Objective: The objective of this study was tocompare 1-year total healthcare costs for patientswith overactive bladder (OAB) initiating treatmentwith extended-release formulations of tolterodineand oxybutynin: tolterodine tartrate extended-releasecapsules (tolterodine ER) versus extended-releaseoxybutynin chloride (oxybutynin ER).
Methods: A model was developed from the payerperspective using data from the PharMetrics Patient-Centric database. Monthly discontinuation rates werederived from a cohort of newly treated patients withOAB (tolterodine ER, n = 15 394 or oxybutynin ER,n = 7934). All were assumed to be receiving therapyfor at least 1 month. Medical management costs werebased on reimbursement for all services for amatched cohort of patients taking tolterodine ER andoxybutynin ER. Medical management costs for thosediscontinuing therapy were based on patients receivingOAB care without pharmacotherapy (n = 29992). Drug costs were from AnalySource (December2004).
Results: After the 11-month follow-up period,21% of patients taking tolterodine ER and 15% ofpatients taking oxybutynin ER remained on originaltherapy. One-year average total costs per patient forthose started on tolterodine ER were $8876 and$9080 for oxybutynin ER, a difference of $204 peryear. Sensitivity analyses indicated results wererobust to changes in drug cost and probability of discontinuation.When discontinuation rates were heldequal, cost differences continued to favor tolterodineER (21%, $272/yr; 15%, $233/yr).
Conclusion: Those taking tolterodine ER hadlower monthly drug and medical management costs.This resulted in a total average annual cost savingsof $204 per patient for those started on tolterodineER. At the end of 1 year, patients with OAB weremore likely to remain on original drug treatment takingtolterodine ER versus oxybutynin ER.
(Am J Manag Care. 2005;11:S150-S157)
Overactive bladder (OAB) is definedby the presence of urgency (with orwithout urge incontinence), usuallywith frequency and nocturia, and theabsence of pathologic or metabolic conditionsthat might explain these symptoms.Recent figures from the NationalOveractive BLadder Evaluation (NOBLE)program show that the overall prevalenceof OAB is 16.5% in the United States—representingmore than 34 million men andwomen.1,2 Studies indicate that the prevalenceof OAB in European nations maybe similar: a survey of 16 776 subjectsaged 40 years or older found the prevalenceof OAB to be 16.6% across 6 Europeancountries.3
In addition to the intangible costs (ie, theemotional and physical toll that OAB takeson individuals), the economic burden of thisdisease is considerable. The direct and indirectcosts of OAB in 2000 were estimatedto total $12.02 billion in the UnitedStates.2 A recent study analyzed US claimsdata from privately insured patients 18 to64 years of age. Compared with patientswithout OAB, patients with OAB had significantlyhigher average total claim costs($1767 vs $5018, 1995 dollars, respectively;= .0001), a significantly higher probabilityof hospital admission (7.0% vs 20.6%, respectively;= .0001), and a significantly higheraverage length of stay once admitted (4.8 vs6.8 days, respectively; = .0001) for thestudy period.4 Other studies indicate thatcosts can be decreased when OAB is properlydiagnosed and treated.5,6
Clinicians and payers need comparativeinformation to support their clinical andformulary decision making. To respond tothis need, a number of pharmacoeconomicmodels have been used to illustratecomparisons among antimuscarinic treatmentsfor OAB. Earlier studies havefocused on immediate-release formulations; and some compared immediate-withextended-release products.7-13
Several recently published models havecompared the extended-release formulationswith each other because these treatmentsare now more commonly used. In one suchmodel incorporating more than 30 resourceand cost inputs, the average 3-month overalltreatment costs for tolterodine tartrateextended-release capsules (tolterodine ER)therapy were 6% less than that of oxybutyninchloride extended release (oxybutynin ER)therapy ($1207 vs $1283) for OAB patientsinitiating long-acting drug therapy in a primarycare setting.14 This model included estimatesof incontinence pad use and lostproductivity based on published data.
