RESUMO
BACKGROUND CONTEXT: Several investigators have identified an explosive increase in spinal injection rates in the Veterans Administration and Medicare populations. Furthermore, utilization of spinal injection procedures appears to vary by geographic location, subspecialty, and practice setting. Medicare claims analysis has shown that a small percentage of physicians perform a disproportionately large number of injections. Although Medicare utilization has been well characterized, the utilization patterns for privately insured individuals are not clearly known. PURPOSE: The primary purpose of this article was to investigate whether relatively few providers are responsible for a disproportionately high percentage of interventional spine procedures in privately insured plans and to quantify any such findings. The secondary purpose was to determine if provider specialty is a relevant variable in any identified patterns of disproportionate utilization. STUDY DESIGN: A descriptive analysis of utilization patterns using the Medstat MarketScan database was conducted between 2003 and 2007. The database contains deidentified medical, pharmacy, and enrollment claims representing 12 to 14 million individuals. PATIENT SAMPLE: A data set was generated based on the following inclusion criteria: all patients aged between 18 and 99 years receiving at least one spinal interventional procedure between 2003 and 2007: epidural steroid injections, intra-articular facet or medial branch blocks, medial branch radio frequency neurotomy, sacroiliac joint injections, and discography. Our inclusion criteria yielded data on nearly 200,000 patients treated by over 20,000 providers. OUTCOME MEASURES: Not applicable. METHODS: The number of procedures was tallied for a 12-month period beginning with a patient's first procedure claim. The total number of procedures per patient and the mean number of procedures per patient were calculated for the study sample. Within each specialty, all spinal procedures were summed for each individual provider within each procedure category and as an overall total. The overall mean number of therapeutic procedures per patient for all physicians within a specialty was calculated. Within each specialty, the total number of procedures performed by each physician was analyzed in percentiles to highlight any disparity between high- and low-using providers. RESULTS: The final therapeutic procedure data set contained 196,332 patients who received 875,627 procedures. The principal nine specialties performing these procedures were anesthesiology (49.2% of the total number of procedures in the final data set), physiatry (12.5%), pain management (12.0%), family practice (10.2%), orthopedics (5.5%), radiology (3.0%), neurology (2.8%), internal medicine (2.8%), and neurosurgery (1.9%). The overall mean number of procedures across all categories performed per patient during the 12-month inclusion period was 4.46±6.44. Neurologists and pain management specialists were the only provider groups in which the mean number of procedures per patient exceeded the overall mean. The highest 10% of providers, which encompasses those providers performing a mean greater than or equal to 5.08 procedures per patient per year, perform 36.6% of the total spinal procedures performed. The highest 20% of providers, which encompasses those providers with a mean greater than or equal to 3.75, account for 57.6% of all spinal procedures. The highest 10% of providers perform nine times more procedures per patient compared with the lowest 10% and 4.5 times more procedures than the median. This same pattern of high utilization by disproportionately few providers was observed across all nine specialties. CONCLUSIONS: These findings demonstrate that relatively few providers are responsible for a disproportionately high percentage of interventional spine procedures. This pattern of marked overutilization by a minority of providers is the dominant characteristic of utilization within all specialties.