Cost Effectiveness of the Use of Prophylactic Mesh To Prevent Parastomal Hernia After Urinary Diversion with an Ileal Conduit

Take Home Message The use of prophylactic mesh is cost-effective in reducing the risk of parastomal hernia after ileal conduit diversion.

Patient summary: In patients having their bladder surgically removed, a mesh implant can be inserted when a portion of the intestine is used to create an opening to drain urine from the body. Our results show that mesh use to prevent development of a hernia at the opening where urine exits the body is cost-effective from the perspective of health care providers.

Introduction
Parastomal hernia (PSH) is a common complication after stoma creation that leads to a decrease in quality of life for patients and increases in health care costs [1]. Metaanalyses have shown that use of a prophylactic mesh reduces the rate of PSH [2,3] for patients receiving end colostomies. In addition, a prospective randomised multicentre study showed that prophylactic implantation of a lightweight mesh decreased the risk of PSH when constructing an ileal conduit in comparison to no mesh [4]. However, decision-makers need to know whether prophylactic mesh in this setting represents good use of limited resources and is cost-effective before its adoption in routine clinical practice for patients receiving ileal conduit diversion. Cost-effectiveness analyses (CEAs) are central to national reimbursement decisions on new health technologies or treatment methods both in general and for high-income countries such as Sweden [5]. Informed and transparent decision-making is important for efficient allocation of scarce health care resources. A previous study on the use of mesh prophylaxis to prevent PSH showed that this is cost-effective for patients undergoing permanent colostomy for rectal cancer [6]. In another CEA, synthetic mesh was the most cost-effective approach in preventing PSH when compared to a biological mesh and no mesh for patients undergoing end-colostomy creation during rectal cancer surgery [7]. However, to the best of our knowledge there has been no research on the cost-effectiveness of mesh in preventing PSH in patients receiving an ileal conduit. Thus, the objective of this study was to perform a CEA on the use of prophylactic mesh to prevent PSH after ileal conduit diversion from a health care provider perspective in Sweden. The exclusion criteria were a previous stoma and lack of informed consent.

Patients and methods
In the experimental arm a lightweight mesh was placed in a sublay position between the rectus abdominis muscle and the posterior rectus sheath, and the conduit was brought out through a cross-shaped incision in the centre of the mesh. The mesh was anchored to the posterior rectus sheath with 2-0 polydioxanone (PDS) sutures in each corner. The conduit was fixed at the 6-and 12-o'clock positions to the anterior rectus sheath with a 4-0 resorbable suture in both the control arm and the intervention arm, and three monofilament 4-0 resorbable sutures were also used to evert and mature the ileal conduit in all patients. Further details of the surgery performed and of the trial, including the clinical effectiveness of the intervention, can be found elsewhere [4].
As the cost data were from the Skåne County Council register,

Effect measures
We used the primary outcome of the trial, which was the incidence of clinical PSH during follow-up [10]. Clinical PSH was assessed at followup visits at 6, 12, and 24 mo postoperatively, as well as at later visits scheduled at the discretion of the treating urologist. Clinical PSH incidence was registered without any a priori definitions applied for clinical PSH, and both symptomatic and asymptomatic findings were reported.
Since this is a dichotomous variable (yes/no), we present the outcomes as percentages and estimated the probability of having PSH to achieve individual variations. To this end, we performed multivariable logistic regression for which PSH incidence was the dependent variable (yes/ no) and the independent variables were age, sex, and prophylactic mesh status (mesh/no mesh).

Statistical analyses
Statistical analyses were performed to assess differences between the mesh and no-mesh groups using t tests for continuous variables, v 2 tests for dichotomous variables, and Fisher's exact test for PSH incidence. We also performed linear and logistic regressions in sensitivity analyses controlled for the variables that might influence costs and PSH for our inves-

Analysis of cost-effectiveness
Cost-effectiveness was estimated as the incremental cost-effectiveness ratio (ICER), which is the ratio of the difference in average costs per patient to the difference in health effects per patient Effects mesh ÀEffects no mesh for the mesh group compared to the nomesh group. Sampling uncertainty was assessed using 5000 bootstrap resamples to estimate the ICERs [12]. These bootstrapped ICERs and point estimates are graphically presented on a four-quadrant costeffectiveness plane (CE-plane), with effect differences plotted on the xaxis and cost differences on the y-axis between the mesh and no-mesh groups [13]. We present the effect differences as the probability of not

Base case analysis
The base-case CEA includes all the costs (inpatient, outpatient, operation time, and intervention costs) and clinical PSH incidence as the outcome.

Sensitivity analyses and subgroup analyses
Several sensitivity and subgroup analyses were performed to capture uncertainties around the base-case estimate.

Controlling for variables that might affect costs
We performed multivariable linear regression analysis to control for variables that might affect costs. These variables were smoking status, American Society of Anesthesiology (ASA) score, gender, operating hospital, body mass index (BMI), use of preoperative chemotherapy, follow-up duration, previous midline laparotomy, and treatment (mesh vs no-mesh group) in the first scenario (scenario 1a), and then controlled for the significant variables, which were ASA score and follow-up duration in scenario 1b.

Controlling for variables that might affect the risk of PSH
We used the same variables as in scenario 1a to control for the effect on risk of PSH. However, we performed multivariable logistic regression analysis with PSH incidence (yes/no) as the dependent variable.

Removal of cost outliers
To exclude low-cost and high-cost outliers for health care costs, participants with the top 5% costs and bottom 5% costs were excluded from the analysis.

Subgroup analysis by sex
Owing to the higher risk of PSH among women [16], we stratified the results by sex.

