Data-Driven Surgical Referrals for Primary Hyperparathyroidism and Associated Surgical Outcomes: A Simulation Study.

Abstract

Background: To minimize complications after parathyroidectomy, it is critical to connect patients with facilities equipped to perform this specialized procedure. This study assessed the effect of "referring" patients to higher-performing facilities for parathyroidectomy.

Methods: A simulation study of adults who underwent parathyroidectomy for primary hyperparathyroidism was conducted using the Healthcare Cost and Utilization Project State Inpatient and Ambulatory Surgery and Services databases. Wilson score estimates were used to generate serious adverse event (SAE) rates for a training cohort to divide surgical facilities into quality quartiles. Using each facility's fixed effect from the regression model, this study simulated the difference in SAE and cost for each patient between treatment at their original facility and treatment at an alternative higher-quality facility based on a lower SAE rate. The set of potential alternative facilities was determined based on proximity or original facility health system affiliation.

Results: Of the 14,738 patients included in the proximity analysis 11,733 were randomized to the training group and 3005 to the testing cohort. The baseline characteristics and outcomes did not differ between the cohorts. Among the patients in the testing cohort, 314 were simulated to receive care at a higher-quality facility. The simulation predicted decreased SAE rates (2% vs. 3%; p < 0.001), with slightly increased total cost ($6391 vs. $6,351; p = 0.032). The results in the two simulations were similar.

Conclusions: Simulation indicated that data-driven parathyroidectomy referrals can reduce SAE and advance surgical health equity. Data-driven facility selection is one way to achieve better surgical outcomes.