Objective
We sought to determine whether conservative or surgical therapy is more cost effective for the initial treatment of stress urinary incontinence (SUI).
Study Design
We created a decision tree model to compare costs and cost effectiveness of 3 strategies for the initial treatment of SUI: (1) continence pessary, (2) pelvic floor muscle therapy (PFMT), and (3) midurethral sling (MUS). We identified probabilities of SUI after 12 months of use of a pessary, PFMT, or MUS using published data. Parameter estimates included Health Utility Indices of no incontinence (.93) and persistent incontinence (0.7) after treatment. Morbidities associated with MUS included mesh erosion, retention, de novo urge incontinence, and recurrent SUI. Cost data were derived from Medicare in 2012 US dollars. One- and 2-way sensitivity analysis was used to examine the effect of varying rates of pursuing surgery if conservative management failed and rates of SUI cure with pessaries and PFMT. The primary outcome was an incremental cost-effectiveness ratio threshold <$50,000.
Results
Compared to PFMT, initial treatment of SUI with MUS was the more cost-effective strategy with an incremental cost-effectiveness ratio of $32,132/quality-adjusted life year. Initial treatment with PFMT was also acceptable as long as subjective cure was >35%. In 3-way sensitivity analysis, subjective cure would need to be >40.5% for PFMT and 43.5% for a continence pessary for the MUS scenario to not be the preferred strategy.
Conclusion
At 1 year, MUS is more cost effective than a continence pessary or PFMT for the initial treatment for SUI.
The Ambulatory Treatments for Leakage Associated with Stress Incontinence (ATLAS) trial was a randomized controlled trial (RCT) of ambulatory therapies for stress urinary incontinence (SUI) that compared continence pessaries, pelvic floor muscle therapy (PFMT), and combined therapy. Conservative treatment with PFMT and continence pessary treatment improved patient symptoms at 12 months, with one third of study subjects reporting feeling “much better” or “very much better” on the Patient Global Impression-Index (PGI-I). Greater than one third of study subjects also reported a subjective cure of SUI, defined as no longer having bothersome stress incontinence symptoms on the Urogenital Distress Inventory.
To date, there is not a standardized algorithm for the initial treatment of SUI. Many physicians offer conservative management to all of their stress incontinent patients due to the low risks and reversible nature of the treatment. However, conservative treatment success rates are low compared to surgical therapy, and improvement and cure rates diminish over the course of a year. Also, conservative management with continence pessaries or PFMT can be labor intensive for patients, leading many patients to stop pursuing conservative therapy. Patients treated with midurethral slings (MUS) for SUI report higher rates of success between 85-95%. In the era of MUS where success rates are high and serious side effects are low, it is unknown which initial therapy should be offered for the initial treatment of SUI.
We aim to model the overall costs, effectiveness, and cost effectiveness of 3 strategies for the initial treatment of SUI: (1) use of a continence pessary, (2) behavioral treatment with PFMT, or (3) surgical therapy with MUS.
Materials and Methods
Decision model
We performed an institutional review board–exempt cost-effectiveness analysis using publically available data. A decision-analytic model was developed to compare the costs and effectiveness for the initial treatment of SUI with the use of: (1) a continence pessary, (2) behavioral treatment with PFMT, or (3) surgical therapy with MUS ( Figure 1 ).
We modeled a representative population of women with uncomplicated, symptomatic SUI. All women had a normal neurologic examination result, normal postvoid residual urine, and no prior surgical treatment for urinary incontinence. Our base case for our decision tree used data from the ATLAS trial, where women were randomized to behavioral therapy, a continence pessary, or both. Behavioral therapy consisted of PFMT that was implemented in 4 visits every 2 weeks. Women were also given a home program prescription at the end of 8 weeks to maintain treatment. Women fit with a pessary had an 89% chance of a successful fit, as reported in a large study of pessary use for the treatment of incontinence. If they did not achieve a successful fit, they would enter either the MUS or PFMT scenario. Our base-case scenario assumed outcomes of “much better” or “very much better” on the PGI-I in 32.9% of the PFMT group and 31.5% of the continence pessary group as reported in the ATLAS trial at 12 months. If the patient was still bothered with her SUI symptoms after a continence pessary, we modeled a choice of no further treatment, PFMT, or MUS. If the patient was still bothered with her SUI symptoms after PFMT, we modeled a choice of no further treatment or MUS. If the patient was still bothered with her SUI symptoms after MUS, we modeled a choice of no further treatment or MUS ( Figure 1 ). All MUS surgeries were modeled as outpatient procedures. We examined a 1-year time frame to reflect the follow-up data available from the ATLAS trial and a RCT of MUS.
