Experimental and control matching technique
A matched stratified design was used to control for the possible confounding effect of previous clinical treatments (epidural steroid blocks within the 6 months preceding the study, epidural steroid blocks undergone before the 6 months preceding the study, and past lumbar surgery) as
well as gender and decade of life. For each patient, a coin was tossed and the assignment to either intervention or con- trol group was made. When two patients in a row were as- signed to one group, the next patient was assigned to the other group. Five of the 40 recruited patients failed to com- plete their pretest appointment and were dropped from the study, thus leaving 35 patients for the study. Table 1 shows the AE (n