Subgroup analysis by ASA score
The results were also stratified by ASA score since the costs and health effects may be affected by comorbidity. We merged patients with ASA scores of III and IV together into one class.

Subgroup analysis by overweight status
The results were stratified by overweight status on the basis of hypothesis-generating findings suggesting a higher risk of PSH for individuals with high BMI [16]. We used BMI of 25 kg/m 2 as the threshold for overweight according to the World Health Organization [17].

Subgroup analysis by age
The results were also stratified by age ( 71 yr vs !72 yr), with the cutoff for dichotomisation chosen according to the median age at cystectomy in Skåne county during 2016-2020, which was 71 yr [18].

Patients and surgical characteristics
Even though this CEA is based on the subgroup of patients residing in Skåne county who participated in the original trial, there were no significant differences in patient characteristics between the two groups except for operation time (Table 1). On average, operation time was longer for the mesh group (p = 0.003) than for the no-mesh group.

Costs and effects
The mean inpatient costs were €60 726 and €63 811 and the mean outpatient costs were €22 337 and €23 758 for the mesh and no-mesh groups, respectively ( Table 2). There were no significant differences in total, inpatient, or outpatient costs, but there was a significant difference in costs related to operation time. The operation time cost was €598 higher for the mesh group than for the no-mesh group (p = 0.032). The mean total cost per patient was €69 837 in the mesh group and €71 884 in the no-mesh group (Table 2), with a nonsignificant cost difference of À€2047 (95% CI À€16 441 to €12 348). Cumulative PSH incidence was lower in the mesh group (n = 7/80, 8%) than in the no-mesh group (n = 17/79, 22%; p = 0.014; Table 3). The number needed to treat to prevent one PSH case was seven patients. No longterm complications related to mesh use in the intervention group, such as mesh infections or erosions, were observed during follow-up. One patient in the mesh group and three patients in the no-mesh group required surgical PSH repair.

Base-case, sensitivity, and subgroup analyses
For the base-case analysis, the mesh group dominated the no-mesh group: the intervention was less costly and more effective. On the CE-plane, 61% of the pairs were in the SE quadrant (less costly and more effective), followed by 39% in the NE quadrant (more costly and more effective; Fig. 2). The base-case results were not sensitive to any of the scenarios except when the cost outliers were removed (

Discussion
We have performed cost-effectiveness analyses (CEA) of using prophylactic mesh to prevent PSH after ileal conduit diversion from a health care provider perspective. The use of a prophylactic mesh reduced cost (although not statistically significant) and the incidence of PSH. The costeffectiveness plane also indicates that 61% of the costeffect pairs were in the southeast quadrant. The subgroup analyses revealed that in men to prevent one PSH, health care needs to spend €18 000. Since there is no established threshold on how much society is willing to pay to prevent one PSH, it is difficult to interpret this finding as costeffective. Similarly, in the subgroups with ASA score above II and patients aged 72 yr of age and above the ICERs were not dominant. Consequently, these findings indicate that using mesh in women, in patients with ASA class I and in patients younger than 71 yr are more beneficial than their corresponding counterparts, respectively. Since our study uses a patient cohort from the first randomised trial examining the use of prophylactic mesh to prevent PSH in patients receiving an ileal conduit and hence is the first CEA in this setting, it is difficult to compare the findings with similar studies. However, researchers in Canada reported that a prophylactic mesh was dominant in preventing PSH in patients with rectal cancer receiving a permanent colostomy, which is in line with our findings [6]. They also found that mesh was not dominant in advanced disease (rectal cancer stage IV), a result also in line with our observation for patients with more advanced comorbidity (ASA score !II).
The main study limitation is the lack of statistical power, as the power calculation was conducted from the clinical outcome perspective instead of a CEA perspective. A post hoc power calculation for cost-effectiveness using cost and outcome data from the present trial was therefore conducted [19,20]. As mentioned earlier, cost-effectiveness depends on societal WTP, and there is no agreed WTP threshold for preventing incidence of one PSH case. For WTP thresholds ranging from €100 to €100 000, the power calculation showed that, a sample size ranging from eight to 244 patients per group would be needed. Therefore, a CEA based on a simulation model in which data from this and/ or similar future randomised trials could be included might be a more suitable approach for evaluating the costeffectiveness.
Another limitation of the present study is that the trial did not include any PSH-specific outcome measures, such  as hernia-related problems with stoma appliances or local pain; however, only a minority of patients with PSH are asymptomatic [21]. Furthermore, the lack of a more structured definition for clinical PSH, such as the European Hernia Society stratification of PSH based on hernia size that was applied in the aforementioned study [21], is a study limitation. Similarly, no generic or disease-specific instrument for measuring patient quality of life was available for the CEA. A generic instrument such as the European Quality of Life 5 Dimension (EQ-5D) would have been helpful in estimating quality-adjusted life year (QALY) [22] gains for patients receiving a mesh. QALYs not only capture patient quality of life resulting from the intervention of interest but also facilitate its comparison with other similar or different interventions in terms of cost-effectiveness. Furthermore, the lack of cost estimates from primary care and medication costs are study limitations. However, since the patients underwent surgery in a regionalised cystectomy setting in two hospitals, it is reasonable to assume that they would contact the specialised unit rather than their primary health care centre in the case of any complications or health-related issues. To the best of our knowledge, this study is the first CEA conducted in this setting, and the preciseness of the cost data and the randomised controlled study design add strength to the results.

Conclusions
The use of prophylactic mesh is cost-effective in reducing the risk of PSH after ileal conduit diversion when compared to standard care. A CEA based on larger randomised trials or a simulation model incorporating patient quality of life would further validate these findings.