International Statistical Classification of Diseases,
9th Revision, Clinical Modification
This study sought to determine the real-worldeffectiveness of tegaserod therapy on gastrointestinal(GI)-related resource utilization in a managedcare population with a retrospective, longitudinalpre-/post-parallel cohort study of tegaserod usersand a matched reference cohort of tegaserodnonusers through medical and pharmacy claimsdata from a large, geographically diverse, managedcare organization. Continuously enrolled benefit-eligiblepatients newly initiated on tegaserod therapy(index prescription) were identified between August1, 2002, and June 30, 2003, and were categorized(using codes) ashaving irritable bowel syndrome (IBS) or another GI-relateddisorder (eg, gastroesophageal reflux disease).GI-related resource utilization (office visits,hospitalizations, emergency department visits, endoscopicand nonendoscopic procedures, and GI drugprescriptions) was determined for the 6-month periodbefore and after the index prescription date fortegaserod users and nonusers. The study populationconsisted of 3365 tegaserod users and 3364matched nonusers. Within-cohort differences beforeand after therapy were tested using the Wilcoxonsigned rank test. The mean age of 3365 tegaserodusers and 3364 matched nonusers was 47 years (+15years); 92% were women, 47% had an index diagnosisof IBS, and 53% had an index diagnosis ofanother GI-related disorder. Within-cohort GIresource utilization comparisons before and aftertherapy initiation showed significant decreases(< .01) in all utilization categories, except GI drugprescriptions, for tegaserod users; these decreaseswere not consistently observed for matched nonusers.Tegaserod use appeared to be associated with consistentdecreases in GI-related resource utilization after 6months of therapy; similarly consistent reductionswere not observed in tegaserod nonusers. These earlyfindings suggest that tegaserod may provide importantclinical and economic benefits.
(Am J Manag Care. 2005;11:S35-
Although the cause of irritable bowelsyndrome (IBS) has eluded researchersfor many years, recentadvances in clinical research have implicatedthe role of serotonin (5-hydroxytryptamine[5-HT]) in gastrointestinal (GI) motilitydisorders. Three main physiologic manifestationsof IBS lead to varied GI symptoms,including impaired GI motility, alteredintestinal secretion, and increased visceralsensitivity.1 More than one of these mechanismsmay account for a patient's symptoms,but no single factor has been shown tofully explain the pathophysiology of IBS.
Although there is no cure for IBS, treatmentstypically aimed at the relief of individualsymptoms include diet modification,changes in lifestyle, pharmacotherapeuticagents, and psychotherapy alone and incombination.2,3 In addition, none of the currentlyavailable pharmacologic agents usedto treat the individual symptoms of IBS,including antispasmodics, antidepressants,and laxatives, have been successful in managingIBS long term.4,5
One recently approved drug that targetsmultiple IBS symptoms is tegaserod maleate,a highly selective serotonin type 4 (5-HT4)receptor agonist that has been approved inthe United States for short-term use (≤12weeks) for women with IBS with constipation(IBS-C) and, more recently, for menand women younger than 65 years old withchronic idiopathic constipation.6
Interestingly, the treatment gap related toIBS remains a significant issue. A survey ofthe American Gastroenterology Associationmembership indicated that IBS accounts for12% of diagnoses made by primary carephysicians and 28% of diagnoses made bygastroenterologists.7 In addition, among gastroenterologists, IBS was found to be themost common diagnosis.8 According toanother study, in 1998 IBS was responsiblefor approximately 3.65 million office visits,500 000 hospital inpatient stays, 150 000hospital outpatient visits, and 87 000 emergencydepartment visits.9
