Cost-effectiveness of hydrophilic-coated intermittent catheters compared with uncoated catheters in Canada: a public payer perspective

Abstract

Study design: A Markov model was used to analyze cost-effectiveness over a lifetime horizon.

Objective: To investigate the cost-effectiveness of hydrophilic-coated intermittent catheters (HCICs) compared with uncoated catheters (UCs) among individuals with neurogenic bladder dysfunction (NB) due to spinal cord injury (SCI).

Setting: A Canadian public payer perspective based on data from Ontario; including a scenario analysis from the societal perspective.

Methods: A previously published Markov decision model was modified to compare the lifetime costs and quality-adjusted life years (QALYs) for the two interventions. Three renal function and three urinary tract infection (UTI) health states as well as other catheter-related events were included. Scenario analyses, including utility gain from compact catheter and phthalate free catheter use, were performed. Deterministic and probabilistic sensitivity analyses were conducted to evaluate the robustness of the model.

Results: The model predicted that a 50-year-old patient with SCI would gain an additional 0.72 QALYs if HCICs were used instead of UCs at an incremental cost of $48,016, leading to an incremental cost-effectiveness ratio (ICER) of $66,634/QALY. Moreover, using HCICs could reduce the lifetime number of UTI events by 11%. From the societal perspective, HCICs cost less than UCs, while providing superior outcomes in terms of QALYs, life years gained (LYG), and UTIs. The cost per QALY further decreased when health-related quality-of-life (HRQoL) gains associated with compact HCICs or catheters not containing phthalates were included.

Conclusion: In general, ICERs in the range of CAD$50–100,000 could be considered cost-effective. The ICERs for the base case and sensitivity analyses suggest that HCICs could be cost-effective. From the societal perspective, HCICs were associated with potential cost savings in our model. The results suggest that reimbursement of HCICs should be considered in these settings.

Keywords:

• Cost-effectiveness analysis
• spinal cord injury
• chronic urinary retention
• hydrophilic catheters
• intermittent catheterization
• Markov model

Introduction

Bladder dysfunction caused by neurologic diseases or damage such as spinal cord injuries (SCI) is usually referred to as neurogenic bladder (NB). After SCIs, 70–84% individuals cannot void and need some form of catheterization¹. Intermittent catheterization (IC) is the recommended bladder-emptying method²; compared with indwelling catheters it reduces the occurrence of urinary tract infections (UTIs) and complications^3–7. Moreover, 55% of all Canadians with SCIs use IC⁸.

Several types of catheters can be used for IC. The main types of catheters available are uncoated PVC-catheters (UCs) requiring manual lubrication with gel before insertion, and hydrophilic-coated intermittent catheters (HCICs) that come with a lubricious surface, and can often be used straight from the package. A Canadian survey study found that 74% of IC-users catheterize with UCs, 15% with HCICs, and the remaining 11% use both UCs and HCICs⁸.

Because of its coating properties, HCICs are intended only for single use, whereas UCs may be reused. The guidelines of the Canadian Urological Association permit the reuse of intermittent catheters for up to a week or until physical damage is noticed in the home setting⁹. However, concerns regarding its efficacy and patient compliance with cleansing techniques have been raised, and the physical damage over time poses a theoretical safety risk^10–12. Additionally, no catheters sold in Canada have been registered or approved for re-use, and the liability associated with off-label reuse has been questioned¹³. Therefore, the current analysis only involves single-use catheters.

Because of its lubricious coating, HCICs have been shown to cause less pain/discomfort and to reduce micro-hematuria compared with UCs^14–16. However, current evidence is limited, and the clinical impact of the different IC types has been debated^10,17. The two most recent meta-analyses have revealed that HCICs reduce UTI incidences; however, additional high-quality randomized controlled trials (RCTs) are warranted^10,18. Several studies reported that HCICs are preferred over UCs because of the added convenience and ease of use^15,19–21; however, some studies have reported no differences in overall satisfaction^22,23.

New HCIC developments include small, telescopic catheters that are easy and discrete to carry around^11,24 and phthalate-free catheters that eliminate the negative impact of phthalates on human health^25,26. Most uncoated PVC catheters contain softeners such as phthalates, whereas an increasing number of HCICs do not contain phthalates²⁷.

