Final Program

Eleventh Annual Meeting of the Society for Clinical Trials

May 6-9, 1990 Harbour Castle Westin, Toronto, Ontario, Canada

SUNDAY, MAY 6, 19g0

8:00 AM - 2:00 PM

8:30 AM - 5:30 PM

12:00 PM - 5:00 PM

4:00 PM - 8:00 PM

7:00 PM - 10:00 PM

• P R O G R A M

Board of D i m

R ~ - , ~ x ~

E-ve~ng Sodal ~

F~m 718

Ha~oour R)Wr

Had)our Ballroom

MONDAY, MAY 7, 1990

8:00 AM - 5:00 PM

8:30 AM - 5:00 PM

8:30 AM - 8:45 AM

8:45 AM - 9:30 AM

9:3p, AM - 10:00 AM

9:30 AM - 4:00 PM

P-01

F l ~ o n Hmbour ~

E~hibit Program ~ ~

~ ~ ~ ~ ~ ~

David ~ M o t s , President ~ c i e ~ for C l i n i~ l Trials

~ ~ ~ ~ 1 ~ S t u ~ Pooo~k London ~ h o o l of Hygiene ~nd T rop i~ l Medicine

~ ~ ~

~ ~ 1 ~ ~ ~ ~ - ~ - 1 1 : ~ ~

I m p ~ of • Changing ~ m p u t e r Environmen~ on ~he O~r~ t ion~ of ~ C l in i~ l Tri~l; Edward Km~k~, dohn~ Hopkin~ M e d i ~ l Institutions

Controlled Cbrucal Trials 11 235-252 (1990) © Elsevier Science Pubhshing C o , lnc 1990 655 Avenue of the Americas, New York, New York 10010

235 0197-2456/1990/$350

236 Final Program

MONDAY, MAY 7, 1990 (Continued)

9.30 AM - 4:00 PM Poster Session I (Continued) (Poster Primetime - 9 : . 4 5 - 1 1 : 0 0 A M )

P-02 Use of a Computerized Data Audit System to Increase the Quality of Data in a Long-Term Multicenter Trial (LTMT); Mary Brown, University of Rochester

P-03 Capturing Cancer Clinical Trials Data at the Source and Communicating them Electronically; Susan Ree Welch, The University of Texas M.D. Anderson Cancer Center

P°04 Development of a Lookup Table within a SIR Database: Adapted from ICD*9*CM, for Classification of Diagnoses; Jennifer Whyte, McMaster University

P-05 Changing Coordinating Centersin Mid-Triah Lessons Learned from the IHDP; Patricia Wilkinson, Johns Hopkins University

P-06 Need for Selectivity of Data Collected in Clinical Trials; Denise Julian, McMaster University

P-07 Industry Involvement in Cooperative Group Trials- Effects on Data Management; Nancy Paul, National Cancer Institute of Canada

P-08 An Analysis of Data Collection Mechanisms; Doris Tyrrell, VA Medical Center, Palo Alto

Use of an External Lookup Table for Coding Textual Data; Harvey Nelson, McMaster University

P-IO Computer Assisted Electroretinogram Grading; Anik Ganguly, University of Wisconsin

P-11 Problems Developing a Disease-Specific Quality of Life Measure for Multiple Sclerosis (MS); Iris Brownscombe, University of Western Ontario

Po12 Effects of the Omission of Known and Unknown Prognostic Covariates on Treatment Comparison in Survival Studies; Sylvie Chevret, Hopital Saint-Louis

P-13 The Requirements of a Multi.Centered Clinical Trial Designed to Evaluate Spinal Implants: Industry and FDA Perspectives; John Carlow, Buckman Company, Inc.

P-14 Outcome Adjudication in an Ongoing Trial; Sheila Hewson, McMaster University

P-15 The Operation of a Central Adjudication Committee and its Effect on the Validity of the Assessment of Treatment Benefit; Jane Sicurella, McMaster University

P-16 Interrater Variability in Patholog,cal Classification: Impact on Prognosis of Non-Hodgkin-Lymphoma; Thomas Zwingers, Biometric Center for Therapeutic Studies

Final Program 237

MONDAY, MAY 7, 1990 (Continued)

9:30 AM - 4:00 PM Po~er Se~ion I (Continued) (Po~e¢ Pdmelime - 9:.45 - 11:00 AM)