This increased utilization of healthcareresources is also associated with a significanteconomic burden. At a national level,the estimated annual direct medical costsassociated with IBS in the United Stateshave been estimated to be as high as $10 billion.10-12 Included in these costs are primarycare and specialist physician visits, outpatientand inpatient care, and diagnostic testing,but not prescription or nonprescriptionmedication costs.9,12,13 A study of Medicaidrecipients found that expenditures wereapproximately 48% to 58% higher inpatients with IBS than in matched controls.11 A similar pattern of increasedhealthcare utilization was also observed in astudy of all adult members who had undergoneflexible sigmoidoscopy and who weresurveyed regarding IBS symptoms in a managedcare setting. This study found that IBSpatients had more outpatient visits, wereadmitted to the hospital more often, andfilled more total outpatient prescriptionsthan non-IBS patients and that their totalcosts were 51% higher.12
IBS has also been found to have a considerableimpact on indirect costs, includingdecreased work and academic productivityand reduced quality of life (QOL).2 Leongand colleagues10 found that the costs of medicallyrelated work absenteeism were significantlyhigher for employees with IBS thanfor employees without IBS. Absenteeism(days missed from work) and presenteeism(decreased productivity at work) rates werereduced 20% among bank employees withIBS and 6% among those without IBS.14,15Furthermore, other studies have shown thatthe decrease in work productivity associatedwith IBS was greater than that associatedwith many other long-term episodic diseases,including asthma16 and migraine,17and was comparable with that associatedwith the long-term persistent conditions ofhypertension18 and congestive heart failure.19 Patients with IBS have been shown toexperience significant impairment in QOL,including the performance of daily activities,20,21 and have reported that IBS symptomsnegatively affect their social lives,sexual and physical relationships,22 andschool attendance.23
Although it has been shown that tegaserodis efficacious in providing globalrelief of the multiple symptoms of IBS andthat it significantly improves patient QOLcompared with placebo,20,24 no effectivenessstudies have been conducted regarding thebenefits of tegaserod from a managed careperspective. In this study, we investigatedwhether tegaserod therapy was associatedwith a decrease in GI-related healthcareresource utilization for a cohort of tegaserodusers and a matched cohort of tegaserodnonusers. We compared GI-related healthcareutilization rates for the 6-month periodsbefore and after the actual tegaserodindex date or an assigned index date basedon user status and cohort matching criteria.
Subjects and Methods
This study was a retrospective, longitudinal,pre-/post-parallel cohort study oftegaserod users and a matched referencecohort of tegaserod nonusers. The studyincluded medical and pharmacy claims datafrom the data warehouse of a large, geographicallydiverse, managed care organization(MCO). The data warehouse containsadministrative claims for more than 14 millionmembers. Four years of current andhistorical longitudinal member-level administrativeclaims data are maintained in thewarehouse. In addition to medical and pharmacyclaims data, the data warehouseincludes such information as member demographicdata, provider data, product information,and laboratory test results. Theprestudy and poststudy design was chosento minimize confounding and selection bias.A similar analysis in a matched cohort ofnonusers was conducted to investigate possibleregression to the mean.
Classification of Diseases, 9th Revision,
Clinical Modification (ICD-9-CM)
Adults 18 yearsand older were identified as tegaserod usersif they had at least 1 pharmacy claim fortegaserod between August 1, 2002, andJune 30, 2003. The date of the firsttegaserod prescription became the indexprescription date. Benefit eligibility and continuousenrollment criteria of 6 monthsbefore and after the index prescription datewere applied. Tegaserod users were assignedan index diagnostic category of IBS or GI-relatedbased on codes onthe medical claim for an office, outpatient,or emergency department visit or an inpatienthospitalization closest (+15 days) tothe index prescription date. The patientwas assigned the index diagnostic categoryof IBS if the medical claim had an code of IBS. If there was no IBS codebut there was at least 1 code foranother GI-related disorder (eg, esophagealdisorders, gastritis/dyspepsia, gastroenteritis,gastroesophageal reflux disease, abdominalpain), the patient was assigned thediagnostic category of GI-related. The dateof the medical claim was defined as theindex diagnostic date.
