5
RESPONSIVENESS OF SYMPTOM SCALES FOR INTERSTITIAL CYSTITIS K. J. PROPERT, R. D. MAYER, Y. WANG, G. R. SANT, P. M. HANNO, K. M. PETERS, J. W. KUSEK, AND THE INTERSTITIAL CYSTITIS CLINICAL TRIALS GROUP ABSTRACT Objectives. To evaluate the responsiveness of composite scales to change over time in a clinical trial of patients with interstitial cystitis (IC). The measurement of symptoms in IC includes the O’Leary-Sant Symptom and Problem Indexes and the University of Wisconsin Interstitial Cystitis Inventory and scales that measure the individual symptom domains of pain/discomfort, urgency, and voiding frequency. Methods. The data were derived from a randomized clinical trial conducted by the Interstitial Cystitis Clinical Trials Group. Participants met the National Institutes of Health-National Institute for Diabetes, and Digestive and Kidney Diseases criteria for IC and reported at least moderate pain and frequency. The primary endpoint was a patient-reported global response assessment (GRA) at 24 weeks. Secondary endpoints included the three composite indexes, pain/discomfort and urgency, and 24-hour frequency. Responsiveness was assessed by comparing symptom score changes against response categories defined by the GRA. Results. Of the 121 subjects in the original trial, 94 with complete data were included. All three composite indexes were sensitive to subject improvement over time as measured by the GRA. A 1.2-point change in the O’Leary-Sant indexes and a 3.1-point change in the Wisconsin IC inventory corresponded to a one-category change in the GRA. Individual symptoms were also responsive. The correlation was high among the changes in the six outcome measures. Conclusions. The three composite symptom scales are responsive to change over time in patients with IC. These indexes provide important insight into symptom changes and are recommended as secondary endpoints in future clinical trials of IC. Additional endpoints addressing individual symptom domains should also be considered to aid in the evaluation of effect mechanisms. UROLOGY 67: 55–59, 2006. © 2006 Elsevier Inc. I nterstitial cystitis (IC) is a chronic debilitating con- dition, characterized by pelvic/bladder/perineal pain, urgency, and urinary frequency. The etiology of IC is unknown and, although recent research has identified urine markers that may eventually lead to a laboratory test for the disorder, 1 the diag- nosis and outcome assessment are currently based on symptoms. The most commonly used instru- ments to evaluate overall symptoms in IC are the O’Leary-Sant IC Symptom (ICSI) and O’Leary-Sant IC Problem (ICPI) Indexes and the Wisconsin IC Inventory (WICI). 2,3 The two O’Leary-Sant indexes were designed to assess symptoms and their impact on patients’ quality of life, respectively. 2 The WICI is based on seven questions related to urinary symp- toms that are embedded in a longer questionnaire, including other body systems. 3 Measures for the individual domains of pain, urgency, and frequency are also used as outcomes in IC research. 4–6 The three composite scales have been previously validated. 2,3,7 The specific types of validity evalu- ated and methods used differed among the reports but included such psychometric properties as in- ternal consistency, test-retest reliability, and con- struct validity (eg, whether a scale measures what A complete list of the participating investigators is provided in the Appendix. This study was supported by cooperative agreements U01 DK54108, U01 DK54125, U01 DK54127, U01 DK54133, U01 DK54138, and U01 DK54158 from the National Institute for Diabetes and Digestive and Kidney Diseases, National Institutes of Health. From the University of Pennsylvania School of Medicine, Phil- adelphia, Pennsylvania; University of Rochester, Rochester, New York; Sanofi-Synthelabo Incorporated, New York, New York; William Beaumont Hospital, Royal Oak, Michigan; and National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland Reprint requests: Kathleen Joy Propert, Sc.D., Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Blockley Hall, 6th Floor, 423 Guardian Drive, Phila- delphia, PA 19104-6021. E-mail: [email protected] Submitted: April 1, 2005, accepted (with revisions): July 11, 2005 ADULT UROLOGY © 2006 ELSEVIER INC. 0090-4295/06/$32.00 ALL RIGHTS RESERVED doi:10.1016/j.urology.2005.07.014 55

Responsiveness of symptom scales for interstitial cystitis

Embed Size (px)