In Canada, IC is performed using either UCs or HCICs, depending on factors such as reimbursement/out-of-pocket costs, clinical practice, and conventions among the users and prescribers. Access to catheters varies between provinces, and, in Ontario (the largest province in Canada), no general reimbursement is currently provided by the public healthcare system for community-dwelling individuals needing IC because of NB. In 2006, the Ontario Ministry of Health and Long-Term Care issued a report on HCICs, which indicated insufficient evidence regarding the effects of hydrophilic coating on UTIs and patient satisfaction²⁸. Since then, additional evidence has been generated, and broader public funding of IC is currently being considered.

HCICs offer several advantages; however, they are costlier per catheter than UCs. A cost-effectiveness analysis (CEA) will provide data to assist decision-makers in determining whether HCICs offer good value for money from the public payer’s perspective. Limited numbers of CEAs have been conducted on this topic; however, none have been performed in a Canadian setting^27,29–33. Therefore, the present study aimed to determine the cost-effectiveness of HCICs compared with UCs in adult Canadian community-dwelling individuals with SCI and NB.

Materials and methods

Structure of the model

The analysis was performed from the perspective of the Canadian public healthcare system and used Ontario as the reference province for costs and resources. Hence, the analysis was performed from the perspective of the Ontario Ministry of Health and Long-Term Care. A probabilistic Markov decision model developed in Microsoft Excel³⁰ for a UK setting was modified using data for the Ontario setting. This model structure was chosen because it was constructed for the same SCI study population and because, unlike the model developed by Bermingham et al.²⁹, it included long-term sequelae of UTI³⁰.

A hypothetical cohort suffering from NB due to traumatic SCI was simulated for a lifetime horizon. Baseline characteristics of the cohort reflected the SCI population in Ontario with 79% of males³⁴. The model structure consisted of different health states that a catheter dependent individual with SCI can experience; it was validated for its applicability to the Canadian setting by three Canadian medical experts with extensive experience of IC and the study population. A 1-month cycle length was used.

The SCI cohort was classified according to renal function (RF), and, similar to the original model by Clark et al.³⁰, three health states were identified: RF1 corresponded to “no or minor renal impairment”, RF2 corresponded to “major renal impairment requiring intervention along with careful medical monitoring”, and RF3 corresponded to “renal failure requiring dialysis or transplantation”. Each of the renal health states contained three UTI-related nested health states, where UTI1 corresponded to “no presence of treatment-requiring UTI”, UTI2 corresponded to the “presence of UTI responsive to initial treatment”, and UTI3 corresponded to the “presence of UTI unresponsive to initial treatment”. Patients were assumed to have equal rates of being in the specific UTI-related health states across all renal states³⁰. Death was considered a possible outcome from all renal- and UTI-related health states. The cohort was assumed to start in the health state “UTI1; RF1”.

In each monthly cycle, the hypothetical cohort could remain in a given health state or move between the health states based on a set of transition probabilities. Movement to a less severe renal health state was not permitted, because renal impairment and major renal failure were assumed to cause some degree of irreversible renal damage³⁵. Progression from one RF health state to the next was based on renal complications, and it was possible to move between the three nested UTI-related health states based on the risk of UTIs and their responsiveness to initial treatment. The model structure is illustrated in Figure 1, and the transition probabilities are listed in Supplementary Tables S1 and S2.

Figure 1. Schematic representation of the Markov model. The three outer boxes represent the renal function health states (RFI-III) and the three inner boxes represent the UTI-related health states (UTI-III). The arrows either illustrate “remain in health state” or the “progression to next health state”.

Data input

A literature search was carried out using PubMed, Medline, and the Cochrane Library using different combinations of the following terms: catheters, intermittent, hydrophilic, coated, urinary, cost-effectiveness, cost-utility, chronic urinary retention, renal, function, sepsis, spinal cord injury, urethral damage, resistance, E. coli, economic evaluation, health economic modeling, and Markov model. A total 2,116 publications were identified; studies focusing on SCI patients using IC and publications in the English language were selected. Meta-analyses relying on RCTs for this patient group were prioritized. Few studies from Canada were identified, and data from international studies were also used.