P-17 Investigator Effects in a Multicenter Trial for Parkinson's Disease; David Oakes, University of Rochester

po18 Need an Endpoint for a Clinical Trial? Just Ask; Lynn Harrelson-Woodlief, Duke University

P-19 Issues in the Design and Evaluation of a Controlled Intervention Trial to Improve the Utilization of Mammography; Phyllis Gimotty, Michigan Cancer Foundation

The Healthy Responder Effect in Nonrandomized Clinical Trials: An Example from the Cardiac Arrhythmia Suppression Trial; AJfred Hallstrom, University of Washington

P-21 The Analysis of a Complex Dental Study with Repeated Measures; Robert Anderson, VA Medical Center, Hines

P-22 A Graphic Method for Evaluating Effects of Treatment on Cardiac Arrhythmias; Federico Dies, Lilly Research Labora- tories

P-23 Screening and Recruitment for the Systolic Hypertension in the Elderly Program (SHEP); Jeffrey Probstfield, National Heart, Lung and Blood Institute

P-24 Staffing Patterns and Recruitment Strategies in Systolic Hypertension in the Elderly Program (SHEP); Nora Cosgrove, Robert Wood Johnson Medical School

P-25 Recruitment Strategies: Studies of Left Ventricular Dysfunction (SOLVD); Barbara Dierenfeldt, University of Illinois

P-26 Trends in Patient Accrual Rates by Clinical Center Over a 10 Year Period; Dawn Blackhurst, Johns Hopkins Medical Institutions

P-27 Systems of Logging Exclusions in Multicentre Clinical Trials; Annette Seip, McMaster University

P-28 Determining Ineligibility: The Role of Central Review of Supporting Documentation; Anna Sadura, National Cancer Institute of Canada

Problems in Initiating Recruitment in a Clinical Trial; Susan Margitic, Bowman Gray School of Medicine

Is Standardization Possible in Multicenter Clinical Trials in Which 45 Medical Centers and 500 Clinical Professionals are Involved: The COMS Experience; Jan Markowitz, Johns Hopkins Medical Institutions

P-31 Structure, Function and Costs of Site Visits; Nancy Fink, Johns Hopkins Medical Institutions

238 Fina~ Program

MONDAY, MAY 7, 19g0 (continued)

9:30 AM - 4:00 PM PoSi t ~ I (Continued) ( P o = = P r i m = ~ - e:.4s- 1 1 : 0 o / ~ )

P-32 Issues in the Quality of Clinical Trial Protocols; David Moher, Ottawa Civic Hospital

Issues in the Design of Reproducibility Studies for Quality Assurance in Clinical Trials; Joan Jefferys, Johns Hopkins University

Limitations in Generalizing Results from Clinical Trials; Linda Sabella, VA Medical Center, Hines

P-35 The Impact of Management Guidelines on Quality of Breast, Colo-Rectal and Ovarian Cancer in Community Hospitals; Roberto Grilli, Instituto Mario Negri

P-36 VA Cooperative Studies Program: A Registry of Studies Conducted since the Program's Inception; David Weiss, VA Medical Canter, Perry Point

10:00 AM - 11:00 AM Contributed Paper Session I (four [4] simultaneous sessions)

Session IA: Phase I T~l~is Piers 2/3 Chairperson: Richard Simon

10"00 AM 01 Drug Development for AIDS: How Much Can We Get Out of Phase I?; Anne Cross, Bristol-Myers Squibb

10 20 AM 02 Bayesian Analysis of Phase I Clinical Trials; Joel Greenhouse, Carnegie Mellon University

10:40 AM 03 Continual Reassessment Models in Phase I Clinical Trials in Cancer; John O'Quigley, Hopital Saint-Louis

Semlion IB: ~ o f Clinical Tda~

Had:)our Ballroom A

Chairperson: William Biackwelder

10:00 AM The Impact of Stopping Rules on Meta.Analyses; Laurence Freedman, National Cancer Institute

10:20 AM 05 The Design of a Meta-Analysis Collecting Individual Patient Data in Advanced Ovarian Cancer; Lesley Stewart, British Medical Research Council

10'40 AM Risk Factors for Publication Bias: Results of a Followup Study; Kay Dickersin, Johns Hopkins School of Public Health

Se=k~ IC: Tdal MBlageme~t

Cltairperson: Joseph Collins

Harbour B~dlroom C

Final Program 239

MONDAY, MAY 7, 1990 (Continued)