Citation preview

Page 1: Responsiveness of symptom scales for interstitial cystitis

OpSmMCDeiwRiOciCTic

Ipo

AA

DDDo

aYWII

Bod

2

ADULT UROLOGY

©A

RESPONSIVENESS OF SYMPTOM SCALES FORINTERSTITIAL CYSTITIS

K. J. PROPERT, R. D. MAYER, Y. WANG, G. R. SANT, P. M. HANNO, K. M. PETERS, J. W. KUSEK,AND THE INTERSTITIAL CYSTITIS CLINICAL TRIALS GROUP

ABSTRACTbjectives. To evaluate the responsiveness of composite scales to change over time in a clinical trial ofatients with interstitial cystitis (IC). The measurement of symptoms in IC includes the O’Leary-Santymptom and Problem Indexes and the University of Wisconsin Interstitial Cystitis Inventory and scales thateasure the individual symptom domains of pain/discomfort, urgency, and voiding frequency.ethods. The data were derived from a randomized clinical trial conducted by the Interstitial Cystitis

linical Trials Group. Participants met the National Institutes of Health-National Institute for Diabetes, andigestive and Kidney Diseases criteria for IC and reported at least moderate pain and frequency. The primaryndpoint was a patient-reported global response assessment (GRA) at 24 weeks. Secondary endpointsncluded the three composite indexes, pain/discomfort and urgency, and 24-hour frequency. Responsivenessas assessed by comparing symptom score changes against response categories defined by the GRA.esults. Of the 121 subjects in the original trial, 94 with complete data were included. All three composite

ndexes were sensitive to subject improvement over time as measured by the GRA. A 1.2-point change in the’Leary-Sant indexes and a 3.1-point change in the Wisconsin IC inventory corresponded to a one-categoryhange in the GRA. Individual symptoms were also responsive. The correlation was high among the changesn the six outcome measures.onclusions. The three composite symptom scales are responsive to change over time in patients with IC.hese indexes provide important insight into symptom changes and are recommended as secondary endpointsn future clinical trials of IC. Additional endpoints addressing individual symptom domains should also beonsidered to aid in the evaluation of effect mechanisms. UROLOGY 67: 55–59, 2006. © 2006 Elsevier Inc.

hlnomOIIwoitiia

vabt

nterstitial cystitis (IC) is a chronic debilitating con-dition, characterized by pelvic/bladder/perineal

ain, urgency, and urinary frequency. The etiologyf IC is unknown and, although recent research

complete list of the participating investigators is provided in theppendix.This study was supported by cooperative agreements U01K54108, U01 DK54125, U01 DK54127, U01 DK54133, U01K54138, and U01 DK54158 from the National Institute foriabetes and Digestive and Kidney Diseases, National Institutesf Health.

From the University of Pennsylvania School of Medicine, Phil-delphia, Pennsylvania; University of Rochester, Rochester, Nework; Sanofi-Synthelabo Incorporated, New York, New York;illiam Beaumont Hospital, Royal Oak, Michigan; and National

nstitute of Diabetes, Digestive and Kidney Diseases, Nationalnstitutes of Health, Bethesda, Maryland

Reprint requests: Kathleen Joy Propert, Sc.D., Department ofiostatistics and Epidemiology, University of Pennsylvania Schoolf Medicine, Blockley Hall, 6th Floor, 423 Guardian Drive, Phila-elphia, PA 19104-6021. E-mail: [email protected]

Submitted: April 1, 2005, accepted (with revisions): July 11,

s005

2006 ELSEVIER INC.LL RIGHTS RESERVED

as identified urine markers that may eventuallyead to a laboratory test for the disorder,1 the diag-osis and outcome assessment are currently basedn symptoms. The most commonly used instru-ents to evaluate overall symptoms in IC are the’Leary-Sant IC Symptom (ICSI) and O’Leary-Sant

C Problem (ICPI) Indexes and the Wisconsin ICnventory (WICI).2,3 The two O’Leary-Sant indexesere designed to assess symptoms and their impactn patients’ quality of life, respectively.2 The WICIs based on seven questions related to urinary symp-oms that are embedded in a longer questionnaire,ncluding other body systems.3 Measures for thendividual domains of pain, urgency, and frequencyre also used as outcomes in IC research.4–6