Key model input parameters are shown in Table 1. The most common IC-related complication is UTI³⁶. The results from a Canadian postal survey by Woodbury et al.⁸ were used to determine the baseline risks of UTIs. According to this study, UC users experience an average of 2.62 UTIs per year, corresponding to an average of 0.21 UTIs per month.

Table 1. Key input parameters.

Parameters	Mean^a	Source/assumption
Monthly adverse event rates (uncoated catheters)
Baseline risk of UTI	21.8%	Woodbury et al.^8
UTI unresponsive to initial treatment	0.32%	Perrouin-Verbe et al.^64; Chai et al.^65; Weld and Dmochowski^3
Major renal impairment	0.020%	Clark et al.^30
Renal failure	0.004%	Clark et al.^30
Bladder stones	0.12%	Perrouin-Verbe et al.^64; Chai et al.^65
Kidney stones	0.12%	Perrouin-Verbe et al.^64; Chai et al.^65
Urethral damage	0.19%	Perrouin-Verbe et al.^64; Chai et al.^65
Treatment effect (HCICs vs UCs)
UTI responsive to initial treatment	0.84	Rognoni and Tarricone^32
UTI unresponsive to initial treatment	0.90	Clark et al.^30
Bladder stones	0.90	Clark et al.^30
Kidney stones	0.90	Clark et al.^30
Urethral damage	0.90	Clark et al.^30
Renal impairment	1	Clark et al.^30
Renal failure	1	Clark et al.^30
Utility decrements
Baseline utility of catheterization (UCs)	0.450	Ara and Brazier^52; Liu et al.^51
UTI responsive to initial treatment	0.092	Bermingham et al.^29; Szende et al.^53
UTI responsive to initial treatment, antibiotic resistant	0.125	Bermingham et al.^29; Szende et al.^53
UTI unresponsive to initial treatment	0.16	Bermingham et al.^29; Szende et al.^53
Bladder stones	0.05	Clark et al.^30
Kidney stones	0.05	Clark et al.^30
Urethral damage	0.147	Bermingham et al.^29; Szende et al.^53
Major renal impairment	0.178	Clark et al.^30; Szende et al.^53
Renal failure	0.264	Clark et al.^30; Szende et al.^53
Utility benefits
Using HCICs instead of UCs	0.028	Igawa et al.^54
Using compact HCICs	0.031	Igawa et al.^54
Using HCICs without phthalates	0.037	Igawa et al.^54
Mortality multipliers
UTI responsive to initial treatment	1	Assumed the same mortality as in the general population
UTI responsive to initial treatment, antibiotic resistant	140.270	Klevens et al.^66; Clark et al.^30
UTI responsive to initial treatment, weighted^b	30.810	Weighted average
UTI unresponsive to initial treatment	797.600	Montgomerie^56; Clark et al.^30
Major renal impairment	18.000	Clark et al.^30
Renal failure	54.000	Clark et al.^30
Other parameters
Proportion of cohort with antibiotic resistance^c	20.6%	Karlowsky et al.^41
Daily catheterization frequency	4.0	Assumption
Sick leave because of UTI responsive to initial treatment	2.0 days	Assumption
Annual sick days because of UTI with sepsis	26.0 days	Schmid et al.^67, adjusted for SCI gender distribution in Ontario
Length of admission for UTI unresponsive to initial treatment	3.9 days	CIHI code 812

^aAll parameters are assumed to be gamma distributed, whereas utility parameters are assumed to be beta distributed.

^bCalculated as: Resistance rate * UTI responsive to initial treatment (resistant) + (1 – resistance rate) * UTI responsive to initial treatment (not resistant): 45.270*21% + 0*79% = 30.8.

^cResistance rates to E. coli are 22.1 pct. for TMP-SMX, and 19.1 pct. for ciprofloxacin.