10:00 AM - 11:00 AM ConldbutBd Paper ~ I (Continued) (four [4] simultaneous sessions)

10.00 AM 07 Low-Tech Profile with Individualized Options for a Multi- national Screening Trial; Sophie Alexander; Universite libre de Bruxelles

10:20 AM 08 Pre-Randomization Drug Withdrawal, Yield and Process: The SHEP Trial; Jacqueline Smith, Medical Research Institute of San Francisco

10:40 AM 09 Recruitment of Infants for Randomized Clinical Trial; Jean- Paul Collet, Clinical Pharmacology Unit, Lyon

10:00 AM S-lO

10:20 AM S-11

10:40 AM S-12

11:10 AM - 12:40 PM

See=d~n ID: Sludent Soholar~ip Piers 4/5

Cltairpemon: Frank Rockhold

Examination of Designs for Phase I Clinical Trials; Sherryl Baker, University of Southern California

Sampling Design of Multiwave Studies with an Application to the Massachusetts Health Care Panel Study; Rick Chappell, University of Chicago

AIDS Trials, Civil Rights, and the Social Control of Therapy: Should We Embrace New Drugs with Open Arms?; David Salisbury, University of British Columbia

Plenary 8e~ion I Fmnl~mao Ballroom

Gliniol~ Tria~ of 8oreening for Eady Del~i ion of Dieea~

There is considerable public health interest in screening the population for early detection of a variety of diseases. Such screening has several possible advantagesand disadvantages. Among the possible benefits are improved prognosis for some individuals whose disease is detected early by screening, and resource savings, while possible disadvantages include longer morbidity for cases detected by screening whose prognosis is unaltered, anxiety and morbidity for the individuals with false positive test results, and resource costs. Because a favorable benefit to cost ratio in a population is not assured, many feel that careful evaluation of screening intervention is required. Randomized trials offer the surest way of determining their value, since several subtle forms of bias can lead to misleading results from observational studies. This session will explore the many methodological issues involved in conducting randomized trials in the context of screening for three diseases: breast cancer, lung cancer and cystic fibrosis. Investigators will discuss study organization, subject recruitment, quality control of the screening test, compliance with screening, diagnostic and treatment procedures, population follow-up and ascertainment of endpoints, and analysis of results.

240 Final Program

MONDAY, MAY 7, 1990 (continued)

11:10 AM - 12:40 PM Rlnmy Se~ion I (Conlanued)

Chairperson: Philip Prorok, National Cancer Institute Anthony Miller, University of Toronto Robert Fontana, Mayo Clinic Christopher Green, University of Wisconsin L.J. Wei, University of Wisconsin

12:40 PM - 2:00 PM Lunch Metro Ce~b-e

2:00 PM - 3"30 PM Three Simultaneous Wodcshops

Wodcshop I Ctmosing Appmpda~ Dose I.mmis in Clinical Trials

Piers 2/3

Selection of a proper dose level for drugs and other modalities (e.g., radiotherapy or hyperthermia in cancer treatment) is an important goal of early phase clinical trials, both single-agent, single-modality (phase I) trials and combination-agent, combination-modality trials. The statistical aspects of the design and analysis of phase I clinical trials have received considerably less attention in the literature than later phase trials. Also, although pre-clinical animal studies often use response surface methodology in the design and analysis of similar studies, this methodology is seldom used explicitly in human studies. Speakers in this session will review the current state-of-the-art in the design and analysis of phase I clinical trials and in the application of response surface methodology to clinical trials and discuss those areas needing further work.

Chairpemon: Stephen George, Duke University Barry Storer, University of Wisconsin Gary Rosner, Duke University Hans Carter, Medical College of Virginia

Wodcshop I Qual~y Comml bsues k~ Multicentm Clinical Trials

~ Balkoom A

In order to achieve and maintain an acceptable level of data quality, it has been noted that certain principles are important. Clinical trials that use laboratories, coding or reading units for processing data collected in the Clinical Center have additional complications in achieving high data quality; quality control for these types of trials is essential. By quality control we mean a system of controlling bias and decreasing variability. The issue for consideration is how much is necessary. The systems that will detect bias, including secular trends and other deviations, are expensive and may detract from the main objectives. This workshop will explore the cost and efficacy of quality control systems for a fundus photograph reading center, dietary coding unit and estrogen-receptor laboratories.

