Preferences for antimuscarinic therapy for overactive bladder


Paul Swinburn, Oxford Outcomes Ltd, Seacourt Tower, West Way OX2 0JJ, UK. e-mail:


Study Type – Preference (discrete choice experiment)

Level of Evidence 2a

What’s known on the subject? and What does the study add?

Whilst antimuscarinic treatments are widely used little work has been done to understand how patients consider the relative benefits and costs associated with their use.

This study provides data which demonstrates both the perceived value of symptom reduction and burden associated with common antimuscarinic AEs. These findings may prove useful in informing prescribing decisions.

OBJECTIVE • To examine patient preferences and strength of preferences for treatment for the various symptoms of overactive bladder and adverse events associated with the use of antimuscarinic treatments.

PATIENTS AND METHODS • A discrete choice experiment (DCE) survey was developed that detailed treatment choices in terms of attributes relating to their efficacy in reducing symptoms and the likelihood of experiencing typical adverse events. Levels for each attribute were based on a literature review, qualitative interviews and a meta-analysis of clinical trial data.

• Attributes were combined into choice sets using a fractional orthogonal design that had been folded over. Pairs of choice sets were presented to overactive bladder (OAB) patients (n= 332), who indicated which treatment alternative they preferred. Data were analysed using the conditional logit model.

RESULTS • Participants expressed the strongest preference for the avoidance of urgency incontinence episodes, followed by preference for a reduction in the experience of urinary urgency and the number of micturition episodes. The influence of the likelihood of experiencing an adverse event on treatment preference was also estimated.

• Finally, marginal rates of substitution were calculated to demonstrate the relative value of trade-offs between the various attributes.

• Treatment preferences were found to be broadly similar across two patient age groups (i.e. under 45 s and 45 and over).

CONCLUSION • The study demonstrates that individuals with OAB place significant emphasis on the prospect of reduction in symptoms. Avoidance of incontinence episodes is particularly valued and equivalent to a much greater reduction in the frequency of micturition or experience of urgency. However, even a modest increase in the likelihood of experiencing an adverse event could easily motivate a change in treatment preference.


overactive bladder


discrete choice experiment


randomized controlled trials


health-related quality of life


stress urinary incontinence


marginal rates of substitution


The International Urogynecological Association (IUGA) and the International Continence Society (ICS) define overactive bladder (OAB) as ‘urinary urgency, usually accompanied by frequency and nocturia, with or without urgency urinary incontinence, in the absence of urinary tract infection (UTI) or other obvious pathology’[1]. OAB is a widespread chronic condition that affects 455 million people worldwide [2]. The total direct economic impact associated with the condition has been estimated at over €1 billion in the UK alone [3].

Although the aetiology of OAB is poorly understood, disturbances in nerves, detrusor smooth muscle and urothelium are all thought to lead to OAB [4]. Whilst the overall prevalence of the condition appears to be similar for both men (10.8%) and women (14.6%) [5], women appear to have an increased likelihood of experiencing urgency incontinence. It has been suggested that hormonal changes occurring around and after the menopause may exacerbate symptoms [6].

OAB presents a serious challenge to the preservation of health-related quality of life (HRQL) [7,8]. In order to cope with the symptoms of OAB, many sufferers develop elaborate coping behaviours aimed at hiding and managing urine loss. These include such behaviours as limiting daily travel to places and routes where the locations of toilets are known, avoiding sexual intimacy, reducing fluid intake and carrying extra clothes [2].

OAB can often be effectively managed with bladder and behavioural training, biofeedback, electrical stimulation, pharmacologic treatments, or with a combination of these therapies [9]. Antimuscarinic agents have become the first-line pharmacotherapy of choice for OAB [10,11]. However, efficacy and tolerability vary among agents and between patients [12].