Six relevant meta-analyses were identified^{10,18,29,37–39} from the literature search, three of which concluded that, compared with the use of UCs, the use of HCICs is associated with a statistically significant reduction in the risk of UTIs^10,18,38. One of the remaining meta-analyses²⁹ was conducted prior to the publication of the, to date, largest RCT investigating HCICs vs UCs in adults with SCIs¹⁵. The remaining two meta-analyses were Cochrane reviews, which found no evidence that the incidence of UTIs is affected by the use of HCICs or UCs. However, the most recent Cochrane review³⁷ was appraised in 2017 and withdrawn in August 2017 because of errors in data extraction and analysis¹⁰. The treatment effects of HCICs were determined based on the meta-analysis by Rognoni and Tarricone¹⁸, because this analysis was recent and only included studies on adults with SCIs. According to this analysis, the relative risk of having a UTI is 0.84 when using HCICs instead of UCs, corresponding to a treatment effect of 16%.

Mortality in the fictive cohorts was attributed to either NB complications, such as renal failure and complicated UTIs, or to all-cause mortality of the background population adjusted for age and gender. For all-cause mortality, an age-specific standardized mortality table for the general population was derived from Statistics Canada, and the most recent mortality rates from 2011–2013 were used⁴⁰.

Mortality rates were calculated by multiplying the mortality rate of the age- and gender-adjusted background population by a mortality multiplier for IC-dependent SCI patients. The mortality multipliers and their sources are listed in Table 1. Mortality attributable to being in the IC-dependent SCI population are particularly caused by multidrug-resistant UTIs, UTIs associated with urosepsis, major renal impairment, and renal failure. The health states UTI1 and RF1 contained no IC-attributable mortality. UTI2 and UTI3 included attributable mortality multipliers because of the increased risk of mortality with complicated urinary infections and the increased need for inpatient treatment; this can lead to other serious complications such as pressure ulcers, or hospital-acquired infections such as pneumonia.

An antibiotic resistance rate of 20.6% was used in the model and was calculated as the average rates of resistance to E. coli (22.1%) for TMP-SMX and ciprofloxacin (19.1%); these data were sourced from a Canadian study of a non-SCI-specific population with UTI caused by E. coli⁴¹. It is likely that the resistance rates for the SCI NB population are higher because these patients are repeatedly exposed to healthcare settings and antimicrobial agents, increasing the risk of infection with multidrug-resistant organisms²⁹.

Catheter-associated complications included in the model are listed in Table 1. Limited clinical evidence exists regarding the impact of HCICs and UCs on catheter-associated complications, and, similar to Clark et al.³⁰, an assumed 10% risk reduction was applied to urethral damage, urinary/kidney stones, and UTIs unresponsive to initial treatment.

Cost data

A weighted average of the prices obtained from three online sales websites specializing in continence care^42–44 was used to determine the acquisition costs for both the catheter types and the lubricant. The weighted average was calculated based on the commercially available sales volume of the three most sold HCICs and UCs in 2017. A hydrophilic catheter was found to cost an average of $3.77 per catheter compared with an UC, which costs an average of $1.07 per catheter. Unlike HCICs, UCs do not come packaged with a lubricant, and users, therefore, need to apply a lubricant to the catheter before each catheterization. Additionally, all IC users have to pay a monthly dispensing fee. In the scenario analysis, it was assumed that compact and phthalate-free catheters carry similar costs as HCICs. Guidelines recommend 4–6 daily catheterizations², and, based on expert input, a daily requirement of four catheters was assumed. The main costs are listed in Table 2.

Table 2. Main cost inputs.

Healthcare costs	Mean	Source
Cost per catheter (UC)	$1.07	^42–44
Cost per catheter (HCIC)	$3.77	^42–44
Daily acquisition cost (lubricant)	$0.15	^42–44
Monthly dispensing fee	$8.83	^42–44
UTI responsive to initial treatment (incl. antibiotic resistant)^a	$1,039.18 per event	Calculated according to Bermingham et al.^29; ^42–44
UTI unresponsive to initial treatment	$5,715.84 per event	CIHI Patient Cost Estimator
Major renal impairment	$2,058.93 per event	Calculated using stages 3 and 4 from Vekeman et al.^68
Renal failure	$7,070.00 per event	CIHI Patient Estimator
Kidney stones^b	$9,923.29	^42–44
Bladder stones^b	$5,349.29	^42–44
Urethral damage^c	$738.54	^42–44
Societal costs
Daily wage (79/21 pct. Male/female)	$165	Canadian Statistics (Ontario specific data) Table 202-01202, 282-0087

^aIncludes costs incurred on GP visit, urine test, 1st line antibiotic treatment, dispensing fee, non-elective hospital admission, and urine test.