Tegaserod nonusers were adults 18 yearsof age and older with 1 or more medicalclaims for office visits, outpatient visits,emergency department visits, and inpatienthospitalizations between July 15, 2002, andJuly 15, 2003. Benefit eligibility and continuousenrollment criteria during the studyperiod were applied. Medical claims wereclassified as IBS or as GI-related using themethodology that was developed for thetegaserod cohort. If there was an code for IBS on the medical claim, it becameclassified as IBS; if there was no IBS codebut there was at least 1 code for a GI-relateddisorder, the medical claim was classified asGI-related. Dates of IBS/GI-related medicalclaims were identified as index diagnosticdates. Nonusers without at least 1 IBS orGI-related medical claim were deleted fromthe analysis.
Tegaserod users were matched withnonusers 1:1 for sex, age, index diagnosticcategory, and index diagnostic date. Thematching procedure identified exact matcheson sex and index diagnostic category andthen selected the nearest possible nonusermatch based on age and index diagnosticdate. After matching, tegaserod nonuserswere assigned the index prescription date oftheir user match. The baseline evaluationperiod was defined as the 6-month periodbefore the actual index prescription date forthe tegaserod user cohort and the "assigned"index prescription date for the nonusercohort, and the follow-up evaluation periodwas defined as the 6 months including andafter the actual or assigned index prescriptiondate.
Outcome Measures and Statistical
The primary outcome measurewas GI-related resource utilization by categoryof service (ie, office visits, hospitalstays, emergency department visits, endoscopicand nonendoscopic procedures, andGI medications), as determined by the numberof claims during the 6-month periodbefore and after the index date. Each personserved as his or her own control in assessingchange before and after the actual orassigned index prescription date.
The significance of before and after differencesfor tegaserod users and nonusers wasdetermined by means of the Wilcoxonsigned rank test. SAS version 8.2 (SASInstitute, Cary, NC) was used to extract andprepare the data files, and SPSS version 11.0(SPSS, Chicago, Ill) was used to generatedescriptive analytic tables and to performstatistical calculations.
Demographic and Clinical Characteristics.
The study initiation period was tied tothe US Food and Drug Administration (FDA)approval of tegaserod in July 2002. We identified5023 tegaserod users with at least 1prescription for tegaserod between August 1,2002, and June 30, 2003. After applying benefiteligibility and continuous enrollmentcriteria, 3365 (67%) tegaserod users remained.Matching resulted in 3364 matchedpairs of tegaserod users and nonusers; 1tegaserod user had no reference groupmatch. Thus, the study population consistedof 3365 tegaserod users and 3364 matchedtegaserod nonusers.
Demographic characteristics of the 2cohorts are summarized in Table 1. Themean age of participants in each cohort was47 years + 15 years. Most (54%) patientswere between 41 and 64 years of age, and11% were 65 years or older. Ninety-two percentof participants in each cohort werewomen. In terms of the index diagnostic category,47.3% of participants in each cohortwere categorized as having a documentedIBS claim, and 52.7% were categorized ashaving a GI-related diagnostic claim.However, within the GI-related diagnosticcategory, tegaserod users had more medicalclaims with codes for intestinaldisorders than nonusers (40.8% vs 21.5%),whereas nonusers had more claims for GIand digestive disorders (39.3% vs 28.6%).
Although the tegaserod user and nonusercohorts were found to be similar in terms ofthe prevalence of individual common coexistingconditions, including hypertension,back pain, migraine, arthritis, and asthma,tegaserod users were more likely to have 1or more coexisting comorbid conditionsthan nonusers (62.4% vs 54.9%).
related Resource Utilization.
Tegaserodusers had a higher frequency of GI-relatedoffice visits, emergency departmentvisits, inpatient hospital stays, endoscopicprocedures, and nonendoscopic proceduresthan nonusers during the 6 months beforethe initiation of tegaserod therapy. Comparingusage rates before and after therapyinitiation, all GI-related resource categoriesshowed a significant and absolute decrease(<.01) for tegaserod users that was notconsistently observed for nonusers (Table 2).Figure 1 presents the relative change inresource utilization rates for tegaserod usersand nonusers. Tegaserod users had at least a20% decrease in rates for all medical GI-relatedresource utilization categories.Relative changes in rates for nonusers werenot as consistent and included a 25%decrease in emergency department use anda 24% increase in endoscopic procedure use.
related Drug Usage.