ChaJq=erson: Patricia Cleary, The George Washington University

Final Program 241

MONDAY, MAY 7, 19g0 (Continued)

2:00 PM - 3:30 PM Three Simultammu~ Wodmlm~ (Continued)

Larry Hubbard, University of Wisconsin Jennifer Gassman, Cleveland Clinic Foundation Rebecca Gelman, Harvard University Jim Neaton, University of Minnesota

Wodcshop III When a Coordinating Canter Assumes an Ongoing Trial

Harbour Ballroom C

A mid-course change in the Coordinating Canter, whether the director alone or the entire Coordinating Center, can result in trauma or chaos for any study. Regardless of the cause of the change, the new Coordinating Center director will face a number of issues that are remarkably similar. Unless there has been promotion from within, the Coordinating Center director debuts as the only newcomer in an existing collaborative structure. When the entire Coordinating Canter is replaced, this is true for the entire staff. The new staff must learn the details of the study and must take over responsibility for all ongoing Coordinating Canter activities without any lapses. The new director must rapidly and accurately determine what problems, if any, exist with regard to adherence to study design; patient recruitment, retention and compliance; data collection and processing; performance monitoring; sequential analysis; and personnel. The patterns and the pitfalls inherent in this potentially difficult situation will be discussed from the point of view of individuals who have dealt with it. The focus will be on the activities and problems faced, rather than on the conditions that precipitated a succession of leadership. Some suggestions will be offered for those considering taking on such responsibility.

Chairperson: Mary Foulkes, National Institute of Neurological Disorders and Stroke Barbara Tilley, Henry Ford Hospital Herman Mitchell, New England Research Institute John Crowley, Fred Hutchinson Cancer Research Canter

3:30 PM - 4'00 PM

4:00 PM - 5:30 PM I : ~ Smmion II

The "Parallel Track" for Uf~-Threatmdng ~ : From '/heon] to Practk~

Recently, considerable discussion has ensued concerning this approach, in which patients who are ineligible or unable to participate in a controlled clinical trial may receive an investigational therapy for which a favorable risk/benefit assessment has not been demonstrated. Although originally developed within the context of trials for AIDS therapies, it may be applied in other disease categories in the near future. Speakers and discussants will focus on two key topics: First, what are the medical, ethical, scientific, logistical and legal difficulties in conducting these trials? Second, what

242 Final Program

MONDAY, MAY 7, 1990 (Continued)

2:00 PM - 3:30 PM Three Simultaneou~ Wodmhop~ (continued)

information can be derived, and what are the potential strengths and limitations of the data obtained? Recant experience from a clinical trial using this approach will be presented.

Chairperson: Richard Schwarz, Jr., Sterling Research Group Paul Worrall, Bristol-Myers Squibb Frank Rockhold, SmithKline Beecham Baruch Brody, Baylor College of Medicine

Discussants: Martin DeLaney, Project Inform Robert Temple, Food and Drug Administration

6'00 PM Frontenac Ballroom

TUESDAY, MAY 8, 19g0

8.00 AM - 5:00 PM Rogislmtio. I-I=tx~r Foyer

8:00 AM - 5"00 PM Exhibit Program Mebo East

8:30 AM - 10:00 AM I ~ Session III Frontenac Ballroom

Measuring Qtadity of Ufe in Clinical Trials

Although measures of quality of hfe (QOL) have received considerable attention in recent years, their use as endpoints in clinical trials is still relatively rare. This situation appears to be changing rapidly as more experience is gained with the application of various QOL instruments in a clinical trial setting However, the problems involved in using QOL endpoints are formidable. Some of these problems arise directly from issues inherent in defining and measuring subjective variables expressing attitudes, moods, and emotions; statistical problems also arise related to repeated measures of multiple items over varying time points and with missing data. Speakers in this session will address these design and analysis issues with emphasis on examples of actual clinical trials.

