Table 2.

Comparison of groups after adjusting for body mass index (BMI)

Variable	Mean difference between posteroanterior and oblique time	T-statistic	p-value
Procedure time (sec)	109.5	2.33	0.03
Needling time (sec)	46.6	3.08	0.001
Fluoroscopy time (sec)	1.2*	1.24†	0.26

When adjusting for BMI, mean procedure time and needle time were significantly less in the posteroanterior group than the oblique group, but no difference was seen in fluoroscopy time using median regression analysis. Unlike the first 2 times (procedure and needling times), we used median time for fluoroscopy time. We followed it with median regression, which simply models the median instead of the mean. We tried to employ the mean regression for fluoroscopy time, but there was no statistically significant mean difference when adjusting for BMI. Again, the outliers appear to bias modeling the mean. Therefore, for fluoroscopy time we reported the results from the median regression.

^*Median time.

^†Chi-square statistic.