The three composite scales have been previouslyalidated.2,3,7 The specific types of validity evalu-ted and methods used differed among the reportsut included such psychometric properties as in-ernal consistency, test-retest reliability, and con-

truct validity (eg, whether a scale measures what

0090-4295/06/$32.00doi:10.1016/j.urology.2005.07.014 55

Page 2: Responsiveness of symptom scales for interstitial cystitis

icsstcutsit

otbTactm

bsddItacp

ptwwiPwfnIucwrgw

rdrrwct

im

ataI(

cAttcfUp20c

ftgBwAcstmotaopmdgssGd

S

IIWPU2

KIID*s†

s

5

t is intended to measure).8 These properties indi-ate the appropriateness of using a scale for cross-ectional evaluations. Responsiveness refers to theensitivity of a scale to identify true changes overime and is a necessary property of endpoints forlinical trials. Although the composite scales aresed as primary and/or secondary outcomes in ICrials, limited information is available on the re-ponsiveness of these scales in the context of clin-cal interventions. One recent report demonstratedhe validity and responsiveness of the ICSI.7The present study evaluated the responsiveness

f the ICSI, ICPI, WICI, and three individual symp-om measures, as part of a clinical trial conductedy the Interstitial Cystitis Clinical Trials Group.5his trial was the first clinical study that includedll the outcomes above and, although no statisti-ally significant differences were found among thereatments, demonstrated overall changes for mosteasures, allowing an evaluation of responsiveness.

MATERIAL AND METHODS

The study population included participants in a double-lind, two by two factorial pilot study of oral pentosan poly-ulfate sodium, oral hydroxyzine, and the combination of bothrugs.5 Eligible participants had a diagnosis of IC, using aefinition similar to the National Institutes of Health-Nationalnstitute for Diabetes, and Digestive and Kidney Diseases cri-eria,9 and moderate or severe symptoms of urinary frequencynd pain/discomfort. All subjects provided written informedonsent, and the institutional review board at each site ap-roved the trial protocol.The primary endpoint for the trial was a participant-re-

orted global response assessment (GRA) at 24 weeks, relativeo overall baseline symptoms.10 A seven-point symmetric scaleas used with the following possible responses: markedlyorse, moderately worse, slightly worse, no change, slightly

mproved, moderately improved, and markedly improved.articipants who reported moderate or marked improvementere defined as treatment responders and those who withdrew

rom the study without providing a final GRA were consideredonresponders. Secondary outcome measures included theCSI and ICPI,2 the WICI,3 symptoms of pain/discomfort andrgency (each on a 0 to 9 Likert scale),4 and voiding frequencyollected by way of a 24-hour voiding diary.4 All outcomesere evaluated at baseline and 3, 7, 10, 17, and 24 weeks after

andomization. For this report, the data from all treatmentroups were combined, and only data from baseline and 24eeks are presented.The Guyatt statistic11 was used as the primary measure of

esponsiveness to change. The Guyatt statistic is calculated byividing the mean change in symptom score in each GRAesponse group by the standard deviation for the group thateported “no change” on the GRA. Linear regression analysisas used to calculate the slopes as a function of the GRA

ategory. Spearman rank correlations were calculated amonghe symptom outcomes.

RESULTS

Among the 121 patients randomized to the orig-nal clinical trial, 94 subjects had complete infor-

ation for the three composite indexes at baseline s

6

nd 24 weeks and are included in this report. Ofhese patients, 87% were women, 85% were white,nd more than 90% had experienced symptoms ofC for at least 1 year. The median age was 46 yearsrange 18 to 79).Summary statistics for the symptom-related out-

omes at baseline and 24 weeks are shown in Table I.verage symptom decreases were approximately

wo points for the ICSI and ICPI and 6.5 points forhe WICI. The mean pain and urgency scores de-reased by approximately one point; the 24-hourrequency decreased by a mean of one void per day.sing a scale of standard deviations to permit com-arisons across measures, with the exception of4-hour frequency, the change ranged from 0.56 to.76 standard deviations. The 24-hour frequencyhange was smaller, around 0.16 standard deviation.The GRA was treated as the reference standard

or the overall response to treatment. Changes inhe six outcomes from baseline to 24 months,rouped by GRA category, are shown in Table II.ecause few patients reported on the GRA that theyere worse, these three categories were combined.ll scales are designed such that a negative value ofhange indicates improvement. A “responsive”cale would demonstrate an increase in the magni-ude of the change as the category of the GRAoves toward improvement; this feature was dem-

nstrated for all indexes. For example, among par-icipants who reported “no change” on the GRA,n average improvement of 0.8 points was foundn the ICSI, and patients in the “markedly” im-roved group demonstrated an average improve-ent of 6.4 points. The three composite scores

emonstrated improvements in all five GRAroups. This was in contrast to the individualymptoms, which worsened, on average, in thoseubjects who reported they were worse on theRA. However, the sample size in the group whoeclined during study was quite small, and the re-

