Home > Clinical Guides > Infection: Prevention and Control of... > Cost-utility analysis: Intermittent self...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Clinical Guideline Centre (UK). Infection: Prevention and Control of Healthcare-Associated Infections in Primary and Community Care: Partial Update of NICE Clinical Guideline 2. London: Royal College of Physicians (UK); 2012 Mar. (NICE Clinical Guidelines, No. 139.)

Appendix JCost-utility analysis: Intermittent self catheterisation

J.1. Introduction

Catheter-associated urinary tract infection (CAUTI) is the most common healthcare acquired infection in the world, accounting for 20% to 45% of all nosocomial infections 376. While most urinary tract infections (UTIs) are mild and easily resolved with appropriate antibiotic treatment, more severe infections can be devastating, resulting in bacteraemia, sepsis and death. Due to the frequency with which they occur, they also impose a substantial economic burden on the NHS 377.

The most important risk factor for the development of CAUTI is the prolonged use of an indwelling catheter. For this reason, intermittent self catheterisation (ISC) has become the preferred method of catheterisation for patients in which it is clinically indicated 369 247. ISC aims to reduce CAUTIs and promote greater independence among people who have bladder emptying problems. Nevertheless, CAUTI remains the most frequent and serious complication of ISC 515.

There are several different approaches to ISC. Patients may use disposable catheters with a hydrophilic polymer surface coating, disposable catheters with pre-packaged water based lubricant (gel reservoir), or non-coated catheters. Non-coated catheters may be discarded after use, or washed and re-used for up to one week. Which material and method constitutes the best approach is an issue of considerable uncertainty.

Our aim in constructing the model was to determine the most cost-effective type of catheter for patients performing ISC in the community. The relative effectiveness of each type of intermittent catheter was based on the results of the randomised controlled trials included in our systematic review. Several different versions of the model were built to reflect the diversity of patient groups using ISC. The model was built probabilistically in order to take into account uncertainty and imprecision around parameter point estimates.

J.2. Methods

J.2.1. Model overview

J.2.1.1. Comparators

There are several types of catheters available for ISC. The catheters included in the model are all those that are available for patients residing in the community:

  • Hydrophilic catheters are coated with a hydrophilic polymer coating. Hydrophilic catheters must be immersed in water prior to use or may be packaged in a casing of water or saline. These catheters are designed for single use.
  • Gel reservoir catheters are pre-packaged with a small sachet of sterile water-soluble lubricant which must be released and spread over the catheter before use. These catheters are also designed for single use.
  • Non-coated catheters do not have a surface coating and patients often apply a water-based or anaesthetic lubricant before use. These catheters may be washed and reused for up to one week, although some patients choose to use them as single use catheters. In the model we chose to explore both methods of non-coated catheter use:
    • Non-coated catheters which are discarded immediately after use – sometimes referred to as ‘sterile’ non-coated ISC.
    • Non-coated catheters which are washed, dried and reused multiple times – sometimes referred to as ‘clean’ non-coated ISC.

The decision to include multiple use non-coated ISC as a treatment alternative was made in consultation with the GDG, expert continence advisor, NICE commissioning managers, Medicines and Healthcare products Regulatory Agency, British Association of Urological Nurses, staff at Stoke Mandeville Hospital, and the manufacturers of each non-coated catheter listed on the Drug Tariff (Bard, Teleflex Medical, Pennine Healthcare, and Hunter Urology). The conclusion from these conversations was that in the community, clean ISC remains a valid method of catheterisation. However, in settings where facilities are not available, patients are catheterised by others, or patients are below 16 years of age (see below), re-use is not advisable. Therefore, two sets of models were built:

  • One for when clean ISC is an option, and
  • One for when it is not.

J.2.1.2. Population

There are multiple causes of bladder dysfunction which affect a heterogeneous population. ISC may be used by patients with neurogenic bladder, dysfunctional voiding syndromes, and patients recovering post-operatively for procedures to the urinary tract or reproductive system 219.

Because the majority of the included clinical effectiveness studies were conducted in patients with spinal cord injury (SCI), the base case model considered a population of adult patients with neurogenic bladder due to SCI.

In order to create a model that would be broadly applicable to all individuals using ISC in the community, separate cost-utility analyses were conducted for adult patients with bladder dysfunction caused by a condition other than SCI as part of the sensitivity analysis.

The GDG noted that in children and young people (≤ 16 years old), symptomatic UTI can cause progressive renal scarring which may lead to renal failure later in life. Renal failure carries a high risk of mortality and morbidity, is associated with very high cost and decreased quality of life. The most recent NICE guideline for Urinary Tract Infection in Children 314 concluded that it was not possible to estimate the true risk of renal failure as a result of childhood UTI, did not identify any quality of life values for children with UTI, and did not consider economic modelling a valid option in this population. The current GDG agreed with this decision and noted that none of the studies included in the clinical review which contained symptomatic UTI as an outcome were conducted in children. Given the uncertain risk of harm as a result of symptomatic UTI in childhood, the GDG decided to employ the precautionary principle in their approach to ISC in children. Therefore, only single use catheters were considered an option for ISC in children and modelling was not explicitly undertaken in this population.

J.2.1.3. Time horizon, perspective, discount rates used

The analysis was undertaken from the perspective of the NHS and personal social services, in accordance with NICE guidelines methodology 315. Relevant costs consisted of the cost of catheters (and lubricant, where applicable) and treatment for UTIs of varying severity at the primary and secondary care level. All costs are reported in 2009/10 British pounds. The primary measure of outcome is the quality-adjusted life-year (QALY). The model was evaluated over a lifetime horizon with both costs and QALYs discounted at a rate of 3.5% per year.

J.2.2. Approach to modelling

Symptomatic UTI is the most meaningful outcome for evaluating the efficacy and costs of intermittent catheterisation. Although asymptomatic bacteriuria is common in patients using ISC over the long term, it has little clinical impact and treatment is not recommended 491. As in the clinical review, symptomatic UTI was defined one or more symptom suggestive of UTI and/or self-reported UTI requiring treatment.

Current management of symptomatic UTI usually includes a clinical assessment of symptoms and dipstick urinalysis, followed by empiric treatment (referred to as ‘first-line antibiotic’ treatment throughout the model). The most clinically relevant outcome following treatment is the resolution of symptoms. In the model, this state is referred to as ‘clinical cure’.

Although empiric treatment is effective in the majority of cases, a small proportion of these patients will experience persistent symptomatic infection and contact their healthcare provider for further treatment. ‘First-line antibiotic resistant UTI’ was used to describe patients with symptomatic relapse who require a further antibiotic prescription within 28 days of the initial prescription. Because antibiotic resistance is a key cause of treatment failure, at this point in the treatment pathway the healthcare provider will normally obtain a urine specimen and initiate targeted treatment based on the results of the culture.

UTIs may be caused by a number of different strains of bacteria. Over the past several years, antimicrobial resistant strains have emerged as important causes of UTI in the UK and around the world 370,398. In order to accurately capture the full impact of UTI on patient morbidity, mortality and cost, the GDG considered it important to incorporate the effects of antibiotic resistance into the model.