^bIncludes costs of GP visit, urine test, 1st line antibiotic treatment, dispensing fee, non-elective hospital admission, and procedure cost for removal of stones.

^cIncludes cost of GP visit, urine test, dynamic studies of urinary tract, and bladder intermediate endoscopic procedure.

The schedule of benefits for physician services under the Ontario Health Insurance Plan was applied to physician services for 2016⁴⁵. Hospital costs were determined from the Patient Cost Estimator of the Canadian Institute for Health Information (CIHI) and the Ontario Case Costing (OCC) Analysis Tool, which provides both patient-level costs for inpatients and ambulatory care cases in Ontario^46,47. Because data from the CIHI and OCC were available for 2014 and 2015, respectively, the prices from these two sources were inflated to 2016 prices by using relevant information from the consumer price index (CPI) healthcare component of Statistics Canada.

Societal costs because of short- and long-term sick leaves, early retirement, and early death were calculated and included in a scenario analysis as lost salary contribution using the human capital method⁴⁸. Labour force was defined as all employed and unemployed persons younger than the assumed retirement age of 65 years. Unemployment was defined as persons who were without a job, but were available to the labour market and were actively seeking work⁴⁹. When calculating the lost value for society, the average income per month for the labour force was applied. The wages for 2016 were calculated as a weighted average corresponding to the shares of males and females in the fictional cohort with Ontario-specific general labour rates from Canadian Statistics⁵⁰. Costs attributable to sick leave were calculated as the number of sick days that the UC users were absent from work because of catheter-related complications. Presumably, UC and HCIC users took an equal number of sick days when they experienced a catheter-related complication. The societal benefit in terms of fewer sick days because of using HCICs rather than UCs was expressed by the reduced number of complications.

Because of limited data, production loss due to sick leaves and catheter-associated early retirement was only assessed when it occurred in one of the UTI-related health states or because of treatment related adverse events (TRAEs). Production loss attributable to sick leaves and early retirement caused by renal impairment were not included. Production loss due to death because of all complications associated with catheter use, here under renal impairment, was included.

Estimates of quality-of-life

The baseline utility decrement for the IC dependent SCI population compared with the general population was obtained from a study conducted in the UK wherein the SF-36 questionnaire was used⁵¹. This estimate was mapped to EQ-5D using a mapping algorithm⁵². Mapping of the SF-36 data yielded a utility decrement of 0.45.

During each cycle in the Markov model, the cohort would accrue a utility benefit/decrement corresponding to their current health states and possible adverse events. These utility decrements were derived from different sources and were applied to Canadian age- and gender-adjusted EQ-5D population norms in the Markov model⁵³.

A utility benefit attributable to the use of HCICs instead of UCs was added to the baseline utility value for catheterization. The utility benefit associated with HCIC use was related to fewer steps/less time required for the catheterization process, as well as less pain and fewer UTIs. The utility value was elicited from a time trade-off (TTO) study performed among the general Canadian population⁵⁴.

A scenario analysis was also performed wherein the utility gains associated with compact and phthalate-free catheter use were added to the HCIC utility. These two utility gains were also elicited from a TTO study, but this was conducted among the general population in the UK⁵⁴.

The utility values used in the model are listed in Table 1.

Model output

The outputs of the applied Markov model were incremental costs, quality-adjusted life years (QALYs), life years gained (LYG), and number of UTIs avoided. An annual discount rate of 1.5% was applied to costs, QALYs, and LYG, as per CADTH guidelines⁵⁰, and results were reported as incremental cost-effectiveness ratios (ICERs). To investigate the impact of variation in parameters on the ICERs, a deterministic one-way sensitivity analysis was performed. To further assess the robustness of the model, a probabilistic sensitivity analysis was performed using 1,000 Monte Carlo simulations.