Table 3 presentsabsolute changes in GI drug prescriptionsfor tegaserod users and nonusers, and relativechange data are presented in Figure 2.Tegaserod users demonstrated significantdecreases in the mean number of prescriptionsfor antispasmodic and promotilitydrugs compared with nonusers. Mean numberof proton pump inhibitor (PPI) prescriptionsincreased significantly for tegaserodusers and nonusers alike. The relativeincrease in total mean numbers of prescriptionsbefore and after therapy was only 2%(mean absolute change, 0.04) for tegaserodusers, which was not statistically significant;however, the relative increase of 15% (meanabsolute change, 0.14) for nonusers was statisticallysignificant (<.001).
Clinical Appropriateness of Tegaserod
A subsidiary analysis related tothe clinical appropriateness of tegaserodprescribing in this managed care populationfound that 82% of tegaserod users had atleast 1 diagnostic claim consistent with IBSor the cardinal symptoms of IBS (ie, abdominalpain, bloating, or constipation) duringthe 12-month study period and that 92%were women.25 Tegaserod was prescribed for80% of users for 12 or fewer weeks and for20% of users for more than 12 weeks.
The 3365 tegaserodusers filled 7646 prescriptions betweenAugust 1, 2002 and December 31, 2003. Themean number of prescriptions per user was2.3 + 1.6 prescriptions, and the median numberof prescriptions was 2. Forty-eight percentof users filled only 1 prescription each.There was no difference between the meannumber of prescriptions for patients with theindex diagnostic category IBS and the meannumber of prescriptions for patients withthe index diagnostic category GI-related(2.30 vs 2.24). However, the mean numberof tegaserod prescriptions for women (N =3108) was significantly greater than it was formen (N = 257) (2.28 vs 2.07; = .04).
Tegaserod is a drug that was approved inthe United States for short-term use in womenwith IBS-C. In clinical trials, tegaserod hasprovided global symptom relief andimproved the QOL of patients with IBScompared with placebo.20,23 The purposeof this study was to evaluate whether theefficacy observed in clinical trials translatedinto real-world effectiveness in alarge geographically diverse managed carepopulation.
This study is the first assessment of GI-relatedhealthcare resource utilization associatedwith tegaserod in a usual care setting.Principal outcomes demonstrated statisticallysignificant decreases in all medical GI-relatedresource categories for tegaserodusers that were not consistently observed fortegaserod nonusers. In particular, a meaningfuldecrease of almost half (0.42) a physicianvisit in a 6-month period for tegaserodusers was observed.
In this study, as in other studies,7-9,13 GI-relatedhealthcare resource utilization primarilyinvolved physician office visits; therewere relatively few inpatient hospital staysor emergency department visits. Significantdifferences in resource utilization wereobserved between the tegaserod user andnonuser cohorts during the 6-month periodbefore the actual/assigned date associatedwith drug initiation. GI-related utilizationrates of tegaserod nonusers were approximately30% to 50% lower than rates oftegaserod users for all GI-related utilizationcategories, suggesting that tegaserod usersand nonusers may have been at differentstages of disease in terms of symptom severityand length of time since diagnosis. Adecrease in healthcare utilization amongpatients with more severe symptoms (ie,tegaserod users) could potentially offset thecosts of greater utilization observed in previousstudies.9,10,12,13
We found that prescription drug use wasconsistent with the types of drugs used totreat the symptoms of IBS. Drug classes withthe greatest use were antispasmodic drugs,laxatives, and PPIs. However, as with theconsumption of healthcare resources, therewere significant differences in drug usebetween the tegaserod user and nonusercohorts during the 6-month period beforetherapy, with nonusers requiring significantlyfewer drugs than tegaserod users.Significantly greater decreases in the meannumber of prescriptions for antispasmodicand promotility agents were observed fortegaserod users than for nonusers.Interestingly, PPI prescriptions increased forboth tegaserod users and nonusers. Thiswould suggest that, regardless of tegaserodtherapy, patients with IBS might also havemultiple GI-related healthcare needs.
