Benefits of inferential statistical methods in radiation exposure studies: another look at percutaneous spinal cord...

Two studies, each consisting of large sample sets, were recently published on radiation exposure in percutaneous spinal cord stimulation (SCS) trialing procedures. A more rigorous use of statistical methods in the second study more accurately defined benchmark reference levels. Principally, one physician implanter-considered an advanced interventional pain physician-performed all such procedures to nullify inter-physician variability. However, the literature is sparse in articles comparing exposure levels of radiation in pain procedures conducted by novice and advanced interventionally trained physicians, and inferential statistical analysis is seldom included in radiation exposure studies.

The aim of this study was to compare fluoroscopy times between novice and expert physician implanters performing SCS trialing procedures, and to the benchmarked reference level, using inferential statistical methods. As a secondary objective, the importance of statistical concepts in interpretive-imaging and image guidance studies for interventional pain procedures will be outlined.

An observational study.

A non-university outpatient Interventional Pain Management practice in the United States.

Fluoroscopy time (in seconds) was retrospectively studied in 18 SCS trialing procedures (with dual lead placement in the low thoracic spine) performed over a 3-month period. The procedures were categorized by physician experience: one novice physician implanter with n = 5 cases and one expert physician implanter with n = 13 cases. All procedures were conducted with the same fluoroscope operator and the same mobile C-arm fluoroscopy system. A two-tailed t-test was used to compare mean fluoroscopy times between physician categories. Left-tailed t-tests were used to compare mean fluoroscopy times for each physician category separately to the benchmark level (μ = 71.7 seconds). Incident air kerma (KERMA) was assessed by nonsimplistic modeling.

No statistical difference was found in mean fluoroscopy times for SCS trialing procedures between the novice- and expert-implanter, χnovice = 63.5 seconds and χexpert = 53.9 seconds. In the case of the novice implanter, although mean fluoroscopy time was lower than the benchmark reference level, χnovice = 63.5 seconds compared to μ = 71.7 seconds, this was not significantly relevant. In the case of the expert implanter, a statistically relevant reduction in mean fluoroscopy time was observed compared to the benchmark level, χexpert = 53.9 seconds versus μ = 71.7 seconds. KERMA ranged from 5.3 mGy to 9.1 mGy with a mean and standard deviation of 6.5 mGy and 1.5 mGy, respectively, in the novice implanter sample set. KERMA ranged from 2.6 mGy to 13.1 mGy with a mean and standard deviation of 5.8 mGy and 3.2 mGy, respectively, in the expert implanter sample set.

Given that reference levels for radiation exposure in SCS trialing procedures are established, combined with comparisons in fluoroscopy times based on physician experience, expanding the physician database will assist in data validation.

Radiation exposure levels in SCS trialing procedures remain negligible. While no differences in fluoroscopy times for such procedures were detected based on physician experience, the expert implanter demonstrated the ability to use less fluoroscopy time than that of the benchmark reference level.