The aim of this study is to explore OAB patients’ strength of preferences for outcomes associated with antimuscarinic therapies, both in terms of their efficacy and associated adverse events. By understanding the relative value placed on certain benefits conferred by the treatment as well as the likelihood of experiencing side-effects, it should be possible to better consider the overall patient treatment experience. This data could potentially prove particularly valuable in informing treatment selection.

Patient preferences were elicited through the use of a discrete choice experiment (DCE) survey [13]. DCEs (also known as conjoint analysis) are an attribute-based survey method for assessing preferences. They entail presenting respondents with a series of hypothetical scenarios (choice sets) composed of two or more competing alternatives that vary along several attributes [14]. In recent years, DCEs have been increasingly utilized to help understand preferences in the field of health and healthcare [15–18].



The purpose of the study was to capture patient preferences and trade-offs for a number of symptoms and adverse event-related attributes of therapy for OAB (Fig. 1). In order to identify the attributes for inclusion in the survey, a literature review and in-depth qualitative interviews with patients were undertaken.

Figure 1.

Study design.


A literature search was conducted for studies detailing the experience of undergoing OAB treatment in terms of symptoms and adverse events. The search was restricted to studies published in the English language between the years 1997 and 2008. A particular emphasis was placed upon qualitative research that examined the impact of OAB on an individual’s HRQL.

The review found that OAB was associated with a considerable decline in HRQL [19–21]. The studies identified in the review revealed a number of symptoms and common antimuscarinic-related adverse events that could impact the social, psychological, occupational, domestic, physical and sexual functioning of sufferers [2].

A systematic review of randomized controlled trials (RCTs) demonstrated that antimuscarinic agents were efficacious for the treatment of overactive bladder and improved HRQL for patients [11]. A recent review summarized the mostly milder adverse events that were relatively frequent with antimuscarinic treatments [12]. The adverse events most commonly reported were dry mouth and constipation.


A number of draft attribute descriptions were created based on the literature review findings. These attributes included three of the most common symptoms of OAB (i.e. increased micturition frequency, urinary urgency and urge incontinence) and the two most common adverse events associated with its treatment (i.e. dry mouth and constipation). Descriptions were provided of the experience of adverse events in terms of their typical severity and physical impact. Attributes were defined at multiple levels, which specified the level of symptoms experienced (for example, 8 episodes of micturition in a 24-h period) or the likelihood of the adverse event occurring (for example, 10% probability of constipation). Although these attributes were based upon recognized medical definitions, simplified naming conventions were adopted to aid participant comprehension.


A semi-structured interview guide was developed that focused on exploring the attribute descriptions in relation to participants’ own personal experiences of the condition. The ability of the descriptions to characterize an individual’s condition accurately was evaluated, and participants were encouraged to provide their own thoughts and feelings regarding the descriptions’ wording. Patients who had been diagnosed with OAB were identified through the use of a specialist patient recruitment agency.

The interviews (n= 7) indicated that participants generally experienced very little difficulty in relating the attribute descriptions to their own experiences. Individuals could readily identify the extent to which they had experienced such events and felt that the descriptions adequately captured the intended phenomena. The incontinence attribute description underwent a minor revision to reduce ambiguity.


The attributes and levels were combined into choice sets using a published orthogonal array ( and folded over to ensure zero overlap and orthogonality in differences. The order of presentation for the choice sets was randomized so as to avoid possible ordering effects, and then they were presented as pairwise choices. Each participant undertaking the survey was presented with 18 choice sets and asked to state whether treatment A or treatment B was preferred within each choice set (Fig. 2). Individuals were also asked to assess their own current symptom experiences using the same attributes. Sociodemographic data were also collected.

Figure 2.

Example discrete choice experiment survey question.


A pen-and-paper version of the DCE survey was piloted with seven OAB patients using a specially prepared cognitive debriefing guide [22]. The intent of this process was to determine if there were any significant cognitive challenges in completing the survey or if there were any discernable sources of error resulting from the survey design. Individuals reported no difficulties in completing the survey, and this final design was adapted to an electronic web-based version for use in the main data collection phase. Careful attention was paid to the adaptation process to ensure that the appropriate usability concerns were addressed.