‘Multidrug resistant UTI’ was defined as resistance to two or more classes of antimicrobial agents. It was assumed that all patients with a multidrug resistant infection are admitted to hospital for treatment with intravenously administered carbapenem antibiotics. Catheter-associated bacteraemia occurs when a patient’s blood and urine cultures reveal growth of the same organism. All patients with catheter-associated bacteraemia were assumed to have symptomatic UTI and it was assumed that they were immediately admitted to hospital upon diagnosis.

Long-term studies have demonstrated that the incidence of urethral complications such as structures and false passages tend to increase over time 515. Although proponents of hydrophilic catheters often cite the lower surface friction associated with their coating detected by cytological investigation 475 449 as evidence of a reduction in urethral complications, no comparative clinical studies have been published. Therefore, in the base case analysis it was assumed that the incidence of urethral complications does not vary between the different catheter types. The model was built to allow exploration of this assumption in sensitivity analysis.

J.2.2.1. Key assumptions

The main simplifying assumption of the model is that the probability of antibiotic resistance does not change over time. The decision to build a static model was based on a lack of available data about current and historical resistance rates, the complexity of forecasting antibiotic resistance trends over time and within populations, and a lack of examples on which to base methodological approaches. The GDG deemed the assumption of a static model to be reasonable and the impact of extreme scenarios was explored in sensitivity analysis.

J.2.2.2. Model structure

A Markov model was constructed to calculate lifetime costs and QALYs for each comparator. Figure 97 illustrates the key health states in the model and possible transitions between them in each cycle. The model is divided into one year cycles, which was thought to be a reasonable cycle length based on available evidence of clinical efficacy and baseline risk. The model was built in TreeAge Pro 2009.

Figure 97. Markov model structure.

Figure 97

Markov model structure. Schematic diagram of the Markov model designed to analyse the cost-effectiveness of different types of intermittent catheter. The Markov modelling approach involves a transition between different health states over time. The model (more...)

The hypothetical SCI population entering the model had an average age of 40 years and was 80% male; this is the average age at injury and gender ratio of spinal cord injury patients according to the US National SCI Database 318.

The model structure did not explicitly account for patients who experience more than one UTI within a one year cycle length. Because the data used to inform the clinical effectiveness for each type of catheter measured the occurrence of ‘one or more UTI’, it was assumed that recurrent infections were implicitly included in the baseline and relative risk estimates. In the absene of more specific randomised evidence of comparative efficacy for recurrent UTI, this was a necessary assumption.

In addition, the analysis also did not explicitly model the transition from first-line or multidrug resistant UTI to bacteraemia. Again, this structural assumption was necessary due to data limitations. A search of the literature only identified the probability of developing bacteraemia after symptomatic UTI of non-specific severity. It was therefore assumed that this value represents the cumulative probability of bacteraemia as a result of all UTI and was only included once in the model.

J.2.2.3. Uncertainty

The model was built probabilistically to take account of the uncertainty surrounding each input parameter. In order to characterise uncertainty, a probability distribution was defined for each parameter based on error estimates from the data sources (e.g. standard errors or confidence intervals). When the model was run, a value for each input was randomly selected from its respective distribution. The model was run repeatedly to obtain mean cost and QALY values.

The number of simulations used to obtain the probabilistic results was chosen according to methods described by Koehler and colleagues 228. The model was set to ensure that the Monte Carlo error was not more than 1% of the standard error of the mean incremental cost and QALY estimate for each type of catheter. For this model, the number of simulations necessary to obtain this level of accuracy is approximately 10, 000.

Various sensitivity analyses were also undertaken to test the robustness of model assumptions and data sources. In these analyses, one or more inputs were changed and the analysis was rerun in order to evaluate the impact of these changes on the results of the model.

J.2.3. Model inputs

J.2.3.1. Summary table of model inputs

The probability of acquiring a CAUTI was based on clinical evidence identified in the systematic review undertaken for the guideline. All other model inputs were identified by supplementary literature reviews and were validated with members of the GDG. A summary of the probability, cost, and utility inputs used in the base-case analysis is provided in the tables below. More details about sources, calculations and rationale underpinning data selection can be found in the section preceding each summary table.

J.2.3.2. Baseline event rates

Symptomatic UTI

The baseline probability of developing symptomatic UTI was calculated from the studies included in the clinical review (Table 23)59,95,110,150,224. The annual rate was obtained by dividing the total number of events observed in patients using single use non-coated catheters by the total number of patient years (Equation 1).

Table 23. Baseline risk of symptomatic UTI in patients with SCI using single use non-coated catheters.

Table 23

Baseline risk of symptomatic UTI in patients with SCI using single use non-coated catheters.

Equation 1. Rate

A standard error for the rate was derived using the delta method as described by Kirkwood and Sterne 2003 (Equation 2)226.

Equation 2. Standard error of the rate

For the purpose of clinical validation, the 95% confidence interval for the rate was derived from the standard error. In order to take account of the constraint that the rate must be greater than or equal to zero, it is preferable to work on the log scale and to derive a confidence interval for the log rate, then calculate the exponential to give a confidence interval for a rate 226. The formula for the standard error of the log rate is derived using the delta method (Equation 3)226.

Equation 3. Confidence interval for a rate
95%CI(rate)=exp(Log(rate)±1.96×1d)95%CI(rateforoneormoreUTIsperyear)=exp (Log(1.14)±1.96×1115)=0.933to1.347

A gamma distribution was applied to the rate according to the method of moments approach described by Briggs et al 2006 (Equation 4)51.

Equation 4. Gamma distribution (α, β)

In order to transform the baseline rate of symptomatic UTI to a probability the following equation was used (Equation 5) 133.

Equation 5. Converting a rate to a probability

Therefore, based on the rate of symptomatic UTI observed in the included studies, the baseline probability of symptomatic UTI associated with sterile non-coated catheter use was 68% (Table 24). This is consistent with other epidemiological and observational studies in the literature 515,518.

Table 24. Baseline event rate and relative treatment effects.

Table 24

Baseline event rate and relative treatment effects.

Urethral complications

In the base case analysis, the baseline probability of developing a urethral complication was derived from an observational study of patients using ISC over an average length of 9.5 years 359. Over this time, 19% of this group developed urethral strictures. According to the equations described above, this results in a 2.38% annual probability of developing a urethral complication (Table 24).

This value is on the upper end of estimates reported by other papers 515. It was chosen to represent the possibility of developing urethral complications of any type, whether they are strictures, false passages, urethritis, or any other complication that could be expected as a result of urethral trauma.

J.2.3.3. Relative treatment effects

Symptomatic UTI

The between-strategy differences in costs and QALYs are driven by the relative risk (RR) of symptomatic UTI for each catheter compared to single use non-coated catheters. The RR for each catheter is based on the results of the systematic review and meta-analysis of randomised controlled trials identified in the clinical review (see section I.4), where single use non-coated catheters were used as the baseline comparator.

The probability of symptomatic UTI associated with each catheter strategy was calculated by multiplying the baseline risk of symptomatic UTI by the RR of symptomatic UTI for each catheter. The results of the meta-analysis and the distribution assigned to each RR are reported below in Table 24.

