Surgeon Factors Rather Than Patient Factors Account for Variation in Recommended Treatment Strategy for Patients With Multiligament Knee Injury

OBJECTIVES: To seek the factors associated with timing, staging, and type of surgery in the management of multiligament knee injuries. METHODS: Design: Cross-sectional scenario-based experiment. Setting: Fifteen fictional patient scenarios with randomized elements. Participants: Fracture surgeons of the Science of Variation Group, an international collaborative of musculoskeletal surgeons who studies variation in care, were invited to participate. Surgeons with limited experience treating multiligament knee injuries were asked to self-exclude. Outcome Measures and Comparisons: Surgeon recommendations for operative treatment, timing of surgery, and use of open surgery in addition to arthroscopy were measured. Patient factors (age, time from injury, contralateral fracture, knee dislocation, combinations of ruptured ligaments, and preexisting osteoarthritis) and surgeon factors (gender, practice location, years of experience, and supervision of trainees) associated with surgeon recommendations were assessed. RESULTS: Eighty-five surgeons participated, of which most were men (89%) and practiced in the United States (44%) or Europe (38%). Operative treatment was less likely among older patients (odds ratio [OR] = 0.051) and preexisting osteoarthritis (OR = 0.32) and more likely in knee dislocation (OR = 1.9) and disruption of anterior cruciate ligament, posterior cruciate ligament, and lateral collateral ligament with or without medial collateral ligament (MCL; OR = 5.1 and OR = 3.1, respectively). Disruption of anterior cruciate ligament, posterior cruciate ligament, and medial collateral ligament was associated with shorter time to surgery (β = -11). Longer time to surgery was associated with contralateral fracture (β = 9.2) and surgeons supervising trainees (β = 23) and practicing in Europe (β = 13). Surgeon factors accounted for more variation in timing than patient and injury factors (5.1% vs. 1.4%, respectively). Open surgery was more likely in patients with lateral collateral ligament injury (OR = 2.9 to 3.3). CONCLUSIONS: The observation that surgeons were more likely to operate in younger patients with more severe injury has face validity, while the finding that surgeon factors accounted for more variation in timing of surgery than patient or injury factors suggests that treatment variation is based on opinion more so than evidence.