Chairperson: Stephen George, Duke University Sally Shumaker, National Heart, Lung and Blood Institute P.K. Tandon, Sterling Research Group George Williams, Cleveland Clinic

Discussant: Robert Temple, Food and Drug Administration

10'00 AM - 10:30 AM

10.00 AM - 4.00 PM P o s t ~ S e e ~ o e II poeter Primelkne - 10:.15 - 11:30 AM)

P-37 Proteo: An Expert System for Quality Control of Patient's Eligibility Criteria in Clinical Trials; Marisa De Rosa, CINECA- Italian Northeast Inter-University Computer Center-Bologna

Final Program 243

TUESDAY, MAY 8, lg90 (continued)

10:00 AM - 4:00 PM Po~ter ~__o,e~=km_" II (Continued) Memo East ( P o ~ Pdme~ime- 10:15- 11:30/~I)

P-38 Data Management for Phase II Studies: The Experience of the EORTC Data Center; Martine Van Glabbeke, EORTC Data Center

P-39 Design of a Distributed Data Processing System for a Large- Scale, Short-Term Trial in Infectious Disease; Irene Voynick, VA Medical Center, West Haven

P-40 Utilizing the Relational Data Base Data Dictionary for Applications; Fran LoPresti, Maryland Medical Research Institute

P-41

P-42

P-43

The UNIX Operating System in the Clinical Trials Environ- ment; Elizabeth Yetisir, McMaster University

Experience Using a Drug Database to Classify Concomitant Medications Taken During Clinical Trials; Gary Anderson, Innovus Inc. and McMaster University

An Automated Participant-Tracking System for Clinical Trials; Pierre-Andre La Chance, Kaiser Permanente Center for Health Research

P-44

P-45

P-46

A Microcomputer Based Patient Tracking System; Paul Utwin, University of Washington

The DATAFAX Project; D. Wayne Taylor, McMaster University

Estimating Rates of Change in Randomized Studies; Fong Wang, Syntex Laboratories

P-47 Difficulty of Retrospectively Correlating Dose with Time to Response Based on a Phase II Clinical Trial; Cheryl Hagerty, City of Hope Medical Center

P-48 Biometric Issues in Measuring and Analyzing Compliance in Clinical Trials; Joerg Hasford, Biometric Centre for Therapeutic Studies

P-49 Does Centre Size Affect Protocol Compliance on Multi- Centered Clinical Trials? (MCCT) The National Cancer Institute of Canada Clinical Trials Group's (NCIC CTG) Experience; Stephanie Loomer, National Cancer Institute of Canada

Outcome Measurements in Clinical Trials for Stroke Patients; Eileen Brown, Henry Ford Hospital

P-51 Trial Implementation Rating Scale; Kathryn Taylor, University of Toronto

P-52 An Evaluation of Human Panelists in Assessing Coronary Atherosclerosis; Janice Pogoda, University of Southern California

244 Final Program

TUESDAY, MAY 8, 1990 (Continued)

10:00 AM - 4:00 PM Poplin Session II (continued) (Poel~ Pdmetime- 10:.15- 11:30 AM)

Methodological Guidelines for Randomized Clinical Trials of Surgical Procedures; John Kestle, McMaster University

Training Levels Comparison (TLC) Trial: Design Issues; Jeannette Lee, University of Alabama at Birmingham

Dosage Adjustments in Response to Monitored Plasma Concentrations: Can Unblinded Staff Adhere to Objective Criteria? J. Todd Sahlroot, National institutes of Health

A Model for Comparing the Efficacy of Randomization Procedures in Controlled Clinical Trials; Franck Chauvin, Centre Leon Berard

P-57 Sample Size for Trials where Survival Depends on First Response Rate; Benny Zee, National Cancer Institute of Canada

Power and Sample Size Calculations: A Raview and Computer Program; William Dupont, Vanderbilt University School of Medicine

P-59 Conditional Power Calculations through Computer Simulation as an Aid to Data Monitoring; Kelvin Lee, VA Medical Center, Palo Alto

P-60 Automatic Drug Replenishment System for Multi-Center Clinical Trials; Loretta Malone, VA Medical Center, Albuquerque

P-61 Drug Accountability System for Clinical Trials; Mark Jones, VA Medical Center, Albuquerque

P-62 Drug Distribution in a Complex Multi-Arm, Multi-Drug Trial: The Postmenopausal Estrogen/Progestin Interventions (PEPI) Experience; Carol Wasilauskas, Bowman Gray School of Medicine

P-63 Coordination and Patient Management of a Multicenter Clinical Trial; Karen Clark, University of California, San Diego

P-64 DATATOP: Managing an Accelerated Trial; Rita Pelusio, University of Rochester

