Surgical specialty and preoperative medical consultation based on commercial health insurance claims

Abstract

BACKGROUND: Surgical patients are sometimes referred for preoperative evaluations by consultants in other medical specialties, although consultations are unnecessary for many patients, particularly for healthy patients undergoing low-risk surgeries. Surgical specialty has been shown to predict usage of preoperative consultations. However, evidence is generally limited regarding factors associated with preoperative consultations. This study evaluates surgical specialty and other predictors of preoperative consultations. METHODS: This retrospective cohort study analyzed surgery claims of 7400 privately insured patients in Washington, United States, from eight surgical specialties. We estimated log-Poisson generalized estimating equation models that regress whether a patient received a consultation on surgical specialty and covariates accounting for the data's hierarchical structure with patients nesting within surgeons, and surgeons nesting within provider organizations. Covariates include age, gender, Deyo comorbidity index, surgical risk, and geographic factors. RESULTS: Overall, 485 (6.6%) patients had a preoperative consultation. The incidence of preoperative consultation varied significantly by surgical specialty. Orthopedics, neurosurgery, and ophthalmology had 3.9 (95% CI 2.4, 6.5), 2.3 (95% CI 1.1, 4.5), and 2.3 (95% CI 1.1, 4.6) times greater adjusted likelihoods of preoperative consultation than general surgery, respectively. The adjusted likelihoods of consultation for gynecology, urology, otolaryngology, and vascular surgery were not statistically different from general surgery. The following covariates were associated with greater likelihood of preoperative consultation: greater age, higher surgical risk, having one or more comorbidities vs. none, and small rural towns vs. urban areas. More than 75% of all consultations were provided to patients with a Deyo comorbidity index of 0 or 1. Low surgical risk patients had 0.3 (95% CI 0.3, 0.5) times the likelihood of preoperative consultation of intermediate and high-risk patients overall. CONCLUSIONS: The likelihood of preoperative consultation varied fourfold (an absolute 9% points) across surgical specialties. Most consultations were provided to patients with low comorbidity and with low or intermediate surgical risk. To improve usage of preoperative consultations as an evidence-based practice, future research should determine how the health outcomes effects of preoperative consultations vary depending on comorbidity burden and surgical risk.