The effect of operator variability and experience in vertebroplasty outcomes

Purpose: To determine if operator variability, prior vertebroplasty experience, or acquired vertebroplasty experience affects clinical outcomes. Materials and Methods: Informed consent was obtained from all patients and all data were handled in a manner consistent with institutional review board guidelines and the Health Insurance Portability and Accountability Act. Outcomes from 841 vertebroplasties, sorted by operator, were studied; two operators had previous vertebroplasty experience and five were neurointerventionalists who were initially new to the procedure. Objective (Roland-Morris Disability Questionnaire [RDQ] and analog pain scales) and subjective (mobility, narcotic use) scores were monitored before and after vertebroplasty at specified intervals following treatment. Perioperative cement volume utilization, complications, and number of treated levels were recorded. Random-effects and repeated-measures analyses of variance were used to assess operator differences and variability in clinical outcomes while generalized linear mixed-model regression analyses were used to track changes in clinical outcomes over time. Results: All operators provided similar average durable clinical improvements in postoperative pain and disability with non-significant interoperator variability. However, generalized linear model regression suggests that four of five initially novice operators showed significant changes in several procedural measures and clinical outcomes over the study timeframe, including cement volume utilization (reduction of 3.64 to 4.63 cm³), 1 week RDQ score (reduction of 4.79 to 8.62 points) and postoperative rest pain (reduction of 1.18 to 2.03 points). Conclusion: Changes over time in measured outcomes suggest the presence of a training effect among novice operators. Cement volume utilization and immediate postprocedural pain decrease with experience but long-term clinical outcomes are insensitive to operator experience.