P-65 Impact of the Monitoring of Centre Performance; Beverly Koski, National Cancer Institute of Canada

P-66 Testing Protocol Knowledge of Clinic Coordinators; Donna Boone, University of Southern California

P-67 Intermediate Drug Approval - A Proposal for Expediting the Conduct of Randomized Clinical Trials while Containing Cost; Alfred Hallstrom, University of Washington

P-68 Changes in NHLBI Guidelines for Clinical Trials, Community-

Final Program 245

TUESDAY, MAY 8, IggO (continued)

10:00 AM - 4:00 PM Poplin ~==ion II (Continued) (Po~B¢ Pdm~ime- 10:.15- 11:30/1~1}

Based Studies, and Demonstration and Education Projects; David Monsees, Jr.. National Heart, Lung,and Blood Institute

P-69 Effects on Patients and Staff of Altering the Protocol of a Clinical Trial after Adverse Effects - Experience in the Cardiac Arrhythmia Suppression Trial (CAST); James Myles, University of Washington

P-70 Coordinating a Pivotal Clinical Trial (PCT): An Educational Experience; Dianne Johnston, National Cancer Institute of Canada

P-71 The Reluctance of Physicians to Recommend Clinical Trials: Factors which may Affect the Attractiveness of Trials for Phys- icians; Laura Siminoff, University of Pittsburgh

P-72 A Crossover Analysis of the Educational Benefits of a Patient Information Handbook; Dorothy Cline, VA Medical Canter, Palo Alto

P-73 A Directory of Clinical Trials Registries; Philippa Easterbrook, Johns Hopkins University

10:30 AM - 12:30 PM

10:30 AM 13

10:50 AM 14

11:10 AM 15

11:30 AM 16

11:50 AM 17

Cor~bu=d Pap~ Se=aon N (four [4] simultaneous sessions)

Session IIA: _k~=_ jes of Bias ancl Clinical Trials

I - lad~r Ballroom A

Chairperson: Fred Ederer

Selection Bias in a Randomized Controlled Clinical Trial; Dalene Stangl, Carnegie Mellon University

Can Observational Studies Ever Replace the Randomized Clinical Trial (RCT)? The Case of Coronary Bypass Surgery with the Internal Mammary Artery (IMA) Graft; William Henderson, VA Medical Canter, Hines

Efficacy Versus Intent-to-Treat Analysis in the Canadian American Ticlopidine Study; Robin Roberts, McMaster University

Analysis by Intent to Treat: Is there Really an Option?; Young Lee, National Institutes of Health

To Blind...Or Not to Blind?; Cindy Coiling, VA Medical Canter, Albuquerque

12:10 PM 18 Must the Evaluator be Blinded?; John Noseworthy, University of Western Ontario

246 Final Program

TUESDAY, MAY 8, 1990 (Continued)

10:30 AM - 12:30 PM

10:30 AM 19

10:50 AM 20

Contdbuto¢l Papo¢ So~lioct H (Continued) (four [4] simultaneous sessions)

Harbour Baikoom B

C I ~ : Allan Donner

Rasampling Methods for Repeated Measures in Clinical Trials; Gregory Evans, Bowman Gray School of Medicine

Analysis of Longitudinal interval Count Data; Therese Stukel, Dartmouth Medical School

11:10 AM 21

11:30 AM 22

11 '50 AM 23

12:10 PM 24

Application of Longitudinal Data Analysis in a Clinical Trial: The Macular Photocoagulation Study; Maureen Maguire, Johns Hopkins Medical Institutions

Analyzing the Number of "Rejection Episodes" in Human Transplantation; James Rochon, University of Western Ontario

Interpreting the Effects of Mediating Variables on Blood Pressure in the HPT; Mark Van Natta, Johns Hopkins University

Statistical Design of the Postmenopausal Estrogen/Progestin Intervention Trial; Mark Espeland, Bowman Gray School of Medicine

10.30 AM 25

10,50 AM 26

11'10 AM 27

Session IIC: Data Monitoring I I-kabour Ballroom C

Chaicperson: Judith Goldberg

Early Termination of the VA Cooperative Study of Sterogd Therapy for Systemic Sepsis; Peter Peduzzi, VA Medical Center, West Haven

Statistical Issues in Monitoring a 2X2 Factorial Clinical Trial; W. Jackson Hall, University of Rochester