The main data collection phase of the study was undertaken using the web- based DCE survey. A specialist patient recruitment agency identified individuals who had reported OAB symptoms and provided them with a link to the study website. After information on the study had been provided to potential participants and informed consent obtained, individuals completed the Modified OAB-V8 [23] to establish their symptom status. Individuals who met the scoring criteria for OAB without the presence of stress urinary incontinence (SUI) were deemed eligible to take part in the study. In total, 332 participants completed the online survey (with 3031 being excluded owing to failure to meet the inclusion criteria, to provide consent, or to complete the survey fully).

Prior to the commencement of the main data collection, the study protocol was submitted for ethical review by the Independent Institutional Review Board. No modifications were deemed necessary, and ethical approval was granted prior to the study proceeding.


Demographic and other background data were summarized using frequencies and descriptive statistics as appropriate. The DCE survey data were analysed using a conditional logit (logistic) model. The conditional logit model evaluates choice responses after conditioning them on the attributes of the other treatment alternatives available within the choice set. Hence, if, for example, ‘Treatment option A’ is preferred in the choice set no. 1, this preference is conditional on the attributes of ‘Treatment option B’. The coefficients obtained from the logit model provide an estimate of the (log) odds ratios of preference for treatment attributes.


The demographic data for the study participants are presented in Table 1.

Table 1.  Participant characteristics for the study population (n= 332)
 Under 4545 and older
Female (%) 78.0% 40.0%
Age (SD) 30.2 (7.0) 59 (8.1)
Employed (%) 78.8% 39.5%
Post-age-18 education (%) 74.2% 61.5%
Experience incontinence (%) 13.6% 31.5%
Experience mobility problems (%) 13.6% 34.5%

The demographic data suggest that women in the study were generally younger than men and exhibited fewer problems in terms of experiencing incontinence and issues with their mobility.


Table 2 presents the coefficients (log odds ratios) and odds ratios for conditional logit analysis.

Table 2.  Coefficients and odds ratios for participant age groups using the conditional logit model
AttributesUnder 4545 and over
CoefficientsOdds ratiosCoefficientsOdds ratios
  • ***

    P < 0.01, **P < 0.05. 95% confidence intervals in parentheses.

Micturition −0.112***   0.894***  −0.128***   0.880***
  (−0.151–−0.073)   (0.860–0.930)  (−0.159–−0.097)   (0.853–0.908)
Incontinence −1.090***   0.336***  −0.920***   0.399***
  (−1.178–−1.003)   (0.308–0.367)  (−0.988–−0.852)   (0.372–0.427)
Urgency −0.155***   0.857***  −0.216***   0.805***
  (−0.181–−0.128)   (0.835–0.880)  (−0.238–−0.195)   (0.788–0.823)
Constipation −0.050***   0.951***  −0.044****   0.957***
  (−0.066–−0.035)   (0.936–0.966)  (−0.057–−0.032)   (0.945–0.969)
Dry mouth −0.015***   0.985***  −0.024***   0.976***
  (−0.023–−0.007)   (0.978–0.993)  (−0.030–−0.018)   (0.970–0.982)
Total respondents (A) 132 132 200 200
Choice sets per respondent (B)  18  18  18  18
Total choice sets (A x B)2376237636003600

The odds ratios for each attribute show the likelihood of an individual expressing a preference for one treatment over another after controlling for all the other attributes within a choice set. Negative coefficients (or odds ratio <1) imply that a treatment with more of an attribute would be less preferred, assuming all other characteristics are equal. In adition, the lower the odds ratio, the less is the preference for an attribute. The table shows that incontinence is the strongest determinant of preference, followed by urgency and micturition. For example, a treatment offering one additional incontinence episode per day for individuals in the under-45 group results in a decrease in the odds of preferring that treatment over an alternative by a factor of 0.336. This is significantly less attractive than a treatment that results in one additional micturition episode per day, associated with an odds decrease of 0.894. The odds ratios for the two adverse event attributes demonstrate the effect on the likelihood of expressing preference for a treatment when the probability of experiencing the adverse event increases by 1%.