Urethral complications

In the absence of any comparative clinical evidence, it was assumed that the risk of developing urethral complications did not differ between catheters. This assumption was explored in sensitivity analysis.

J.2.3.4. Cohort probabilities

Antibiotic resistance in UTI

Despite the clinical and political importance of antimicrobial resistant infections, evidence of the prevalence of resistant infections in the urinary tract is scarce. Only one paper which examined the incidence of first-line antibiotic treatment failure among patients with SCI who use ISC was identified 107. In this Canadian study, patients were randomised to receive either a 3-day or 14-day course of ciprofloxacin. At 23-day follow-up, symptomatic relapse was experienced by 5 out of 30 patients in the 3-day treatment group 107. The probability of clinical failure after treatment for symptomatic UTI was therefore 15.4%.

Among individuals with SCI, it is thought that prolonged, repeated exposure to healthcare settings and antimicrobial agents increases the risk of infection with multidrug resistant organisms. The most common mechanism of resistance in UTI-causing organisms is the production of extended-spectrum beta-lactamases (ESBL). These enzymes inactivate certain antibiotics. Like all forms of antimicrobial resistance, the prevalence of ESBL varies by geography, healthcare setting, and patient demographic. Recent studies have found that the annual probability of multidrug resistant UTI observed in the SCI population ranges from 4.3% in community dwelling persons using ISC 487 to 9% acute rehabilitation settings 313. Based on these estimates, it was assumed that on average, 7% of individuals with catheter-associated UTI are infected with a multidrug resistant pathogen (Table 25); this assumption was further explored in sensitivity analysis.

Table 25. Overview of baseline probabilities and probability distributions.

Table 25

Overview of baseline probabilities and probability distributions.

If on average, 15.4% of patients with SCI who use ISC experience treatment failure for symptomatic UTI and 7% of SCI patients using ISC fail treatment by virtue of having multidrug resistant UTI, it was assumed that the remaining patients experience treatment failure due to first-line antibiotic resistant infections.

Mortality due to multidrug resistant UTI

Patients infected with ESBL-producing bacteria are generally sicker than patients who are not infected with ESBL producing strains. However, there are very few studies of mortality in patients with multi-drug resistant UTI. Even among the few studies that addressed the issue in patients with bacteraemia, the question of whether ESBL-production significantly increases the risk of death remains unclear 391.

A retrospective analysis 201 of ESBL-producing bacteria found an overall mortality rate of 12.1% among patients with UTI caused by ESBL-producing E.Coli and Klebsiella bacteria. However, there was no control group for this population and it was not clear whether the analysis controlled for the contribution of antibiotic resistance to the reported mortality rates. A recent retrospective study by Klevens et al (2008)227 determined that 8 out of a total of 43 deaths in patients with UTI caused by ciprofloxacin resistant E.Coli were directly caused or contributed to by the resistant organism. Out of a total of 3112 ciprofloxacin resistant isolates collected from 2000 to 2004, 9.8% were UTIs caused by ciprofloxacin-resistant E.coli. Therefore, the mortality rate in patients with UTIs caused by drug-resistant bacteria was 2.6%. The GDG thought this to be a reasonable estimate of mortality to include in the base case analysis.


In order to estimate the incidence of bacteraemia following UTI, we looked primarily to the economic evaluations retrieved by our systematic reviews and completed a search of PubMED to identify other data. In 2000 Saint et al418 published a systematic review of the incidence of bacteraemia in patients with UTI; this was the most recent and comprehensive source of data identied to inform this parameter. Each of the five studies included in this review reported similar estimates ranging from 2.6% to 4.0%. The pooled estimate for the risk of developing bacteraemia as a result of catheter-associated UTI was 3.6% with a 95% CI of 3.4% to 3.8%418. The studies included in this review were from a heterogeneous hospital-based population. In the absence of any specific data regarding individuals with SCI, the same probability was assumed to apply to both the SCI and non-SCI population (Table 25).

Mortality due to bacteraemia

There have been few studies of bacteraemia in patients with SCI. Two retrospective analyses of deaths occurring within 30 days of diagnosis of bacteraemia in patients with SCI were identified 303 488. The study by Montgomerie and colleagues (1991) reported 4 deaths in 50 bacteraemic episodes were directly related to bacteraemia with a UTI origin (probability of 7.7%), while Wall et al (2003) report a total of 8 deaths in 95 bacteraemic episodes (probability of 8.1%). The former was used to inform the base case analysis as this rate was derived from patients with UTI-associated bacteraemia only. The slightly lower probability of mortality in these patients compared to non-SCI individuals (Table 29) appears to be a well-recognised phenomenon in the literature 303.

Table 29. Summary of probability and utility values for people without SCI.

Table 29

Summary of probability and utility values for people without SCI.

J.2.3.5. Life expectancy

Although there have been dramatic improvements in the care of patients with spinal cord injuries over the past 50 years, life expectancy remains slightly below normal. Mortality rates are significantly higher during the first year after injury than during subsequent years, particularly for more severely injured individuals. For the purposes of this analysis, it was assumed that patients using IC in the community had survived beyond the one-year time point.

To date, there is only one study of mortality among spinal cord injury patients in Britain. Frankel et al (1998) 135 conducted a review of medical records from patients with spinal cord injury of at least one year duration at Stoke Mandeville hospital and the Regional Spinal Injuries Centre in order to calculate standardised mortality ratios (SMRs) for subjects injured between 1973 and 1990. The gender distribution of this cohort (81% male) closely matched that of our baseline demographic and the analysis combined mortality ratios for all levels of disability. Age-dependant annual mortality rates were calculated by multiplying the SMR of 5.41 for patients aged 31–41 at time of injury by central mortality rates obtained from life tables for England and Wales in 2007–2009 337.

J.2.3.6. Utilities

In accordance with the NICE reference case, health outcomes were estimated using the Quality Adjusted Life Year (QALY). In order to calculate QALYs, it is necessary to quantify both the quality of life of each health state and the time spent in each state. A systematic literature search was performed in order to identify all health related quality of life studies related to UTI and UTI-associated bacteraemia. The results of this review are reported in Appendix K.

The literature search revealed two recent studies which measured the impact of UTI in people with SCI using a validated generic measure of health-related quality of life 174,257,483. The authors of these studies were contacted for additional information and both replied. Although Haran and co-workers were unable to provide any further data, Vogel and colleagues granted us access to recent patient-level SF-12 responses collected as part of a longitudinal study of adults who sustained SCI as children and adolescents 483,529. The responses were classified into three groups according to our outcome of interest: no UTI, UTI and severe UTI (requiring intravenous antibiotics or hospitalisation). The recall period for each group was one year (i.e. patients were asked to describe their health over the past year). Using an algorithm developed by Gray et al 2006 164, this data was mapped to EQ-5D values for the UK population. Because of the random component contained within this mapping algorithm, a simulation was run 1000 times in order to calculate a mean value, standard error and confidence interval for each of the three health states measured (Table 26).

Table 26. Health state utility weights for people with SCI.

Table 26

Health state utility weights for people with SCI.

In order to calculate a utility value for first-line resistant UTI, it was assumed that the quality of life associated with this health state is worse than that for UTI but better than that for multidrug resistant UTI. The mean value of these two health states was taken and the standard error was assumed to be 5% of the mean in order to generate the probability distribution (Table 26).