A cost-effectiveness model developed byHughes and Dubois compared the extendedreleaseforms of tolterodine and oxybutynin.15This study included patient-level data onincontinence episodes and pad use from clinicaltrials and from the literature, incorporatingpersistence and adverse effects (eg, drymouth or central nervous system sideeffects). Both extended-release formulationswere found to be cost effective, and their rankordering in the results were influenced by"uncertainty surrounding the health and costconsequences of early discontinuations."15
Economic models of OAB therapy areoften developed so managed care organizations(MCOs) can compare treatments toassist in formulary decision making andinput their own plan-specific informationinto the model. However, in the majority ofcircumstances, most MCOs do not have thespecific clinical information needed for themodel inputs. For example, the number ofincontinence episodes experienced by apatient is often an important input in OABeconomic models, but the majority of MCOsdo not have systems in place to accuratelytrack such variables. Moreover, most plansdo not cover certain costs, such as the costof incontinence pads, diapers, or the cost oflaundry. Thus, the costs associated with theutilization of such materials would likely notfactor into their coverage decision making.Thus, it is unlikely that a typical insurer orMCO has the clinical inputs required to customizethese models, nor would they begreatly concerned with some of these costs.
There are, however, a number of potentialmodel inputs that insurers can trackeasily within their population and that havean impact on reimbursement. For example,payers know their drug costs for pharmacyclaims, and most can track how long anenrollee is persistent with therapy, or if thepatient discontinues the drug. An OAB economicmodel that incorporates these typesof inputs is, therefore, not only more relevantto the insurer in terms of reimbursementcosts, but is more easily customizableand can be used to more accurately reflectthe unique experiences of a given insurer'sOAB population. The objective of this studywas to compare tolterodine ER and oxybutyninER in a cost-minimization model thatis not only user friendly, but also includesdata readily available to most payers asinputs into the model.
A decision-analysis modelwas developed using Microsoft Excel comparingthe 12-month economic impact oftreating OAB. The diagram in the Figuredepicts a schematic of the decision-analysismodel used. Because they are the most widelyprescribed drugs for OAB, the comparisonis limited to the long-acting forms of tolterodineand oxybutynin (tolterodine ER andoxybutynin ER). The model takes thepayer's perspective and a cost-minimizationapproach; treatment outcomes are assumedto be the same for patients who persist withtherapy. Because the frequency of therapyswitching has been found to be relatively low(~6%),16 it was assumed that when treatmentwas discontinued, patients were medicallymanaged without pharmacotherapy. Discontinuationof initial therapy (which alsoincludes patients who switched from therapy)was accounted for on a monthly basis.
Medication costs and medical managementcosts are included. Medication costsare calculated using the average wholesaleprice (AWP). The AWP for both drugs wasobtained in December of 2004 from Analy-Source, an online vendor of drug pricingdata.17 The cost for tolterodine ER was basedon Detrol LA® 4 mg and was assumed to be$3.31 per dose, or $99.30 per month.Oxybutynin ER cost was based on DitropanXL® 10 mg and was assumed to be $3.38 perdose, or $101.40 per month.
International Classification of Diseases,
Ninth Revision, Clinical Modification
Procedural Terminology 4
Common Procedures Coding System
In addition to the costs associated withOAB pharmacotherapy, the model alsoincorporates costs for medical management.These include the costs of non-OAB-relatedpharmacy services, as well as the cost ofinpatient, outpatient, and emergency servicesused by OAB patients over the course of1 year. Both medical management costs andthe rates of OAB therapy discontinuationwere obtained from the PharMetrics Patient-Centric database, which is composed of fullyadjudicated medical and pharmacy claimsfor more than 55 million covered lives frommore than 75 health plans across the UnitedStates.18 The PharMetrics database includesboth inpatient and outpatient diagnoses (informat)and procedures (in and formats),as well as both standard and mailorderprescription records. Data onprescription records include the NationalDrug Code, as well as "days supplied" and"quantity dispensed" fields.
Records in the PharMetrics database arerepresentative of the national managed carepopulation based on a variety of patient andhealth plan demographic measures, includinggeographic region, age, sex, and plantype. The data are also longitudinal, with anaverage member enrollment time of 2 years.Only health plans submitting data for allmembers are included in the database,ensuring complete data capture and unbiasedsamples.
The sample population for this study consistedof all patients with either a new diagnosisof OAB or new use of 1 of themedications of interest (tolterodine or oxybutyninin any form) between January 1,2001, and December 31, 2002. The date offirst OAB diagnosis or medication use servedas the patient's index date. All patients wererequired to have pre-index and follow-upperiods of 12 months each in relation to thisdate. Claims for selected patients were thenaccumulated spanning the period fromJanuary 1, 2000, to December 31, 2003. Itwas found that 80% of all pharmacologicallytreated patients with OAB were usingtolterodine ER or oxybutynin ER, with theremaining 20% taking one of the immediate-releaseformulations.
