 17.0 TTs per year; 30-day hypocalcemia rates began to decrease when surgeons performed ≥ 17.8 TTs per year. No similar associations were observed for vocal cord paralysis and hematoma. To improve quality and patient safety, total thyroidectomies should be directed to surgeons who perform at least 17 of these procedures annually.