Results

According to the Ontario Ministry of Health and Long-Term Care, a 50-year-old SCI patient with NB would live for an average of 12.36 additional years when using UCs and 13.14 additional years when using HCICs. Patients with NB gained 6.09 QALYs using HCICs compared with 5.37 QALYs using UCs, while accruing a cost of $120,639 and $72,622, respectively. Hence, the probabilistic incremental cost per QALY gained was found to be $66,634 for a lifetime horizon in the base case. The model further predicted that the estimated number of UTIs in an IC-dependent SCI patient would decrease by 11% over the patient’s lifetime with the use of HCICs compared with the use of UCs. The results are summarized in Table 3.

Table 3. Probabilistic cost-effectiveness results.

	Costs	QALYs	LYG	UTI events
Base case
Uncoated catheters	$72,622	5.37	12.36	37
Hydrophilic-coated intermittent catheters	$120,639	6.09	13.14	33
Incremental findings	$48,016	0.72	0.79	4
ICER		$66,634.33	$61,037.02	$13,088.91
Scenario including societal costs
Uncoated catheters	$1,884,681	5.37	12.36	37
Hydrophilic-coated intermittent catheters	$1,730,221	6.09	13.14	33
Incremental findings	–$154,460	0.72	0.79	4
ICER		Dominant	Dominant	Dominant
Scenario including utility gain from compact catheters
Uncoated catheters	$72,622	5.37	12.36	37
Hydrophilic-coated intermittent catheters	$120,639	6.49	13.14	33
Incremental findings	$48,016	1.12	0.79	4
ICER		$42,821.05	$61,037.02	$13,088.91
Scenario including utility gain from phthalate-free catheters
Uncoated catheters	$72,622	5.37	12.36	37
Hydrophilic-coated intermittent catheters	$120,639	6.58	13.14	33
Incremental findings	$48,016	1.21	0.79	4
ICER		$39,852.00	$61,037.02	$13,088.91
Scenario including utility gain from compact catheters and phthalate-free catheters
Uncoated catheters	$72,622	5.37	12.36	37
Hydrophilic-coated catheters	$120,639	6.98	13.14	33
Incremental findings	$48,016	1.60	0.79	4
ICER		$29,942.04	$61,037.02	$13,088.91

The deterministic sensitivity analyses demonstrated that the additional utility benefit of HCICs over UCs has a large impact on the results. If the utility benefit was assumed to be 0/0.05, the ICER range becomes $132,485–$46,925. The costs per catheter also had a substantial impact; if the cost per HCIC varied between $3.00–$4.50, the ICER range was $44,851–$85,178, and if the cost per UC was assumed to be $0.50 or $2.00, the ICERs became $79,987–$42,000. The relative treatment effect of HCICs vs that of UCs regarding UTI incidence was tested for the values of 0.79 and 0.90, which gave an ICER range of $57,181–$78,812. Considering that a 14% or 41% monthly risk of UTIs yielded ICERs of $71,414 and $56,736, respectively, the baseline risk of UTIs (UC) also had a sizable impact. Mortality multipliers had a minor impact on the results; testing mortality multipliers values of 1 and 2,000 for UTIs unresponsive to initial treatment revealed ICERs of $66,508–$63,073, and testing multiplier values of 50 and 250 for antibiotic resistance revealed a range of $73,846–$60,793. The remaining parameters analyzed in the sensitivity analysis were the age at onset of SCI (40 and 60 years) and discount rates (0% and 3%), which provided ICERs between $58,645 and $72,022. The results of the deterministic sensitivity analysis are shown in Table 4 and in the tornado diagram in Supplementary Figure S1.

Table 4. Deterministic sensitivity analysis.

