Study design, materials and methods
This is a within-trial comparison of participants in the Refractory Overactive Bladder: Sacral Neuromodulation vs Botulinum Toxin Assessment (ROSETTA) multicenter randomized trial (February 2012- July 2016) at 9 US medical centers. Women with >6 UUIE over 3 days refractory to medications were randomized to 200 units of Botox performed in clinic or SNM performed as a staged procedure. Women in the SNM group with ≥50% UUIE reduction after Stage 1 were eligible for Stage 2 implant. The Botox group were allowed 2 additional injections, performed a minimum of 4 months apart. The analysis includes the intent-to-treat population with 24-month UUI data and was conducted from a health care sector perspective. Costs include treatment and medical utilization during the 24-month trial; results are in 2017 US dollars with a 3% discount rate. A resource-based costing method was utilized; medical service use was multiplied by average unit cost of the service with price weights based on national Medicare payment averages. Effectiveness was measured as mean decrease in daily UUIE. Our primary outcome was cost-effectiveness, defined as the cost per reduction of one UUIE/day. A decision analytic model was developed from trial data to evaluate the treatments through 5 years. Sensitivity analyses were performed to assess whether variable values affected results.
Results
A total of 386 women were randomized; 84% of the SNM group were eligible for stage 2 implant. Over 24 months, 72% of the Botox group requested a second injection and 47% requested a third injection. Baseline characteristics were similar between groups: 82.6% were white, mean age was 63.2 ± 11.39 years, mean BMI was 32.1 ± 8.53 kg/m2, and mean UUIE/day was 5.2 ± 2.60. This analysis includes 141 (81%) of SNM and 152 (80%) of Botox participants. The mean reduction in UUIE/day at 24 months was not significantly different between groups (-2.93 ± 0.28 SNM vs -2.98 ± 0.31 Botox, p=0.8878). The cumulative mean per-person costs over 24 months were $38,260 for the SNM group and $7,836 for the Botox group (P<0.01).
Interpretation of results
The cost-effectiveness results through 24 months favored Botox, showing the SNM group incurred $10,429 higher costs per reduction of one UUIE/day than the Botox group. Estimates through 5 years also favored Botox, showing the SNM group incurred $9,740 higher costs per reduction of one UUIE/day. Threshold analysis was performed; with reduction of SNM device and procedure cost by 85%, Botox remained cost-effective at 24-months but SNM was more cost-effective at 5 years. With increased frequency of Botox injections to 5 per year, SNM was more cost-effective at 5 years.