TABLE I. Change in symptom scales frombaseline to 24 weeks (n � 94)

ymptom Index Baseline Week 24 Change

CSI 14.0 � 3.4 11.9 � 4.1 �2.1 � 3.5CPI 12.5 � 2.6 10.3 � 3.9 �2.2 � 3.2

ICI 30.9 � 7.6 24.4 � 11.0 �6.5 � 8.6ain score* 6.1 � 1.3 5.1 � 2.1 �1.0 � 1.8rgency score* 6.5 � 1.5 5.4 � 2.0 �1.1 � 1.64-hr frequency† 18.0 � 6.1 17.0 � 7.2 �0.9 � 5.7

EY: ICSI � O’Leary-Sant Interstitial Cystitis Symptom Index; ICPI � O’Leary-Santnterstitial Cystitis Problem Index; WICI � Wisconsin Interstitial Cystitisnventory.ata presented as mean � SD.Pain/discomfort and urgency scores assessed twice at baseline and average of two

cores used as baseline measurement.Sample size for 24-hr frequency was 92 subjects at baseline and 24 weeks and 90

ubjects for change.

ults should be interpreted accordingly.

UROLOGY 67 (1), 2006

Page 3: Responsiveness of symptom scales for interstitial cystitis

fliricwTqi

acGtesi�a

o

atibsbss

tiwucssas

S

SI

I

W

P

U

2

KD* prove

SI

SIIW

AD*Po

U

Also shown in Table II are the Guyatt statisticsor each response category. A Guyatt statistic of ateast �1.0 (because a negative change indicatesmprovement) was considered to represent highesponsiveness to change. All three compositendexes indicated significant responsiveness tohange in the two highest response categories,hich denoted responders for the clinical trial.his was also true for pain and urgency, but fre-uency was highly responsive only in the markedlymproved group.

To provide insight into the “minimally detect-ble change,” the slope of the change in each out-ome was calculated as a function of the change inRA category. The calculated slopes were �1.2 for

he ICSI and ICPI and �3.1 for the WICI. Thus, forxample, a three-point change in the WICI corre-ponded to a one-category change in the GRA. Sim-lar calculations yielded slopes of �0.8, �0.6, and

1.6, for the individual pain/discomfort, urgency,nd frequency domains, respectively.The correlations among the changes for the six

TABLE II. Change in symp

ymptom Index Worse No Change

ample size 9 26CSI �0.2 � 1.6 �0.8 � 3.2

Guyatt statistic �0.06 �0.25CPI �0.2 � 1.1 �1.3 � 2.5

Guyatt statistic �0.08 �0.52ICI �2.7 � 4.7 �2.1 � 5.5Guyatt statistic �0.49 �0.38

ain score 0.9 � 1.1 �0.3 � 0.9Guyatt statistic 1.00 �0.33rgency score 0.3 � 1.2 �0.4 � 1.2Guyatt statistic 0.25 �0.33

4-hr frequency* 3.1 � 8.5 0.5 � 2.7Guyatt statistic 1.15 0.19

EY: GRA � global response assessment; other abbreviations as in Table I.ata presented as mean � SD.Sample sizes for 24-hr frequency: 25 subjects in each of no change and slightly im

TABLE III. Change in symptom scores b

ymptomndex

Pain

Worse or NoChange Improved*

WorC

ample size 32 62CSI �0.6 � 3.0 �2.9 � 3.4 �0CPI �0.6 � 2.3 �3.1 � 3.3 �1

ICI �1.6 � 4.4 �9.0 � 9.1 �2

bbreviations as in Table I.ata presented as mean � standard deviation for change from baseline to 24 weeks“Improvement” for individual symptom measures defined as any negative change.value from two-sample t test comparing ICPI value for subjects who were worse o

ther P values �0.005.

utcome measures demonstrated good agreement s

ROLOGY 67 (1), 2006

mong the three composite scores with all correla-ion coefficients greater than 0.5. All three compos-te scales also correlated well with the pain score,ut correlations with urgency and frequency wereomewhat lower. The correlation was high (0.64)etween the pain and urgency scores; this was noturprising as they were measured on similar scalesequentially.To evaluate further whether the composite symp-

om indexes were particularly sensitive to changesn certain symptoms, the indexes were examinedithin subcategories defined by pain, frequency, andrgency (Table III). In all cases, a statistically signifi-ant difference (P �0.05) was found in the compositecores between groups defined by a change in aingle domain. However, no evidence showed thatny of the composite scores were particularly sen-itive to change in one specific symptom area.