A propensity score was calculated for eachpatient with OAB. The estimated propensityscores were intended to represent a particularpatient's probability for receiving a giventreatment option and were calculated bysumming coefficient values for a list of potentialconfounding variables. Use of thesescores confers the advantage of having a singleestimate available to adjust for confoundingin a multivariate analysis. In thiscase, the outcome of interest was thepropensity for treatment with any of theselected medications for OAB.
Using these data, 3 cohorts were createdfor deriving the model inputs:
All patients taking tolterodineER and oxybutynin ER, selected as describedabove, were included in this cohort. Thecohort included 15 394 treated with tolterodineER and 7934 patients treated with oxybutyninER. Using these data, monthlydiscontinuation rates were calculated for eachdrug group (Table 1).
From among those in Cohort 1,the tolterodine ER and oxybutynin ERpatients were matched based on the estimatedpropensity score for OAB pharmacotherapy tocreate a matched cohort. After matching, acohort of 7257 tolterodine ER and 7257 oxybutyninpatients resulted. The average annualmedical costs for each treatment group werethen calculated. Costs were derived from theamounts paid by the insurer for services rendered.These costs were included as the medicalmanagement costs for these patients,excluding the OAB drug treatment costs(Table 2).
All patients who were treatedwith any OAB medication were matched basedon the estimated propensity score to a controlgroup of patients with OAB who were notreceiving any OAB pharmacotherapy. Aftermatching, there were 29 992 patients in theOAB-drug-treated group and 29 992 patients inthe medically managed group (no pharmacotherapy).The individuals not receivingOAB pharmacotherapy were considered to bea "medically managed" group, and the averagecost of medical treatment for this groupwas used to account for monthly medicalcosts for patients who discontinued treatmentwith a study drug for the remainder of thestudy period (Table 2).
All patients who receivedtolterodine ER or oxybutynin ER wereassumed to continue receiving treatment forthe first full month. This assumption wasmade because the costs for care and the costsfor the drug will have been incurred by thepayer, regardless of whether or not thepatient discontinues pharmacotherapy shortlyafter initiation. All patients were followedfor 11 months after the initial treatmentmonth, for a total 1-year follow-up period. Ifthe original therapy was discontinued, costsfor the remainder of the study period wereassumed to be those for medical management(the nonpharmacotherapy group).
case and Sensitivity Analysis.
Thebase case examined the average cost pertreatment group that resulted when the discontinuation rates from the PharMetricsdatabase were used. In one-way sensitivityanalyses, the medical management costswere varied to identify the increase ordecrease needed for parity. One-way sensitivityanalyses also were used to estimatethe increase or decrease needed in the drugAWP for total average costs to be the samefor both groups.
The per-month cost calculations for thebase-case analysis are presented in Table 3.This analysis results in total average costsper patient per year of $8876 for patientsstarted on tolterodine ER and $9080 forpatients started on oxybutynin ER, a differenceof $204 per patient per year, or $17 perpatient per month. Sensitivity analysis indicatedthat when all other costs are held atthe baseline assumptions, the cost of oxybutyninER must decrease from $3.38 to$1.88 per month for the average annualcosts to be equal for both groups; or the costof tolterodine ER must increase from $3.31per dose to $4.57 per dose for the averagecost to be equal.
When persistence rates were held equalacross the 2 treatments, cost differencescontinued to favor tolterodine ER. If the discontinuationrates for tolterodine ER wereused for both treatments, the cost differencewas $272 per patient per year; when rates foroxybutynin ER were used for both treatments,the cost difference was $233 perpatient per year. When the medical managementcosts were varied, tolterodine ER medicalcosts must increase from $662 to $700per month for total costs to be equal.Oxybutynin ER medical costs must decreasefrom $712 to $677 per month for paritybetween treatments. When medical managementcosts for both drug groups were heldconstant at either $662 or $712 per month,the annual cost difference favored oxybutyninER at $20 and $50 per patient per year,respectively.
The model results indicate that whenOAB therapy is initiated with tolterodineER, there is an annual cost advantage to thepayer over initiating therapy with oxybutyninER. The advantage stems from boththe monthly OAB drug costs and monthlymedical management costs being slightlylower for tolterodine ER. Additionally, therapydiscontinuation rates over the courseof the first year were also slightly lower withtolterodine ER. The relative advantageafforded by superiority of tolterodine ERacross these 3 measures results in a costadvantage over the first year of therapy thatis robust to changes in the model's assumptionsregarding inputs.