In the absence of published utility values for UTI-associated bacteraemia, it was assumed that a linear decrease in health-related quality of life applies to those in this health state and that the standard error was 5% of the mean. The implications of this assumption were explored in sensitivity analysis.

The values calculated from the studies by Zebracki et al 2010529 and Vogel et al 2002483 were chosen to inform the base case analysis as they better account for the range of health states within the model and were elicited with a recall period that more accurately matches the model cycle length than the data reported by Lee and Harran 174,257.

A recent Cochrane review of procedures for urethral narrowing did not find any quality of life data among patients treated for urethral strictures 512. A search of the Tufts cost-effectiveness analysis registry 3 also failed to identify any relevant utility weights in the literature. Given that urethral complications would likely involve significant discomfort and stay in hospital, it was assumed that the quality of life associated with this health state would be comparable to that experienced by patients with multidrug resistant UTI.

J.2.3.7. Resource use and cost

Cost of catheters

All catheters available through the NHS Drug Tariff323 were classified as either hydrophilic, gel reservoir or non-coated with the help of the continence expert and manufacturer information provided on-line. In cases where there was uncertainty about catheter type, manufacturers were contacted by telephone. The average cost of each type of catheter was used as the point estimate; the maximum and minimum listed costs formed the range used to inform each distribution (Table 27).

Table 27. Catheter unit costs and annual resource use.

Table 27

Catheter unit costs and annual resource use.

Most individuals using ISC catheterise between four and six times a day regardless of the type of catheter they use 513. In order to calculate the annual cost of gel reservoir, hydrophilic and single-use non-coated catheters, it was assumed that patients catheterise an average of 5 times per day. Depending on personal habits and preferences, individuals using non-coated catheters multiple times use a highly variable number of catheters per month. To ensure consistency with prescribing data from the NHS Drug Tariff323 and the literature478, an average of 5 catheters per month (ranging from 4 to 6 per month) was used to calculate the annual cost of non-coated catheters used multiple times in the base case analysis. This was varied in sensitivity analysis.

Non-coated catheters require an application of lubricant before use. Although most patients use a water-based lubricant, the GDG estimated that an average of five percent of patients who self catheterise regularly use lidocaine lubricant. This estimate was probabilistically incorporated in the cost of lubricant by assuming a range of between 0% and 10%. Because lubricant is applied to the catheter each time it is used, it was assumed that patients with single use and multiple use non-coated catheters consume equal amounts of lubricant.

In order to accurately capture the cost of catheter use in the community, a monthly prescription dispensing fee was added to the cost of catheters and lubricant (i.e. one prescription charge per month for gel reservoir and hydrophilic catheters and a total of two prescription charges per month for noncoated catheters). The range used to inform this distribution was based on the highest and lowest dispensing fee scales for authorised dispensing practitioners.

Cost of treatment for infection

CAUTI treatment costs were estimated based on recommended diagnostic and treatment pathways for UTI in adults 474 181. Costs regarding contact time with primary healthcare workers were obtained from the 2009/10 Personal and Social Services Research Unit 88 Costs incurred in the community were based on data from the 2010 NHS Drug Tariff 323. The cost of secondary care was calculated according to 2009/10 NHS Reference costs. A detailed breakdown of the cost of treating catheter-related infections is presented in Table 28.

Table 28. Cost of treatment.

Table 28

Cost of treatment.

Please note the following for costing purposes:

  • Patients may consult a number of different healthcare professionals for treatment of UTI. It was assumed that the healthcare provider most frequently contacted for UTI was a GP (in 80% of cases), followed by community nurse specialist (in 10% of cases) and hospital emergency room (in 10% of cases) 513. The cost of GP consultations and community nurse specialist were obtained from the Unit Costs of Health and Social Care 2009/10 88, the cost of emergency room visit was obtained from the NHS reference costs 2009/10100. These costs were incorporated into the model probabilistically according to the following distributions:
    • The average cost of a GP consultation was estimated at £30, based on a 12.6 minute surgery consultation with upper and lower confidence intervals based on the mean cost of home visit (£60) and 10 minute surgery consultation (£23) used to inform the distribution parameters (α = 100.0000, β = 0.3000)
    • The cost of a 20 minute home visit from a community nurse specialist (£20) was used as the mean cost per nurse consultation, with the cost of the same length of visit by a community specialist (£23) and clinical support worker (£8) forming the upper and lower confidence intervals (α = 44.4444, β = 0.4500).
    • The mean national unit cost of an emergency room visit is £62 with an inter quartile range of £37 (α = 4.8985, β = 12.6510).
  • First-line therapy for symptomatic UTI in England currently includes the antibiotics trimethoprim, nitrofuratonin, cefalexin, and pivmecillinam; what drug is prescribed varies by region and between practices 13. In the base case analysis, the model assumes an average treatment length of 5 days for each drug (with the exception of pivmecillinam), based on an average treatment duration of 3 and 7 days for women and men, respectively 13. Mean unit cost was calculated as a simple mean based on the following costs listed in the NHS Drug Tariff 2010 323 and dosages from the prescribing support unit 467 (the most expensive and least expensive course of treatment was used as confidence intervals used to inform the parameter distribution):
    • Trimethoprim 200mg twice daily for five days (£0.75)
    • Nitrofuratonin 50mg four times daily for five days (£1.91)
    • Cefalexin 500mg twice daily for five days (£1.30)
    • Pivmecillinam 200mg three times daily for three days (£4.05)
  • The same sources and methods were used to calculate the average cost of second-line antibiotics used to treat first-line resistant UTIs. The cost of second-line antibiotics was calculated as a simple mean of the costs of the following individual drugs:
    • Ciprofloxacin 250mg three times daily for seven days (£2.33)
    • Cefaclor 250mg three times daily for seven days (£5.28)
    • Cefixime 200mg once daily for seven days (£13.23)
    • Norfloxacin 400mg twice daily for seven days (£3.81)
    • Ofloxacin 400mg once daily for seven days (£5.82)
    • Pivmecillinam 400mg four times daily for seven days (£50.40).
  • In both first- and second-line treatment, it is assumed that patients are fully compliant. Given the short duration of the course of antibiotics, this is considered reasonable 131.
  • Increased fluid intake and frequent urination associated with UTI will result in increased catheter use while the patient is symptomatic. Therefore, the cost of additional catheters (and lubricant for non-coated catheters) was added to the cost of each infection treated in the community. The GDG indicated that an average of 12 catheters per infection (and infection exacerbation) would be a reasonable estimation.
  • Patients with multidrug resistant infections are usually admitted to hospital for intravenous drug therapy 13. The cost of treatment for a multidrug resistant infection was calculated as a weighted average reference cost for kidney or urinary tract infection with intermediate complications (LA04E; £2,097 (£1, 681 to £2417)) and without complications (LA04F; £1, 618 (£1, 203 to £1, 822)). The average excess bed day cost for each HRG is £197 (£154 to £224) and £195 (£154 to £222)100, respectively. These costs were weighted according to reported activity, with 73% of the total cost attributed to LAO4E, in order to produce a total average cost for people with multi-drug resistant UTI.
  • The cost of treatment for bacteraemia secondary to UTI was assumed to be equivalent to the non-elective reference cost for kidney or urinary tract infection with major complications (code LA04D) with a national average unit cost of £2938 (£2264 to £3352) and average excess bed day cost of £198 (£152 to £227)100. In the UK, bacteraemia caused by resistant organisms does not appear to have a significant impact on length of hospital stay compared to bacteraemia caused by susceptible organisms (Melzer and Petersen 2007)294.
Cost of treatment for urethral complication

The cost of treating a urethral complication was estimated based on reference cost group LB30B: urethra disorders and intermediate/minor procedures without complications with a national average unit cost of £1,268 and lower and upper quartile unit cost of £908 and £1,399 100. The effect of increased treatment cost due to failed or repeat procedures was explored in sensitivity analysis.