Tables 3 and 4 present the marginal rates of substitution (MRS) of treatment attributes. These MRS values represent the strength of patient preference, by quantifying the trade-offs patients are willing to make to accept an increase in one level of an attribute against another. For instance, the table for patients 45 years and older suggests that patients are willing to trade-off, on average, an increase of 7.19 episodes of micturition against an increase in one episode of incontinence. The 95% confidence intervals for MRS values are presented and were calculated using the parametric bootstrap method with 100 000 repetitions.

Table 3.  Marginal rates of substitution between treatment attributes for patients under 45
 MicturitionsIncontinenceUrgencyConstipationDry mouth
  • *

    Number of episodes occurring during a 24-h period for urinary frequency (Micturitions), urge urinary incontinence (Incontinence) and urinary urgency (Urgency).

  • A 1% increase in the chance of experiencing the adverse event on treatment for ‘Constipation’ and ‘Dry mouth’.

Micturitions*0.100.72  2.22 7.59
(0.07, 0.14)(0.48, 0.99) (1.43, 3.36) (4.52, 15.61)
(7.29, 14.76)(6.09, 8.38)(16.76, 30.93)(48.82, 155.24)
Urgency*1.380.14  3.0710.48
(1.01, 2.09)(0.12, 0.16) (2.31, 4.40) (6.86, 21.67)
Constipation0.450.050.33 3.42
(0.30, 0.70)(0.03, 0.06)(0.23, 0.43) (2.10, 7.03)
Dry mouth0.130.010.10  0.29
(0.06, 0.22)(0.01, 0.02)(0.05, 0.15) (0.14, 0.48)
Table 4.  Marginal rates of substitution between treatment attributes for patients 45 and over
 MicturitionsIncontinenceUrgencyConstipationDry mouth
  • *

    Number of episodes occurring during a 24-h period for urinary frequency (Micturitions), urge urinary incontinence (Incontinence) and urinary urgency (Urgency).

  • A 1% increase in the chance of experiencing the adverse event on treatment for ‘Constipation’ and ‘Dry mouth’.

Micturitions*0.140.59  2.89 5.36
(0.11, 0.17)(0.45, 0.73) (2.11, 4.06) (3.94, 7.38)
(5.83, 9.38)(3.87, 4.69)(16.37, 28.49)(30.86, 51.47)
Urgency*1.690.24 4.89 9.07
(2.21, 1.37)(0.21, 0.26) (3.84, 6.69) (7.23, 12.10)
Constipation0.350.050.21 1.86
(0.25, 0.48)(0.04, 0.06)(0.15, 0.26) (1.32, 2.59)
Dry mouth0.190.030.11 0.54
(0.14, 0.25)(0.02, 0.03)(0.08, 0.14) (0.39, 0.76)


This report describes a study designed to understand the value of different outcomes associated with antimuscarinic therapy for overactive bladder. The study employed a DCE methodology to elicit the preferences of people with OAB in the UK.

The analysis revealed that all attributes included in the DCE survey were highly significant predictors of choice. As would be anticipated, any increase in the severity of symptoms or in the probability of adverse events being experienced was accompanied by a decrease in the likelihood that the treatment would be preferred. The attribute that exerted the most influence on treatment preference was found to be incontinence, followed by urgency, micturition, constipation and dry mouth. This is consistent with previous research that has found that incontinence has the most profound effect on the HRQL of OAB patients, in terms of both its physical and emotional impact [23]. Constipation and dry mouth show limited influence but it should be noted that these values are representative of only a 1% increase in the likelihood of the adverse event being experienced. These values can be used in the calculation of the potential impact on preference of any likelihood of experiencing such an event. For example, a treatment that offers similar efficacy in terms of symptom reduction but is 10% less likely to result in constipation will have odds of preference equal to 1.56 among those 45 years and over (and is therefore preferred 56% more often than the alternative treatment). This is highly indicative of the value placed by patients on medications that offer increased tolerability.