Parameter	Base case (Tables 3 and 4)	Alternatives	ICER	Reference
Base case			$66,634
Catheters per day	4	3	$41,077	Assumption
		5	$73,094
Cost per catheter (UC)	$1.07	$0.50	$79,987	Assumption
		$2.00	$42,000
Cost per catheter (HCIC)	$3.77	$3.00	$44,851	Assumption
		$4.50	$85,178
Baseline risk of UTI/month (UC)	22%	14%	$71,414	Cardenas and Hoffman^69: Community period.
		41%	$56,736	Cardenas and Hoffman^69, Cardenas et al.^15, DeRidder et al.^22: Complete treatment period
Treatment effect (HCICs vs UCs): UTI responsive to initial treatment	0.84	0.79 (21%)	$57,181	Cardenas and Hoffman^69: Hospital period.
		0.90 (10%)	$78,812	Cardenas and Hoffman^69; Cardenas et al.^15, DeRidder et al.^22: Complete treatment period
Utility benefit from using HCICs instead of UCs	0.028	0	$132,485	Assumption
		0.05	$46,925
Age at onset	50	40	$72,079	Assumption
		60	$58,645
Discount rate (costs and benefits)	1.50%	0%	$62,752	CADTH Guideline^50
		3%	$68,252
Mortality multiplier: UTI unresponsive to initial treatment	797.6	1	$66,508	Assumption
		2,000	$63,073
Mortality multiplier: antibiotic resistant	140.270	50	$73,846	Assumption
		250	$60,793

The secondary analysis that included the societal costs revealed that the use of HCICs dominates that of UCs, implying that HCICs cost less than UCs, while also providing superior outcomes in terms of QALYs, LYG, and UTIs.

The secondary scenario analysis that included the utility gains associated with compact catheter use revealed an incremental cost per gained QALY of $42,821. The secondary scenario analysis that included the utility gains from phthalate-free catheter use revealed an ICER of $39,852. The scenario analysis that included the utility gains associated with both compact and phthalate-free catheter use revealed an ICER of $29,942.

The cost-effectiveness acceptability curve (Supplementary Figure S2) and the cost-effectiveness plane (Figure 2), based on the probabilistic sensitivity analysis (PSA) with 1,000 simulations, demonstrated that, when the societal willingness-to-pay was $66,000 or above per QALY, the probability that HCICs would be cost-effective was >50%.

Figure 2. Cost-effectiveness plane.

Discussion

The current study modified a previously published Markov model to analyze the cost-effectiveness of HCICs in a Canadian setting; the base case analysis identified an ICER of $66,462/QALY by using HCICs compared with using UCs. From a societal perspective, HCICs were found to be a dominant treatment strategy as compared with UCs.

A potential weakness of the present study was the limited availability of data on baseline risk, preventive effects of HCICs, and mortality and consumption data regarding IC use in Canada. To overcome this problem, it was necessary to include data from other countries and to make several assumptions, which created some uncertainty regarding the results. Evidence regarding the impact of HCICs use on UTI incidence was particularly limited, which has been the subject of several RCTs, but with conflicting results. All the main RCTs had significant methodological limitations, most notably in terms of insufficient power, attrition, and how they defined and measured UTIs. These limitations can lead to imprecise estimates of the relative treatment effect, and the current results have to be interpreted accordingly. Similar to the results presented in the original study by Clark et al.³⁰, this study detected a rather low average additional life expectancy of 12.75 years in the cohort with a starting age of 50 years. This relatively low average life expectancy could be due to the limited and relatively old evidence of mortality in persons with SCI⁵⁶, which is another limitation of this study.

Considering the aforementioned limitations, the current results appear to be rather robust, because the PSA demonstrated that all ICERs were in the northeast quadrant of the cost-effectiveness plane. To our knowledge, this is the first CEA of IC to include a societal perspective and the catheter-related impact on health-related quality-of-life (HRQoL), thus providing a more comprehensive picture of the benefits associated with IC for the users.