The results of this study are similar tothose found by Noe and colleagues wherethe extended-release formulations of tolterodineand oxybutynin were compared.14 Noe'scost-minimization model assumed equaleffectiveness across treatments and used acombination of published data and expertopinion to construct a detailed model ofresource use. The authors estimated that 3-month treatment costs were $1207 fortolterodine ER-treated patients and $1283for oxybutynin ER-treated patients. If simplymultiplied by 4 to obtain an annual estimate,this would indicate that $4828 wouldbe spent for a tolterodine ER-treated patientand $5131 would be spent for an oxybutyninER-treated patient over 1 year, an annualdifference of $304. Although the differencein the Noe annual estimate and the annualestimate presented here is relatively large, itshould be noted that the claims data used inthis analysis indicate that the majority oftreatment discontinuations occur in the first3 months after initiation of therapy; thus,the simple multiplication of the Noe 3-month cost is an overestimation of the truecost. However, it does provide a crude basisfor comparison.
Additionally, the lower total cost estimatepresented by Noe and colleagues may beexplained by 2 key methodological differences.Noe and colleagues' cost estimatesare for OAB-related treatment only andincluded treatments for urinary tract infection,skin infections, depression, and fractures,because these conditions are oftenassociated with OAB. However, the total costin the model presented here included costsfor all healthcare services (eg, inpatient, outpatient,emergency department), not justOAB-related services. Moreover, becausemany of Noe and colleagues' cost inputswere based on claims submitted toCalifornia's Medicaid program (Medi-Cal), itwould be expected that total cost estimatespresented by Noe would be lower than thosefound in this study, which had costs basedon reimbursements by private insurersmade during 2001 and 2002.
In the Hughes and Dubois model, incrementalcost effectiveness was calculatedconsidering rates of treatment persistenceand adverse events obtained from clinicaltrial data.15 The authors explain thatalthough both tolterodine ER and oxybutyninER were cost effective, their relativeranking as "more" cost effective variedunder different scenarios because of uncertaintyaround the consequences of discontinuation.Compared with tolterodine ER,oxybutynin ER was associated with an incrementalcost per incontinent-free week thatranged from £84.82 to being dominatedwhen full persistence was assumed in thesensitivity analysis.
In sensitivity analyses, the current modelwas found to be robust to changes in drugpricing and discontinuation rates.Tolterodine ER costs had to increase by 38%to $4.57 per dose for parity to exist betweentreatments. Oxybutynin ER costs had todecrease by 55% from $3.38 to $1.88 perdose for parity to exist between treatments.Tolterodine ER results remained favorablewhen discontinuation rates were made equalusing either 15% or 21% discontinuation.Small differences were found when medicalmanagement costs were held equal.When costs were held equal at $662 permonth for both groups, a difference of $20per patient per year, or less than $2 perpatient per month, results, favoring oxybutyninER. When costs were held at $712 permonth for both groups, a difference of $50per patient per year, or less than $5 perpatient per month, results, again favoringoxybutynin ER. This indicates that thesemedical management costs can influencethe total average costs in important ways.Robust methods must be used in estimatingthese costs, as was done here by using apropensity score matched cohort. The $50per month advantage in medical management costs that is seen with tolterodine ER(at $662 vs $712 per month) contributes toa $204 total advantage for the year. Themonthly medical management cost differencemust be reduced to approximately zerofor the total annual advantage to disappear.
It is important to recognize that themodel has several limitations that couldhave an impact on its use and the resultspresented. The methodology used in thisanalysis is such that the efficacy of treatmentsis assumed to be the same. Althoughincluding discontinuation rates in the modelmay serve to account for some differences inefficacy (assuming that lack of efficacy isassociated with treatment discontinuation),discontinuation alone cannot account fordifferences in other treatment outcomes.However, other economic models have alsoassumed equal efficacy in their analyses.
The model does not take into considerationhumanistic outcomes, such as patient-reportedoutcomes or measures of quality of life,which would capture aspects as the socialstigma or embarrassment generated by thecondition. Because patients with OAB havereported that their condition exerts tremendouspsychological burden,19 if there are differencesin humanistic outcomes betweentolterodine ER and oxybutynin ER, these differencescould have an impact on the relativecost effectiveness of these long-acting agents.Moreover, the model does not account for differencesin productivity associated with treatments,which also could have an impact ontotal cost or cost effectiveness.