The values could potentially offer substantial benefit in informing treatment selection. Whilst the odds ratios provide insight into the extent of an attribute’s influence on treatment preference, the marginal rates of substitution demonstrate the typical acceptability to patients of trade-offs between attributes, and are thus indicative of their relative value. For example, the data for patients 45 and older suggest that a reduction of one urge incontinence episode per day is equivalent to a reduction in urinary urgency episodes of 4.2 per day. Alternatively, patients appear to value the experience of one urge incontinence episode per day as roughly equivalent to a 21% likelihood of experiencing constipation. This permits the direct comparison of alternative treatments in terms of the attributes investigated in this study. Whilst this cannot account for other attributes that may influence treatment preference, such as pill burden or dosing regimen for example, it can provide a clearer picture of the overall value of a treatment to patients.

Whilst OAB is a condition that increases with aging, it is present in all age groups. Many individuals view the development of urinary problems as a normal consequence of the aging process and as a result do not seek treatment [24]. An attempt was made in this study to understand the potential influence of age on the treatment preferences expressed by patients. Whilst no formal hypotheses were tested it was anticipated that there may be some disparity in preferences resulting from differing expectations of health and the variable burden placed on individuals in terms of disruption of social and work activities. The detrimental impact of OAB on social interaction and employment is well established [3]. The results revealed a remarkably consistent picture of treatment preference, with attributes being viewed similarly in terms of order of relative importance and strength by both the younger and older patient groups. Where some small differences did exist it was found first that younger patients showed a stronger aversion to the experience of urinary incontinence. This is consistent with the expectation that younger individuals may find such episodes more burdensome in terms of disruption caused to their typical day-to-day activities. Second, older patients were found to be only slightly more inclined to avoid episodes of urinary urgency than their younger counterparts (OR of 0.81 for older than 45 years vs 0.86 for under 45 years), although it was not significant enough of a difference to alter the order of preference of attributes. This may be a result of differences in mobility between the groups and the difficulties associated with reaching toilet facilities within a short timeframe.

The study does have some limitations, which may affect the interpretation of the results. The nature of stated preference methods involves deconstructing properties of a good or service into its constituent parts. Exploring complex issues such as potential treatment benefit versus risk of adverse event is a challenging task for some individuals. An attempt was made to provide clarity for participants in the attribute descriptions; however, we cannot discount that some individuals may not have fully interpreted the risk involved in the intended fashion [25], despite indicating otherwise. Furthermore, efforts were made to examine the influence of a range of participant characteristics with possible involvement in treatment choice. However, the possibility does exist that there may be other salient groupings within the data that exert a significant influence on preference. Finally, it should be noted that comparisons between the frequency of symptoms and the probability of experiencing adverse events may not accurately reflect the reality of clinical decision-making. Comparisons do, however, provide a valuable and valid mechanism for exploring the importance or influence of the attributes on treatment preference.

In conclusion, the present study aimed to understand the preferences of patients with OAB for the symptom relief and side-effects of antimuscarinic medications. It found that treatment choice was significantly influenced by all the symptoms of OAB, especially urge incontinence, with varying strengths of trade-off between the attributes, and that the strength of preferences was not significantly affected by the age of patients.


P. Swinburn and A. Lloyd are both employees of Oxford Outcomes; N. Hashmi is an ex-employee of Pfizer; D. Newal is an employee of Pfizer; H. Najib is an employee and shareholder of Pfizer. Source of Funding: Pfizer UK.