COMMENT

Although a patient-reported global assessment

scores by GRA categoryGRA Category

SlightlyImproved

ModeratelyImproved

MarkedlyImproved

26 17 14�1.3 � 2.6 �3.4 � 2.6 �6.4 � 3.3

�0.41 �1.06 �2.00�1.1 � 2.4 �3.5 � 2.9 �6.4 � 3.4

�0.44 �1.40 �2.56�5.7 � 6.9 �9.5 � 9.2 �16.1 � 9.4

�1.04 �1.73 �2.93�1.1 � 1.8 �1.5 � 2.1 �3.1 � 1.2

�1.22 �1.67 �3.44�0.8 � 1.1 �1.7 � 1.5 �2.9 � 1.8

�0.67 �1.42 �2.42�0.3 � 4.5 �1.6 � 6.8 �5.9 � 4.1

�0.11 �0.59 �2.19

d groups and 13 subjects in markedly improved group.

ain/urgency/frequency change categoryUrgency 24-hr Frequency

r Noe Improved*

Worse or NoChange Improved*

64 42 483.1 �2.9 � 3.4 �1.0 � 2.8 �3.3 � 3.52.4 �2.8 � 3.4 �0.8 � 2.4 �3.5 � 3.35.4 �8.2 � 9.3 �3.1 � 5.3 �9.7 � 9.9

anged in urgency category compared with those who improved was P � 0.023; all

tom

y p

se ohang

30.5 �.2 �.9 �

.

r unch

uch as the GRA is often used as the primary end-

57

Page 4: Responsiveness of symptom scales for interstitial cystitis

pmtspvsuha

hronbIosttfcidsyormi�rmwtcswt

isIudttlpcdcpP

fa

drbwptot

aaitcrufiIctiWpoecrasdeIip

tpdhimeaRcdslaptpIite

5

oint for clinical trials in IC, composite indexesay provide useful secondary outcome informa-

ion. Demonstrating that these indexes are respon-ive to changes measured by way of the GRA wouldrovide evidence of their utility as endpoints. Con-ersely, a symptom scale that did not demonstrateensitivity to global improvements would be lessseful as an outcome measure, although it mightave a role in the evaluation of treatments aimed atsubset of symptoms.The two O’Leary-Sant IC indexes and the WICI

ave undergone comprehensive evaluation of basiceliability and validity,2,3,7 but, to our knowledge,nly one report has been done on their responsive-ess to change in patients with IC.7 Recently, Lu-eck et al.7 evaluated the responsiveness of theCSI in a randomized clinical trial of different dosesf pentosan polysulfate sodium. Their method wasimilar to that used in the current study; however,heir overall assessment of improvement, the Pa-ient Overall Rating of Symptoms, is somewhat dif-erent from the GRA. In that study, the meanhange in the ICSI score within 32 weeks was 2.3n those who worsened, �1.5 in stable respon-ents, and �5.4 in those who improved. Applyingimilar groupings of response to the present studyielded corresponding change scores for this indexf �0.2, �1.1, and �4.8. The Guyatt statistic amongesponders, defined as those who reported at leastoderate improvement in either study, was nearly

dentical: �1.6 for the study by Lubeck et al.7 and1.5 in the present study. This clearly indicates a

esponsive index among patients with improve-ent. In both studies, the number of patientshose symptoms worsened was small, reducing

he opportunity to assess the responsiveness to in-reasing symptoms. For the other two compositecales evaluated in our study, the responsivenessas also good and qualitatively similar to that for

he ICSI.The use of symptom scales as primary endpoints

n clinical studies requires a definition of clinicalignificance to calculate the required sample sizes.n randomized clinical trials, this may entail eval-ation of the “clinically significant change” (ie, theegree of improvement that has meaning for pa-ients) and the “clinically significant difference” inhe changes between treatment groups. The calcu-ated slopes in the present study suggest that a one-oint change in the ICSI or ICPI, or a three-pointhange in the WICI, would represent a clinicallyetectable change, as defined by a one-categoryhange on the GRA. Lubeck et al.7 reported a two-oint change in the ICSI for each category of theatient Overall Rating of Symptoms scale.It should be noted that these data were derived