Finally, the model is constructed at thetreatment level, and does include clinical orpatient-level information. More detailed dataon the experiences of the patient whilereceiving OAB therapy could prove useful toa more accurate estimation of treatmentcosts, but because the cost of resources suchas incontinence pads or the cost of frequentclothes laundering are not covered by managedcare payers, these costs are not relevantgiven the perspective of this analysis.
A user-friendly model was developed thatallows MCOs to use data that are accessibleto them (eg, drug costs and drug discontinuationrates) to compare OAB treatments initiatedwith tolterodine ER or oxybutynin ER.Base-case data from a large claims databaseindicate that patients with OAB were morelikely to remain on original treatment at theend of 1 year when treated with tolterodineER versus oxybutynin ER. When combinedwith lower drug and medical managementcosts, this resulted in a total annual cost differenceof $204 per patient. This model canbe adapted by payers to make formularymanagement decisions among therapiesbased on knowledge of their own patientpopulation.
World J Urol.
1. Stewart WF, Van Rooyen JB, Cundiff GW, et al.Prevalence and burden of overactive bladder in theUnited States. 2003;20:327-336.
2. Hu TW, Wagner TH, Bentkover JD, et al. Estimatedeconomic costs of overactive bladder in the UnitedStates. 2003;61:1123-1128.
3. Milsom I, Abrams P, Cardozo L, Roberts RG, ThuroffJ, Wein AJ. How widespread are the symptoms of anoveractive bladder and how are they managed? A population-based prevalence study. 2001;87:760-766.
Drug Benefit Trends.
4. Zhou Z, Jensen GA. Insurance claims costs for overactivebladder disorder. 2001;13:45-58.
5. Muchmore L, Williamson TE, Gardner H. Incidenceand cost associated with pharmacological treatment foremployees diagnosed with urinary and overactive bladderdisorders. 2002;5:275-276.
6. Bailey KL, Torigoe YR, Zhou S, et al. Overactivebladder cost of illness: analysis of Medi-Cal claims.2000;3:139-140.
Manag Care Interface.
7. Arikian SR, Casciano J, Doyle JJ, Tarride JE,Casciano RN. A pharmacoeconomic evaluation of twonew products for the treatment of overactive bladder.2000;13:88-94.
8. Becker RV, Noe LL, Trotter JP, et al. A cost-effectivenesscomparison of extended release tolterodine (Detrol®LA) versus controlled release oxybutynin (Ditropan® XL).Poster presented at: 80th Annual Meeting of the SouthCentral Section of the American Urological Association;September 12, 2001; Austin, Tex. Poster 236.
9. Bentkover J, Bachleda M, Kulseng-Hansen S, et al.An economic model of overactive bladder in Norway.2002;5:275.
10. Chapple C, Hill S, Corey R, et al. An economicmodel of unstable bladder in the United Kingdom.Poster presented at: Second International Consultationon Incontinence; July 1-3, 2001; Paris, France.Poster 77.
11. Kobelt G, Johnson L, Mattiasson A. Cost effectivenessof new treatments for overactive bladder: the exampleof tolterodine, a new muscarinic agent: a Markovmodel. 1998;17:599-611.
12. Getsios D, Caro JJ, Ishak KJ, El-Hadi W, Payne K.Canadian economic comparison of extended-releaseoxybutynin and immediate-release tolterodine in thetreatment of overactive bladder. 2004;26:431-438.
13. O'Brien BJ, Goeree R, Bernard L, Rosner A,Williamson T. Cost-effectiveness of tolterodine for patientswith urge incontinence who discontinue initialtherapy with oxybutynin: a Canadian perspective. 2001;23:2038-2049.
J Manag Care Pharm.
14. Noe L, Becker R, Williamson T, Chen D. A pharmacoeconomicmodel comparing two long-acting treatmentsfor overactive bladder. 2002;8:343-352.
15. Hughes DA, Dubois D. Cost-effectiveness analysis ofextended-release formulations of oxybutynin and tolterodinefor management of urge incontinence. 2004;22:1047-1059.
16. PharMetrics. Pfizer OAB Draft Technical Report. Onfile, 2004.
17. AnalySource. AnalySource Online. Available at:http://www.analysource.com. Accessed December 2004.
18. PharMetrics. Overview. Available at: http://www.pharmetrics.com/h_overview.html. Accessed February24, 2005.
19. Chiaffarino F, Parazzini F, Lavezzari M, et al.Impact of urinary incontinence and overactive bladderon quality of life. 2003;43:535-538.