J.2.4. Sensitivity analyses

ISC in people who do not have SCI

In the absence of any clinical data, it was assumed that the relative risk of symptomatic UTI for each type of catheter was the same as that observed in the SCI population. This was a necessary assumption in order to explore the cost-effectiveness of intermittent catheter types across a wider group of people with bladder dysfunction. The GDG indicated that it was also a reasonable assumption as there is no clinical reason to suspect that SCI patients would respond any differently to any one type of catheter than any other patient using ISC.

Cohort probabilities

People with bladder dysfunction not caused by SCI are a highly diverse group of patients, with a wide range of ages, health states, disabilities. Several cohort probabilities were changed to reflect the probability of antibiotic resistance and mortality in a more heterogeneous population. There is very little epidemiological evidence about the prevalence and morbidity of UTI in this population as a whole; young women appear to be the most common subject of UTI-related research in the literature. The GDG indicated that if the sample size were large, this population may represent a sufficiently heterogeneous group from which to draw the parameters to inform probabilities for the sensitivity analysis.

A study of over 75,000 patients from the UK General Practice Research Database was used to estimate the probability of treatment failure in this group of patients. This study found that between 12% and 16% of women treated for UTI return within 28 days for a further course of treatment, regardless of the antibiotic initially prescribed 255. This is consistent with the findings of a study of a large pharmaceutical database in the Netherlands 158. Following input from experts at the Health Protection Agency (Neil Woodford and Alan Johnson; personal communication), and review of several other data sources10,79,227,294,379,398,514, it seems likely that between 4% to 8% of community acquired urinary isolates in the UK and USA are resistant to ciprofloxacin or contain extended-spectrum beta-lactamase (ESBL) producing bacteria. Therefore, it was assumed that approximately 6% of UTIs in the UK are multidrug resistant (Table 29). The same probability of developing bacteraemia and of dying from multidrug resistant UTI as in the base case analysis was assumed to apply to this analysis. The probability of mortality from bacteraemia was obtained from a meta-analysis by Bryan and Reynolds (1984) 52.


The life expectancy and utility values informing the model were also updated (Table 29). Three studies were identified through our quality of life review (Appendix K:) which allowed a series of multiplicative relationships to be used to calculate utility values per symptom day for patients without SCI. The per-day utility value for patients who recover from symptomatic UTI after empirical treatment was derived from a study by Ellis and Verma (2000) 120, in which the SF-36 questionnaire was administered to a group of otherwise healthy women suffering from UTI and their matched controls. The algorithm suggested by Ara and Brazier (2008) 21 was used to convert SF-36 responses into EQ-5D health state valuations, which were adjusted based on average mapped EQ-5D values for the UK population 207.

A study by Ernst et al (2005)123 used the Quality of Well Being to evaluate the effect of failed antibiotic treatment compared to clinical cure in patients being treated for UTI. In order to calculate the proportional utility decrease for patients with first line resistant infections, the reported value for patients who failed treatment at 7 days was divided by the score for patients who were cured after 3 days. A multiplicative relationship was assumed to apply to the EQ-5D value derived from Ellis and Verma (2000) in order to estimate the utility value for patients with first-line resistant UTI. The same calculation was applied to patients experiencing treatment failure at 14 days in order to estimate the daily utility value for patients with multidrug resistant UTI. In the absence of any utility values for UTI-associated bacteraemia, a value derived from inpatients with bloodstream infections of unspecified origin was used to inform this health state 428.

The recall period used by Ellis and Verma (2000) asked patients about their quality of life within the past 24 hours. To obtain QALYs, the daily utility value for each health state was multiplied by the duration of the health state, assuming that the rest of the year was lived in a state of full health (Equation 6). For patients who achieve clinical cure after empiric treatment, an average symptom duration of 3.5 days was assumed based on expert opinion. The duration of first-line resistant UTI was assumed to be 8.5 days allowing time for the patient to realise treatment failure, consult a healthcare professional, and begin a second course of antibiotics. Given that patients with multidrug resistant UTI and bacteraemia would be admitted to hospital for treatment, it was assumed that these infections would last an average of 10 days based on expert opinion and NHS Reference Cost data.

Equation 6. QALYs for patients without SCI
Resource use and cost

All costs remained the same as in the base case.

Urethral complications

Currently, there is no comparative clinical evidence to suggest that the use of one type of catheter results in fewer urethral complications compared to another. However, there have been animal and laboratory studies suggesting that the coated catheters reduce removal friction and cell adhesion compared to non-coated catheters 275,489. This is sometimes interpreted as evidence that hydrophilic catheters cause less urethral trauma and may lead to a decrease in urethral complications. The effect of a reduction in urethral complications associated with hydrophilic and gel reservoir catheters was explored in the sensitivity analysis.

Parameter uncertainty

One- and two-way sensitivity analyses were undertaken to evaluate the relative impact of the probability of antimicrobial resistance, mortality, utility, resource use and cost on the outcome of the model.

J.2.5. Value of information analysis

All decisions about the cost-effectiveness of interventions are associated with a certain degree of uncertainty in the evidence base. As a result of this uncertainty there will always be a chance that the wrong decision will be made. A wrong decision would be costly in terms health benefit and resources forgone. The best way to resolve this uncertainty is to gather more information, but this may also be costly and time consuming. Value of information (VOI) analysis provides a framework for determining the expected benefit of future research by taking into account both the probability that further information will change the adoption decision, the sample size necessary to achieve maximal benefit, and the opportunity cost of conducting a research project of this size. VOI aims to answer the question of whether future research should be conducted, and if so, on which uncertain parameters, and provides and estimate of the optimal sample size for each study.

Expected value of perfect information (EVPI)

Per-patient EVPI

The first step of VOI is to estimate the expected value of perfect information (EVPI) per patient. As stated in section J.2.6, the decision rule that we must use when making recommendations is to choose the option that maximises net benefit based on current information. If we had perfect information, we would always choose the correct option and there would be no loss. However, in order to achieve perfect information we would require a study with infinite sample size.

In reality, there will always be a degree of error associated with each data input in the decision problem. The expected cost of uncertainty is determined jointly by the probability that the decision based on based on existing information will be wrong and the consequences of a wrong decision. The expected loss as a result of uncertainty is equivalent to the expected gain from eliminating uncertainty (i.e. the EVPI). Mathematically, the EVPI is the difference between expected maximum net benefit with perfect information and the maximum expected net benefit with current information. The per-patient EVPI for each model was generated directly from the simulated output (over 10 000 iterations) from TreeAge 2009.