Given the complexity of treating patientswith IBS and the specific action of tegaserod,it is important that physicians prescribe themedication as indicated and directed. A preliminaryanalysis of tegaserod users indicatedthat their treatment appeared clinicallyappropriate. Despite the lack of specificityassociated with administrative claims data,82% of patients prescribed tegaserod had atleast 1 diagnostic claim consistent with IBS orthe cardinal symptoms of IBS (ie, abdominalpain, bloating, or constipation) during the 12-month study period, and 92% were women.
In conducting this study, our challengewas to identify an appropriate comparisongroup for our cohort of tegaserod users.Although our constructed reference groupprovided a frame of reference, it might nothave been a true control group related to thepropensity to have potential requirementsfor tegaserod use. Patients whose physiciansprescribed tegaserod might have differedclinically from those whose physicians didnot in ways that could have independentlyaffected healthcare utilization and thereforebiased the observed results. In addition, theseverity and chronicity of the disease andthe existence of other comorbidities werenot tightly controlled within this studyapproach.
Another limitation relates to the timing ofthis study and to the commercial launch oftegaserod. This product was approved by theFDA in July 2002 for use in women withIBS-C and became commercially available inAugust 2002. Initiating the study so quicklyafter product launch might have resulted insome physicians prescribing tegaserod topatients with severe symptoms. Moreover,the follow-up period was not long enough toallow for long-term evaluation of healthcareutilization over the natural progression ofthe disease and its management. For example,office visits might have been promptedby many conditions other than IBS alone. Inthis study, 48% of tegaserod users filled only1 prescription each, which minimized theduration of exposure during follow-up.Consequently, the effect on overall utilizationmay be overstated.
Most important, evaluating this agent atthe time of launch would have been associatedwith high product sampling, affectinguse and our ability to study causal relationshipsrelated to therapy. In addition, wecould not control for the use of other therapeuticremedies, including lifestyle modifications,over-the-counter remedies, andalternative therapies. Although no otherstudies have measured healthcare resourceutilization in tegaserod users, these findingsbuild on earlier research around QOLissues to extend the impact of this medicationon patients with IBS and GI-relateddisorders.
In this study population, tegaserod therapywas associated with consistent decreasesin all GI-related resource utilization categories(office visits, hospital stays, emergencydepartment visits, endoscopicprocedures, and nonendoscopic procedures)after 6 months of therapy. These early findingssuggest that tegaserod may provideimportant clinical and economic benefits forthe population of patients for whom it isindicated. Additional research is necessaryto compare findings in other MCO settingsand to substantiate the longer-term implicationsof these findings, including an economicanalysis to understand the impact ofcost offsets to payers and patients.
Am J Manag Care.
1. Crowell MD. The role of serotonin in the pathophysiologyof irritable bowel syndrome. 2001;7(suppl):S252-S260.
2. Drossman DA, Camilleri M, Mayer EA, WhiteheadWE. AGA technical review on irritable bowel syndrome.2002;123:2108-2131.
3. Brunton SA. Treatment options for patients with IBSwith constipation? 2004;6:1-6.
Am J Gastroenterol.
4. Brandt LJ, Bjorkman D, Fennerty MB, et al.Systematic review on the management of irritable bowelsyndrome in North America. 2002;97(suppl):S7-S26.
Am J Med.
5. Jackson JL, O'Malley PG, Tomkins G, Balden E,Santoro J, Kroenke K. Treatment of functional gastrointestinaldisorders with antidepressant medications: ameta-analysis. 2000;108:65-72.
Am J Gastroenterol.