Previous CEAs have relied on utility values from generic measures, such as EQ-5D and SF-36, to account for the HRQoL in different health states associated with catheter use. Generic tools are useful because they provide consistency and allow for comparisons between the treatment areas. However, the generic instruments have also been found to be insensitive to small but, over time, substantial improvements to quality-of-life, as well as to certain categories of benefit, e.g., convenience and mode of administration^57–59. It has, therefore, been argued that it is impossible to assess the cost effectiveness of specific improvements in catheter design by generic HRQoL measures⁶⁰. For example, although the HRQoL is directly affected by CAUTIs, it is also affected by indirect health-related factors associated with catheter use, such as pain and discomfort during catheterization, worrying about future UTIs, time spent on catheterization, and discreteness of the catheter design⁶⁰. Thus, these factors affecting HRQoL should also be considered when comparing the cost-effectiveness of different catheter types. Recently, the inclusion of patient acceptability/satisfaction with the procedure as well as products in CEA has been listed as a high priority in future research on IC by the International Consultation on Incontinence⁶¹. Hence, utility gains associated with the use of HCICs and catheters with a compact design, as well as catheter materials without phthalates, were included in this study. Deterministic sensitivity analysis demonstrated that these utility benefits have a significant impact on the cost-effectiveness results, and future studies should further investigate the impact of these catheter innovations.

Several CEAs from other countries have concluded that HCICs are cost-effective in their respective settings^27,30–33. Clark et al.³⁰ predicted an incremental 1.4 LYG with HCIC use with an additional cost of £2,100 as compared with UC use in the UK. Furthermore, Truzzi et al.²⁷ determined an ICER of 122,330 BRL per QALY and an additional cost of 31,240 BRL with HCIC use in a Brazilian setting. Moreover, with an ICER of €24,405, HCICs were also estimated to be cost-effective in an Italian healthcare perspective as compared with UCs. Finally, from the Japanese payer perspective, HCICs were considered highly cost-effective, with an ICER of 3.8 million yen, compared with UCs.

Because there is no official willingness-to-pay threshold value in Ontario, it cannot be concluded with certainty whether the base case results are perceived as cost-effective. Laupacis et al.⁶² suggested a threshold range between $20,000–$100,000 per QALY in Canada; however, they also stressed that the values were “arbitrary” and that other considerations, e.g. ethical and political, also play a role in the implementation of new technologies. A threshold of $50,000–$100,000 per QALY gained is often cited in the health economic literature⁶³. The base case result of $66,462 falls within both cited ranges, indicating that HCICs may be considered a cost-effective treatment strategy in Canada. A threshold value of $100,000 is depicted in Figure 2 for illustrative purposes.

Future research could investigate the broader societal impact of different types of catheters in countries other than Canada. Furthermore, additional research is warranted to investigate the impact of different bladder management techniques and devices on the HRQoL. Finally, given the lack of strong evidence regarding the various consequences of performing IC, additional RCTs or “real world evidence” with a sufficient number of participants and an adequate timespan are warranted.

Conclusion

We investigated the cost-effectiveness of HCICs vs UCs from a public payer perspective and, in a secondary scenario, from a societal perspective. We found that, over a lifetime horizon, using HCICs compared with using UCs yields an ICER of $66,462 from the public payer perspective. When the HRQoL-gains associated with compact HCICs, or catheters not containing phthalates, were included, the cost per QALY decreased and made HCICs more economically attractive compared with UCs. From a societal perspective, providing reimbursement for HCIC use could lead to potential cost savings as compared with UC use.

Transparency

Declaration of funding

LHT was employed by Coloplast A/S while modifying the existing Markov model to a Canadian setting. Editorial services from Marksman Healthcare Communications, and a critical scientific review by Rebecca Hancock-Howard (PhD), from Amaris, were additionally supported by funding obtained from Coloplast A/S.

Declaration of financial/other relationships

LHT performed part of the modification of the existing Markov model as part of her master’s thesis in Economics at the University of Copenhagen. AK provides occasional medical expert input for advisory boards at Coloplast A/S. A JME peer reviewer on this manuscript declares that they are the lead author of two papers referenced by this paper and received fees from Wellspect Heathcare for workshop participation. A second JME peer reviewer on this manuscript declares that they are a clinical investigator for a NIHR funded program grant, including a trial of single use vs multi/single use catheters (MultiCath trial), and that they are a co-author of a withdrawn Cochrane review of intermittent catheters with a new review in progress. All peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose than those outline here.

Acknowledgments

We would like to thank Marksman Healthcare Communications for the editorial assistance and Rebecca Hancock-Howard (PhD) from Amaris for critically reviewing the manuscript. Jeppe Sørensen from Coloplast A/S contributed to the interpretation of results and conducted a scientific review of the manuscript.