Population EVPI

The next step in determining the EVPI is to calculate the upper limit for future research expenditure by taking into account both the current and future patient populations who might be expected to benefit from the intervention in question. Multiplying the per-patient EVPI by the number of current and future people using intermittent catheterisation in England and Wales who will be affected by the decision provides us with an upper boundary for future research expenditure (Equation 7).

Equation 7. Population EVPI
Current and future patients affected by the decision problem

Several sources of data and a series of assumptions were used to inform the population estimate for the value of information analysis:

  • Prevalence and incidence of traumatic SCI in England and Wales: There are currently 40,000 people in the UK living with SCI (Kennedy 1998). The majority of these injuries are caused by trauma. The annual incidence of traumatic-SCI is approximately 15 new cases per million per year in Western Europe 86.
  • Prevalence and incidence of non-traumatic SCI in England and Wales: There is little information about the prevalence of other conditions causing SCI such as spinal stenosis, tumours, ischaemia and inflammation, but it is thought that approximately 36% of spinal cord injuries are non-traumatic291. The annual incidence of non-traumatic SCI is estimated at 26.3 cases per million321,322
  • Proportion of patients with SCI who use ISC: Roughly 80% of people with SCI have some degree of difficulty with bladder function269,269; it was assumed that 60% of these patients would use ISC. Approximately 90% of individuals with SCI live in private residences following rehabilitation318 and 40% are in school or employment318. It was assumed that the same proportion applies to those with non-traumatic SCI.
  • Proportion of patients who should not use multiple use non-coated catheters: It was assumed that people who do not live in a private residence and people who are at work or school do not have regular access to facilities needed to wash and dry catheters. Clean multiple use non-coated ISC was assumed to be an option for the remainder of the SCI population.
  • Lifetime of the technology: Current guidelines recommend that the selected time horizon should reflect the effective lifetime of the technology. A search of PubMed and Google Scholar did not reveal any evidence of imminent new developments in catheter material research, but there is active work in this field. Ten years was thought to represent a reasonable estimate of time before a new type of intermittent catheter might be expected to be brought to market.
  • Current and future population of England and Wales: The population of England and Wales is currently 62 million, projected to rise to approximately 67 million over the next 10 years337.

Given current population and incidence estimates and discounting at a rate of 3.5%, over the next 10 years approximately 13, 437 people will have a choice between using clean or sterile ISC. Approximately 11, 500 will have a choice between different types of sterile ISC.

Expected value of partial parameter information (EVPPI)

It is also possible to identify which type of additional evidence is most valuable to the decision problem by calculating the expected value of partial parameter information (EVPPI). The EVPPI is an estimate of the value of eliminating uncertainty regarding a particular parameter or set of parameters (for example, baseline risks or quality of life). This information can be used to indicate which endpoints should be included in further experimental research, or to focus research on obtaining more precise estimates for values which may not require an experimental design. As with the EVPI, the per-patient EVPPI must also be multiplied by the affected population over the appropriate time period to obtain the population EVPPI. The population EVPPI provides an upper boundary for the cost of research into particular parameters.

The method of calculating EVPPI is conceptually very similar to EVPI. It is the difference between the expected value with perfect and current information about a parameter or group of parameters. The crucial difference is that EVPPI requires a two-level, or ‘nested’, Monte Carlo procedure. The procedure begins with an outer loop sampling values from the distribution of the parameters of interest, and for each of these, an inner loop sampling the remaining parameters from their conditional distribution.

The per patient EVPPI for the current model was calculated using TreeAge 2011. The outer loop was run 400 times (that is, each parameter of interest was sampled 400 times) and the inner loop was run 5 000 times (that is, for each of the 400 outer samples, the Monte Carlo analysis was run for 5 000 iterations). Given time constraints, this was thought to represent a pragmatic solution to the suggestion that the inner loop should be run 1000 to 10 000 times 335.

Expected value of sample information (EVSI) & expected net benefit of sampling (ENBS)

Although the population EVPI and EVPPI provide an estimate of the maximum budget for research, a positive value does not mean that such a budget should be set. In order to determine the net benefit of conducting research into a particular topic or specific set of parameters, it is essential to first determine the optimal sample size.

With increasing sample size, the EVSI will reach a ceiling which equals the maximum EVPI/EVPPI (representing an infinite sample size). However, with increasing sample size, the costs of research will also increase. The expected net benefit of sampling (ENBS) is defined as the difference between the expected value of sample information (EVSI) with sample size n and the cost of conducting research with sample size n. The point at which EVBS is maximised is the optimal sample size for the proposed study. If there is no positive sample size for which the EVBS is greater than zero, then additional research is not warranted and the decision should be based on current information only.

The EVSI was calculated by repeatedly running the EVPPI analyses for different n values (outer loops). These analyses were only undertaken for parameters with EVPPI values greater than zero. The analyses were run 5 times for each sample size and an average EVSI obtained for each sample size.

Cost of research

Clinical trial budgets are a mixture of direct, indirect, fixed and variable costs. A search was preformed to identify average research budgets for similar types of trials but no information was identified. It was assumed that a trial of this type would be relatively inexpensive to administer. A fixed cost of £50, 000 was used to account for the estimated full time salary of a study coordinator, to supplement the costs of a clinician/researcher and cover the cost of any additional expertise needed for data analysis. An estimated incremental cost of £500 per patient was also assumed to relate to the costs of administration associated with each patient. It was assumed that the costs of the catheters themselves would be covered by the NHS, not the research grant.

J.2.6. Interpreting results

The results of cost-effectiveness analysis are presented as incremental cost-effectiveness ratios (ICERs). ICERs are calculated by dividing the difference in costs associated with two alternative treatments by the difference in QALYs:

ICERs=Cost of B-Cost of AQALY of B-QALY of A

Where more than two interventions are being compared, the ICER is calculated according to the following process:

  1. The interventions are ranked in terms of cost, from least to most expensive.
  2. If an intervention is more expensive and less effective than the preceding intervention, it is said to be ‘dominated’ and is excluded from further analysis.
  3. ICERs are then calculated for each drug compared with the next most expensive non-dominated option. If the ICER for a drug is higher than that of the next most effective strategy, then it is ruled out by ‘extended dominance
  4. ICERs are recalculated excluding any drugs subject to dominance or extended dominance.
  5. When there are multiple comparators, the option with the greatest average net benefit may also be used to rank comparators.

NICE’s report ‘Social value judgements: principles for the development of NICE guidance’ sets out the principles that GDGs should consider when judging whether an intervention offers good value for money 316. In general, an intervention is considered to be cost-effective if either of the following criteria apply:

  • The intervention dominates other relevant strategies (that is, is both less costly in terms of resource use and more clinically effective compared with all the other relevant alternative strategies), or
  • The intervention costs less than £20,000 per quality-adjusted life-year (QALY) gained compared with the next best strategy

J.2.7. Validation

The model was developed in consultation with the GDG; model structure, inputs and results were presented to and discussed with the GDG for the purpose of clinical validation and technical interpretation.