6. American College of Gastroenterology FunctionalGastrointestinal Disorders Task Force. Evidence-basedposition statement on the management of irritable bowelsyndrome in North America. 2002;97(suppl):S1-S5.
7. Mitchell CM, Drossman DA. Survey of the AGAmembership relating to patients with functional gastrointestinaldisorders. 1987;92:1282-1284.
8. Everhart JE, Renault PF. Irritable bowel syndrome inoffice-based practice in the United States. 1991;100:998-1005.
Burden of Gastrointestinal Diseases.
9. American Gastroenterological Association. Bethesda, Md:American Gastroenterological Association; 2001:1-86.Available at: http://www.gastro.org/clinicalRes/pdf/burden-report.pdf. Accessed November 16, 2004.
Arch Intern Med.
10. Leong SA, Barghout V, Birnbaum HG, et al. Theeconomic consequences of irritable bowel syndrome: aUS employer perspective. 2003;163:929-935.
11. Martin BC, Ganguly R, Pannicker S, Frech F,Barghout V. Utilization patterns and net direct medicalcost to medicaid of irritable bowel syndrome. 2003;19:771-780.
Am J Gastroenterol.
12. Longstreth GF, Wilson A, Knight K, et al. Irritablebowel syndrome, health care use, and costs: a US managedcare perspective. 2003;98:600-607.
Am J Manag Care.
13. Martin R, Barron JJ, Zacker C. Irritable bowel syndrome:toward a cost-effective management approach.2001;7(suppl):S268-S275.
Am J Manag
14. Dean BB, Aguilar D, Barghout V, et al. Impairmentin work productivity and health-related quality of life inpatients with irritable bowel syndrome. 2005;11(suppl):S17-S26.
Health Productivity Management.
15. Groves D. The impact of irritable bowel syndromein the workplace: a study to improve health and productivity.2003;2:8-10.
Sheets: Facts About Asthma.
16. Centers for Disease Control and Prevention. Atlanta, Ga: Department ofHealth and Human Services; August 8, 1997.
Arch Intern Med.
17. Hu XH, Markson LE, Lipton RB, Stewart WF,Berger ML. Burden of migraine in the United States: disabilityand economic costs. 1999;159:813-818.
Heart and Stroke
18. American Heart Association. Dallas, Tex: American HeartAssociation; 1998.
Heart Disease and
Stroke Statistics—2004 Update.
19. American Heart Association. Dallas, Tex: AmericanHeart Association; 2003:1-48.
Am J Gastroenterol.
20. Reilly MC, Barghout V, Ruegg P, Pecher E, Ricci JF.Tegaserod significantly reduces work productivity lossand daily activity impairment in patients with IBS withconstipation [abstract]. 2004;99:S241.
Am J Gastroenterol.
21. Hungin APS, Tack J, Mearin F, Whorwell PJ,Dennis E, Barghout V. Irritable bowel syndrome (IBS):prevalence and impact in the USA—the truth in IBS (TIBS)survey [abstract]. 2002;97(suppl):S280-S281.
Bowel Syndrome in Women: the Unmet Needs.
22. Society for Women's Health Research. Washington, DC: Society for Women's Health Research;2002:1-37.
23. Talley NJ, Weaver AL, Zinsmeister AR, Melton LJ3rd. Onset and disappearance of gastrointestinal symptomsand functional gastrointestinal disorders. 1992;136:165-177.
24. Nyhlin H, Bracco A, Wagner A. Tegaserodimproves quality of life of patients with non-diarrheairritable bowel syndrome: outcomes of the TENOR(TEgaserod in NORdic countries) study. Presented at:69th Annual Scientific Meeting of the American Collegeof Gastroenterology; October 29-November 3, 2004;Orlando, Fla.
Am J Gastroenterol.
25. Stephenson JJ, Fernandes J, Joo S, et al. Tegaserodtherapy reduces GI-related resource utilization [abstract].2004;99:S239-S240.