rom a trial limited to subjects with at least moder-

te symptoms and the responsiveness of these in- u

8

exes among subjects with milder symptoms willequire evaluation in other contexts. Furthermore,ecause complete data on changes from baselineere required for the current analysis, only theatients who did not withdraw from the clinicalrial were included. This subgroup may have anverrepresentation of subjects who improved andhe results should be interpreted accordingly.To our knowledge, no formal methods are avail-

ble to determine which of these scales is mostppropriate for the assessment of symptom changesn patients with IC. Although our results indicatehat all three composite indexes are responsive tohange as measured by the GRA, overall responseates could differ depending on which index wassed. For example, suppose the response were de-ned as at least a one-point improvement in theCSI, corresponding to the clinically detectablehange. In the present study, 60 (64%) of 94 pa-ients would be considered responders by this def-nition. Alternatively, if a three-point change in the

ICI were used to define response, 59 (63%) of 94atients would be considered responders. Theseverall response rates are nearly identical; how-ver, they represent different patient subsets. Spe-ifically, 27 patients (29%) would be consideredesponders by one definition but not by the other,nd only 46 patients (49%) would meet the re-ponse criteria for both. This lack of correspon-ence reflects the heterogeneous symptoms andffects on quality of life in patients with IC. TheCSI is also weighted more toward irritative void-ng symptoms, rather than pain, and this may ex-lain in part these differences.In addition to the choice of scale, a second po-

ential difficulty exists in using symptom scales asrimary endpoints.9 Missing data and subject with-rawal are inherent to all clinical trials and canave important consequences for the validity and

nterpretation of results. Methods for addressingissing data often require assumptions that are

ither invalid or unevaluable in a given study. Onedvantage of outcomes such as the Patient Overallating of Symptoms and GRA, for which responsean be easily defined, is that subjects who with-raw from the study can be included as nonre-ponders for the calculation of response rates, al-owing an intention-to-treat analysis. Therefore,lthough composite symptom indexes play an im-ortant role as secondary endpoints, it is importanthat changes, if any, be interpreted in the context ofotential biases that might arise from missing data.t is also helpful to have the ability to assess thendividual domains of IC symptoms, especially inhe evaluation of therapies that target specific ar-as. This can be done by including individual pain,

rgency, and/or frequency measures as secondary

UROLOGY 67 (1), 2006

Page 5: Responsiveness of symptom scales for interstitial cystitis

eo

cwoIatvss

smf9

C5

yi1

sf1

tt

ct1

vIo

qs

I

WA

d7

oJ

A(AMRGMCwMUgTMLBtLmRNJviDiKRFgDBMGMDLlf(

U

ndpoints or by evaluating of individual questionsf the composite scales.

CONCLUSIONS

The ICSI, ICPI, and WICI are responsive tohange over time as measured by GRA in patientsith IC. These indexes are recommended as sec-ndary endpoints for future clinical trials of IC.ndividual symptoms should also be measured toid in the evaluation of effect mechanisms for newherapies. Newer symptom scales will need to bealidated in future clinical studies for their respon-iveness to change and correlation with global re-ponse assessments.

REFERENCES1. Keay SK, Zhang C-O, Shoenfelt J, et al: Sensitivity and

pecificity of antiproliferative factor, heparin-binding epider-al growth factor–like growth factor, and epidermal growth

actor as urine markers for interstitial cystitis. Urology 57:–14, 2001.

2. O’Leary M, Sant GR, Fowler FJ, et al: The Interstitialystitis Symptom Index and Problem Index. Urology 49(supplA):58–63, 1997.

3. Goin JE, Olaleye D, Peters KM, et al: Psychometric anal-sis of the University of Wisconsin Interstitial Cystitis Scale:mplications for use in randomized clinical trials. J Urol 159:085–1090, 1998.

4. Propert KJ, Schaeffer AJ, Brensinger CM, et al: A pro-pective study of interstitial cystitis: results of longitudinalollow-up of the Interstitial Cystitis Data Base cohort. J Urol63: 1434–1439, 2000.

5. Sant GR, Propert KJ, Hanno PM, et al: A pilot clinicalrial of oral pentosan polysulfate and oral hydroxyzine in pa-ients with interstitial cystitis. J Urol 170: 810–815, 2003.