The model was systematically checked by the health economist undertaking the analysis; this included inputting null and extreme values and checking that results were plausible given inputs. The model was also peer reviewed by the lead health economist at the NCGC; this included systematic checking of many of the model calculations.

J.3. Results

J.3.1. Base case analysis

For patients who are able to wash and re-use catheters, this represents the most cost-effective option for intermittent self catheterisation. For patients who may not be in a situation that allows them to wash and re-use catheters, gel reservoir catheters are most cost-effective. Results of the base case probabilistic analysis are summarised in Table 30 and shown graphically in Figure 98.

Table 30. Base case analysis results (probabilistic).

Table 30

Base case analysis results (probabilistic).

Figure 98. Base case analysis results (probabilistic).

Figure 98

Base case analysis results (probabilistic). Results for each subgroup are plotted on the incremental cost-effectiveness ratio axis. The non-coated multi-use catheter is the least costly strategy and has been used as the baseline comparator. Therefore, (more...)

In both scenarios, gel reservoir catheters are the most effective type of catheter (i.e. associated with more QALYs than the other catheter types). However, they are not always the most cost-effective option. According to NICE decision making rules (page 387), an intervention can only be considered cost-effective if its ICER falls below the £20,000 to £30,000 threshold. According to the results of our model, when gel reservoir catheters are compared to multiple-use non-coated catheters, the ICER is £51, 345. In other words, the QALY gain associated with gel reservoir catheters compared to multiuse non-coated catheters is not enough to justify the large difference in cost.

When it is not possible to re-use non-coated catheters, gel reservoir is the most cost-effective type of catheter. Compared to hydrophilic catheters, gel reservoir catheters are more effective and slightly more expensive, with an ICER of approximately £3, 270 per QALY.

As outlined in Table 31, the main cost driver in the model is the cost of the catheters (and lubricant where applicable). The cost attributed to treating infections is lowest for gel reservoir catheters, however these catheters are associated with the greatest catheter cost. The opposite is true of multiple use non-coated catheters.

Table 31. Discounted total cost per patient with SCI over a lifetime horizon (deterministic).

Table 31

Discounted total cost per patient with SCI over a lifetime horizon (deterministic).

J.3.2. Sensitivity analysis

ISC in people who do not have SCI

The results of the model are unchanged in patients with bladder dysfunction that is not caused by SCI, assuming the same relative effectiveness as observed in the SCI population. Where it is possible to wash and re-use non-coated catheters, gel reservoir catheters are not recommended on the basis that the ICER is £149, 559. When re-use of non-coated catheters is not an option, gel reservoir catheters represent the most cost-effective option. In both cases, single-use non-coated catheters are excluded from the analysis by dominance (Table 32, Figure 99).

Table 32. Sensitivity analysis results (probabilistic) – patients without SCI.

Table 32

Sensitivity analysis results (probabilistic) – patients without SCI.

Figure 99. Sensitivity analysis results (probabilistic) – patients without SCI.

Figure 99

Sensitivity analysis results (probabilistic) – patients without SCI.

Baseline risk of infection in people without SCI

The baseline risk of infection in people without SCI is likely to differ according to the specific population in question. Older women in particular are likely make up a large proportion of people performing ISC and are very susceptible to UTIs448. The baseline probability of infection used in the base case model was 67.8%, based on an annual risk of 1.14; no higher estimates were identified in the literatrure. In exploratory analysis, the baseline risk of UTI was increased to 2 and 4, with an associated annual probability of 86% and 98%, respectively (Table 33). In both cases, noncoated multiple use catheters remain the most cost-effective option for ISC.

Table 33. Baseline risk of UTI in people without SCI – exploratory analysis (probabilistic).

Table 33

Baseline risk of UTI in people without SCI – exploratory analysis (probabilistic).

In situations where non-coated catheters can be washed and reused (in patients with SCI)

Urethral complications

When the relative risk of urethral complication associated with the use of hydrophilic catheters is half that of other catheters, they are still excluded from the analysis by extended dominance. This remains the case when the probability of urethral complications associated with hydrophilic catheters is eliminated and the cost associated with urethral complications is doubled. The same is true for gel reservoir catheters (i.e. when the risk of urethral complication associated with the use of gel reservoir catheters is reduced by half or eliminated and cost doubled, the ICER remains well above the £20,000 cost-effectiveness threshold). The results of these exploratory analyses are presented in Table 34.

Table 34. Results of one- and two-way sensitivity analyses (probabilistic) – Clean ISC.

Table 34

Results of one- and two-way sensitivity analyses (probabilistic) – Clean ISC.

Probability of antimicrobial resistance and mortality

Antimicrobial resistance is dynamic and difficult to predict. The probability of treatment failure, multidrug resistance and mortality were each examined at the upper limit of their confidence intervals in one- and two-way sensitivity analysis. In each case, clean non-coated catheterisation is the most cost-effective strategy (Table 34).

Threshold analysis – catheter use

The number of clean non-coated catheters used per year was varied between an average of 60 per year (average 5 per month) and 1825 per year (average 5 per day) in a threshold analysis. Clean ISC ceases to be the most cost-effective option when an average of 208 non-coated catheters is used per year; this equivalent to approximately 4 catheters per week. Therefore, if on average patients use more than four non-coated catheters per week, gel reservoir catheters are the most cost-effective option for ISC.

In situations where non-coated catheters cannot be cleaned (in patients with SCI)

Urethral complications

When the probability of urethral complications associated with hydrophilic complications is halved, gel reservoir remain the most cost-effective option in situations where clean ISC is not an option. Gel Reservoir catheters are also the most cost effective option when the probability of urethral complications associated with the use of hydrophilic catheters is eliminated and the cost is doubled (Table 35).

Table 35. Results of one- and two-way sensitivity analyses (probabilistic) – Probability and cost of urethral complications in situations where non-coated catheters cannot be washed and reused.

Table 35

Results of one- and two-way sensitivity analyses (probabilistic) – Probability and cost of urethral complications in situations where non-coated catheters cannot be washed and reused.

J.3.3. Value of information analysis

The per-patient and population EVPI is presented in Table 36. At a threshold of £20, 000, the maximum budget for research into the cost-effectiveness of different types of catheter for ISC is approximately £2.5 million. Source/Note: At a threshold of £20, 000 per QALY.

Table 36. Expected value of perfect information.

Table 36

Expected value of perfect information.

Table 37 presents the EVPPI for each group of parameters. Of the five general parameter groups across each of the two models, only one had a nonzero EVPPI. Note that EVPPI is not expected to sum to EVPI due to interaction between parameters (for example, collecting information about one parameter may affect the value of collecting information on another with which it is closely related). Calculating EVSI and ENBS for the parameter distributions of the relative risk of symptomatic UTI associated with gel reservoir and hydrophilic catheters revealed that under our estimates of the cost of research, conducting additional research into this decision question will not yield a net benefit (Table 38).

Table 37. Expected value of perfect parameter information.

Table 37

Expected value of perfect parameter information.

Table 38. Expected value of sample information and expected net benefit of sample information: Relative effectiveness of gel reservoir vs. hydrophilic catheters.

Table 38

Expected value of sample information and expected net benefit of sample information: Relative effectiveness of gel reservoir vs. hydrophilic catheters.