We chose to model a retropubic MUS for surgical treatment of SUI. Our base-case rate of subjective cure was 90.8% as measured by the PGI-I in a randomized trial of MUS vs PFMT. Morbidities included in this model have been previously published and include mesh erosion, urinary retention requiring operative take back, de novo urge incontinence, and recurrent SUI. Patients who developed de novo urge incontinence from a MUS were treated with anticholinergic medication. Most outcomes were modeled after a RCT of MUS. Mesh exposure removal within 1 year of MUS placement was estimated at 1.3%. In patients undergoing MUS removal for mesh exposure, the rate of SUI was assumed to be 36%. For patients who developed urinary retention needing sling release, the rate of recurrent SUI was assumed to be 13%, as reported in the largest retrospective study to date. Patients with recurrent incontinence after sling release were not offered a repeat sling procedure. Patients with SUI after failed or removed MUS were able to undergo 1 additional MUS. Of those undergoing a second MUS, the same outcome algorithm was applied, with the exception that no further MUS would be offered if SUI persisted. All base-case assumptions can be found in the Table .
| Parameter estimates | Base case | Range | Source for base case |
|---|---|---|---|
| Probability of cure with PFMT | 0.329 | 0.2–0.6 | 1 |
| Probability of cure with pessary | 0.315 | 0.0–0.54 | 1 |
| Probability of successful pessary fit | 0.89 | 0.79–1.0 | 4 |
| Probability of choosing MUS after failed PFMT or pessary | 0.49 | 0.29–69 | 2 |
| Probability of de novo urge incontinence after MUS placement | 0.04 | 0.02–0.06 | 3 |
| Probability of urinary retention after MUS placement that requires sling release | 0.058 | 0.01–0.08 | 3 |
| Probability of SUI after sling release | 0.13 | 0.0–0.20 | 8 |
| Probability of mesh exposure requiring sling removal/excision | 0.013 | 0.0–0.06 | 6 |
| Probability of SUI following MUS removal for mesh exposure | 0.36 | 0.20–0.50 | 7 |
| Cost of pessary treatment | $220 | $100–350 | 9 |
| Cost of PFMT treatment | $660 | $500–1900 | 9 |
| Cost of outpatient MUS | $3938 | $2170–6510 | 9 |
| Cost of outpatient sling release for urinary retention | $3290 | $1645–4935 | 9 |
| Cost of outpatient sling removal for mesh exposure | $3460 | $1730–5190 | 9 |
| Cost of anticholinergic medication for 1 y | $520 | $250–1050 | 10 |
| Health utility value for no postoperative incontinence | 0.93 | 0.7–.93 | 11 |
| Health utility value for SUI | 0.71 | 0.5–.93 | 11 |
| Health utility value for urge incontinence on anticholinergic medication | 0.82 | 0.60–.93 | 12 |
In addition to base-case analysis, 1- and 2-way sensitivity analysis was used to examine the effect of varying rates of subjective cure of SUI for all 3 treatment scenarios.
Model descriptions and assumptions
Costs
We conducted our analysis from the perspective of a third-party payer; we specifically used Medicare reimbursement to provide cost estimates. We used only direct costs and did not use charges billed or indirect costs such as travel or lost time from work. We based physician costs on Current Procedural Terminology codes and the associated 2012 physician fee schedules. Outpatient costs were based on the ambulatory procedure code for hospital Medicare reimbursement. Costs related to pessary fitting and maintenance for 1 year were estimated at $220. PFMT was modeled for 4 sessions over 8 weeks at an estimated cost of $660. Costs related to outpatient MUS were estimated at $3940. Costs related to outpatient release of MUS for urinary retention were estimated to be $3290 and outpatient removal of MUS for mesh exposure were estimated to be $3460. De novo urge incontinence was treated with anticholinergic medication and was estimated to be $520 for a patient on Medicare. All base-case costs and ranges for sensitivity analysis are listed in the Table .
Health outcomes
Effectiveness was measured by quality-adjusted life years (QALYs). QALYs were calculated based on health state utility scores, which can range from 0-1 (0 representing death and 1 representing perfect health). Based on published estimates, for patients cured of SUI we assumed a utility of 0.93 and for patients with SUI, a utility of 0.71. For those treated with anticholinergic medications secondary to de novo urge incontinence the base-case utility score was 0.88. The time horizon for this model was 1 year; therefore we did not use a discounted rate per year.
The primary outcome modeled was the incremental cost-effectiveness ratio (ICER)–the ratio of the difference in costs to the difference in QALYs between scenarios. A commonly used ICER threshold of <$50,000/QALY was considered cost effective in this study.
Results
Using our base-case assumptions, MUS for initial treatment of SUI was the preferred strategy with an ICER of $32,132/QALY when compared to initial treatment with PFMT. One-way sensitivity analysis across our credible ranges of probability estimates found that if subjective cure of SUI with PFMT fell <33%, then the scenario would be dominated (cost more, less effective) and never be cost effective. If subjective cure of SUI with PFMT was >44% then it would be the preferred scenario over MUS. Initially treating patients with a continence pessary was only cost effective if subjective cure rates of SUI were above that of PFMT, and in this base case would need to be >35%. Two-way sensitivity analysis found similar results where the MUS scenario would be the most preferred strategy unless subjective cure was >40.5% for PFMT and 43.5% with a continence pessary ( Figure 2 ).