6. Mayer R, Propert KJ, Peters KM, et al: A randomizedontrolled trial of intravesical bacillus Calmette-Guérin forreatment refractory interstitial cystitis. J Urol 173: 1186–191, 2005.

7. Lubeck DP, Whitmore K, Sant GR, et al: Psychometricalidation of the O’Leary-Sant Interstitial Cystitis Symptomndex in a clinical trial of pentosan polysulfate sodium. Urol-gy 57: 62–66, 2001.

8. Lohr KN, Aaronson NK, Alonso J, et al: Evaluatinguality-of-life and health status instruments: development ofcientific review criteria. Clin Ther 18: 979–992, 1996.

9. Gillenwater JY, and Wein AJ: Summary of the National

nstitute of Arthritis, Diabetes, Digestive and Kidney Diseases 2

ROLOGY 67 (1), 2006

orkshop on Interstitial Cystitis, NIH, Bethesda, Marylandugust 28–29, 1987. J Urol 140: 203–206, 1988.10. Propert KJ, Payne C, Kusek JW, et al: Pitfalls in the

esign of clinical trials for interstitial cystitis. Urology 60:42–748, 2002.11. Guyatt G, Walter S, and Norman G: Measuring change

ver time: assessing the usefulness of evaluative instruments.Chron Dis 40: 171–178, 1987.

APPENDIX

complete list of the Interstitial Cystitis Clinical Trials GroupICCTG) membership follows: University of Pennsylvania:lan J. Wein, M.D. (primary investigator), George W. Drach,.D., Philip Hanno, M.D., Eric Rovner, M.D., Marilou Foy,.N., Gloria McNamara, R.N.; New England Medical Center:rannum R. Sant, M.D. (primary investigator), Erol Onel,.D., T. C. Theoharides, Ph.D., M.D., Patricia Radgowski,arolyn Shea-O’Malley, R.N.; University of Rochester: Ed-ard M. Messing, M.D. (primary investigator), Robert Mayer,.D., Kay Rust, R.N., M.S.N., F.N.P., Elizabeth Smith, B.S.;niversity of Maryland: John Warren, M.D. (primary investi-ator), Toby Chai, M.D., Susan Keay, M.D., Linda Horne,heresa Jackson; University of Oklahoma: Daniel J. Culkin,.D. (primary investigator), James F. Donovan, JR, M.D.,

ynda Kelsey, R.N., M.S., Karen Mataranglo, R.N.; Williameaumont Hospital: Ananias C. Diokno, M.D. (primary inves-igator), Kenneth Peters, M.D., Eleanor Anton, R.N., Loniampkins; Henry Ford Hospital: David Burks, M.D. (co-pri-ary investigator), Rifaat Dagher, M.D., Michelle Peabody,.N., Jill Sullivan, R.N., B.S.N.; Queen’s University: J. Curtisickel, M.D. (primary investigator), Alvaro Morales, M.D.,

oe Downey, B.Sc., M.Sc., Laurel Emerson, R.N.; Stanford Uni-ersity Medical Center: Christopher K. Payne, M.D. (primarynvestigator), Kathryn Azevedo, Ph.D., Gilbert Rigaud, M.D.,ebra Clay, R.N.; University of Pennsylvania Data Coordinat-

ng Center: J. Richard Landis, Ph.D. (primary investigator),athleen J. Propert, Sc.D. (co-primary investigator), Deborah. Erickson, M.D. (Pennsylvania State University), John T.arrar, M.D., John E. Tomaszewski, M.D., Rosemary A. Madi-an, R.N., M.P.H., Ted Barrell, B.A., Denise Cifelli, B.S., Jamesattilo, B.S., Stephen B. Durborow, B.S., Lori Fanelli, R.T.,.A., Christine Hardy, M.S., Raymond W. Hilderbrand, M.S.,.B.A., Chris Helker, R.N., M.S.P.H., Gina Norwood, B.S.,argi Parikh, M.S., Yanlin Wang, M.S., Yawei Zhang, M.D.,.S.; National Institute of Diabetes and Digestive and Kidneyiseases: John W. Kusek, Ph.D., Christopher Mullins, Ph.D.,eroy M. Nyberg, JR, Ph.D., M.D.; University of Virginia: Wil-

iam A. Steers, M.D. (Chairman of ICCTG Steering Committeerom 1998 to 2001); Yale University: Harris E. Foster, M.D.Chairman of ICCTG Steering Committee from 2001 to

004).

59