J.4. Discussion

J.4.1. Summary of results

This analysis combines the best available evidence about the costs and consequences of each type of catheter used for intermittent catheterisation. Based on the results of the model, we can conclude that the small decrease in symptomatic infections associated with single-use gel reservoir and hydrophilic catheters is not enough to justify the large increase in the cost of these catheters compared to multiple use non-coated catheters. As a result, clean multiple use non-coated catheters represent the most cost-effective type of catheter for ISC. This conclusion was robust to a wide range of sensitivity analyses, including the increased probability of urethral complications that may be associated with the use of non-coated catheters. However, multiple use non-coated catheters cease to be the most cost-effective choice when patients use an average of more than two catheters per day. Compliance and behaviour are therefore important factors for healthcare workers to consider when prescribing an ISC regime.

Healthcare workers must also consider other patient-specific situations when deciding which catheter to prescribe. Washing and re-using non-coated catheters may not be an appropriate option for all patients. When clean ISC is not an alternative, gel reservoir catheters may be considered the most cost-effective choice for ISC. If hydrophilic catheters are preferred to gel reservoir catheters, they may also be considered as an option.

J.4.2. Patient preference and compliance

Under the current decision rule, the recommended treatment is identified as that with the highest ICER that falls below the cost-effectiveness threshold. Preferences are incorporated into the cost-utility analysis through the values that are attached to each health state; these values represent the average weight attached to each health state by the general population and are assumed to be independent of factors related to the health care process.

The use of societal values creates the potential for conflict where individual patients hold a strong preference for a particular treatment that is not reflected in the decision made at the societal level49. It has been suggested that one way to incorporate individual patient preference into cost-effectiveness decisions would be to adopt a two-part decision process which gives the patient the choice of the most cost-effective treatment plus all cheaper options 103.

Of the five RCTs included in our review of clinical efficacy, three included a measure of patient preference and comfort; none found any difference between catheter types. Nevertheless, it is still possible that patients may find one type of catheter more comfortable or easier to use than another and therefore derive a benefit from the catheter that is not captured in the model102. When deciding between gel reservoir and hydrophilic catheters for patients who cannot use multiple non-coated catheters, the GDG did not wish to force the consumption of more costly gel reservoir catheters. If a patient has a strong preference for hydrophilic catheters then the GDG agreed that they should be able to choose this less costly option.

It is important to note that under this rule patients should not be given a choice of therapies that are more expensive and more costly than the most cost-effective treatment 103. In other words, this line of reasoning cannot be extended to patients who are able to use clean multiple use non-coated catheters but prefer not to, nor to patients who prefer single use non-coated catheters to single use gel reservoir or hydrophilic catheters.

J.4.3. Limitations & interpretation

This analysis did not take into account the dynamic and extremely complex nature of antimicrobial resistance. Although we sought to use the most current, relevant estimates to inform this analysis, data about the prevalence and mortality associated with antibiotic resistant UTIs is limited and it is impossible to predict the future of this phenomenon. If the prevalence, clinical and economic impact of antimicrobial resistance increases beyond the upper estimates used in this model, then the cost-effectiveness of clean intermittent catheterisation in this population may have to be re-visited.

The clinical review undertaken as part of this analysis was not designed to evaluate the most effective method of cleaning non-coated catheters. There are many different methods of cleaning advocated in the literature (such as soap and water, boiling, microwave sterilisation, and peroxide application) and no consensus as to which is best. Only two of the manufacturers contacted during the development of this guideline provided any direction as to how to clean and store non-coated catheters – both advised washing with soap and water and leaving to dry in a clean area, using paper towels to absorb excess water if necessary.

J.4.4. Generalisability to other populations / settings

The analysis presented in this report compared all four options for performing ISC from a UK NHS perspective, taking into account a wide range of considerations with extensive sensitivity analyses. It is directly applicable to this guideline and the current UK NHS.

This analysis was designed to assess the cost-effectiveness of different types of intermittent catheters for patients performing intermittent self catheterisation in the community. Outside of the community and primary care setting, there may be other considerations which must be taken into account when considering the cost-effectiveness of each strategy.

The main driver of cost differences in the model is the cost of the catheters themselves. Therefore, the results of this model are only applicable to healthcare systems in which a single payer is responsible for both the cost of the catheter regime and the cost of treatment for UTI and UTI-associated complications.

J.4.5. Comparisons with published studies

Several studies have noted similar effectiveness and lower costs with the use of a clean multiple use non-coated catheters compared to single use catheters 110,182,383. However, none have attempted to evaluate the costs and quality of life associated with symptomatic UTI or its downstream consequences. To the best of our knowledge, this represents the first cost-utility analysis of intermittent self catheterisation. By combining the best available evidence about the relative efficacy and costs of the different methods of ISC, this analysis aimed to address an issue which has been a source of debate for many years516.

Clean intermittent self catheterisation was first introduced in the 1970s as the preferred method of intermittent catheterisation for patients in the community. Lapides et al (1972)246 proposed that bladder distension was the main contributing factor to UTI rather than the introduction of bacteria to the bladder. Partly on the basis of this theory (which still holds sway within the urological literature) and partly based on non-systematic reviews of the clinical evidence, it is interesting to note that several evidence- and consensus-based guideline groups have recently made recommendations which are very similar to the conclusion reached by our analysis:

  • In 2010, the Infectious Diseases Society of America195 published clinical guidance recommending the use of multiple-use catheters in outpatient and institutional settings, while recognising that multiple use catheters may not always be an option if patients find it inconvenient to clean their catheters when away from home.
  • The European Association of Urology Nurses 144 further specifies that catheterisation should be sterile when preformed by someone other than the patient.
  • In 1996, the Agency for Healthcare Policy and Research8 clinical practice guideline on the management of urinary incontinence supported the use of clean intermittent self catheterisation.

J.4.6. Conclusion = evidence statement

Washing and re-using non-coated catheters is the most cost-effective option for intermittent self catheterisation. In situations where it may not be feasible or appropriate to wash and reuse non-coated catheters, gel reservoir catheters appear to be the most cost-effective catheter type. However, if patients prefer hydrophilic catheters to gel reservoir catheters, they may also be considered cost-effective. Single use non-coated catheters are never a cost-effective option for intermittent self catheterisation.

J.4.7. Implications for future research

The expected value of future research is a function of the amount of uncertainty associated with the current adoption decision. Based on best available evidence, the current model reveals that among patients for whom multiple use non-coated catheters are an option, there is very little uncertainty associated with the optimal choice of intermittent catheter. Concequently, the results of our value of information analysis suggest that obtaining more information about this decision would not be a cost-effective use of NHS resources.

Copyright © 2012, National Clinical Guideline Centre.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of National Clinical Guideline Centre to be identified as Author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

Cover of Infection: Prevention and Control of Healthcare-Associated Infections in Primary and Community Care
Infection: Prevention and Control of Healthcare-Associated Infections in Primary and Community Care: Partial Update of NICE Clinical Guideline 2.
NICE Clinical Guidelines, No. 139.
National Clinical Guideline Centre (UK).

NICE (National Institute for Health and Care Excellence)

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...