Interim Analyses for Sample Size Reestimation Without Unblinding Treatment Group Identification in Double-Blind Clinical Trials; Weichung Shih, Merck Sharp & Dohme Research Laboratories

11 '30 AM 28

11:50 AM 29

Data Monitoring Through Stochastic Curtailing When the Outcome Proportions are Small - An Exact Approach; Ronald Thomas, VA Medical Center, Pale Alto

Study Duration for Stratified Clinical Trials with Survival Data and Early Stopping Rule; Kyungmann Kim, Dana-Father Cancer Institute

10:30 AM 30

S o ~ lil~. S t ~ Topics I P ~ 4/5

C~rpersoo: Barbara Hawkins

The Statistical Analysis of Causes of Death in Randomized Clinical Trials; Marc Buyse, EORTC Data Center

Final Program 247

10:30 AM - 12:30 PM

TUESDAY, MAY 8, lgg0 (Continued)

Contdl=~m~ Papm" ~ II (Continued) (four [4] simultaneous sessior~s)

10:50 AM 31 Grouped Residuals for Proportional Hazards Models; Myrto Lefkopoulou, Dana-Farber Cancer Institute

11:10 AM 32 The Analysis of Failure Time Data in Crossover Studies; John Lewis, ICI Pharmaceuticals

11 "30 AM 33 Comparison of Methods for Analyzing Paired Survival Datawith Censoring; Pamela Hartigan, VA Medical Center, West Haven

11:50 AM 34

12:10 PM 35

12:30 PM - 2:00 PM

2:00 PM - 3:30 PM

Two Methods for Comparing Staging Systems for Chronic Diseases; Jacques Benichou, Hopital Saint-Louis

Plans for Analysis of the SHEP Trial; Janet Wittes, VA Medical Center, New Haven

~ Monitoring in Clinical TriaLs: "Wl~m ~l'muicl W~ ~ o p The Trial~

Habour Ballroom A

When sufficient-or nearly sufficient-data have accumulated, data monitoring committees face a critical decision on stopping. The decision to stop usually involves data not only on the primary response variable, but also on secondary response variables and on adverse reactions. The time to stop the trial is when the evidence has become convincing to our peers that a clear-cut conclusion about efficacy and safety can be drawn. If the trial is stopped prematurely, the evidence will fail to convince our peers and the trial will, therefore, fail to achieve its objective. If, on the other hand, the trial continues until the evidence is more than sufficient, then patients in- or outside the study will continue to receive the inferior treat- ment. Achieving data monitoring committee consensus on when to stop is often not an easy task. The clinicians, statisticians, ethicists or representatives of other disciplines on the committee may have prolonged disagreements before reaching a consensus. In this workshop data from three recent trials will be presented as they were presented to the data monitoring committees for these trials at a time when the committees interpreted the evidence to be either sufficient or not quite sufficient for a decision to stop. The audience will be invited to discuss the data and make recommendations on stopping as if they were members of the data monitoring committee. After 15 minutes of discussion for each trial, the decision, and reasons therefor, reached by the data monitoring committee of the trial will be presented.

Chaiq=eflmn: Fred Ederer, The EMMES Corporation Lawrence Friedman, National Heart, Lung and Blood Institute Richard Parrish, University of Miami David DeMets, University of Wisconsin

2 4 ~ ~:~e,~, P;,~;am

TUESDAY, MAY 8, 1990 (Continued)

2:00 PM - 3.30 PM Wododmp V Hadmur Ballroom B Data Entry in Multk:entm Clinical Studies: "How We Have ~ To Solve The Problem"

Clinical trials and epidemiologic studies all utilize some form of intelligent data capture such as PC-based data entry, with or without paper data forms. Some studies capture data in a distributed fashion (entered into machine readable form at the collaborating Clinical Centers), others in a centralized fashion (data forms sent to the Coordinating Center for entry). Although each study has different requirements for the amount, complexity and timeliness of data collected, some standard requirements for data capture are certainly common to all. Each of the collaborators has been involved in major studies funded by NIH that have spent a significant amount of effort developing data capture systems for these studies. Some groups have chosen to use commercially available software "off-the-shelf" for data entry. Others have endeavored to develop from scratch data entry systems tailored to the clinical setting. And still others have used commercially available software, but invested in a significant amount of programming effort to tailor the software to the clinical setting. This workshop will compare the merits of each of these three strategies, and in the process, describe many of the requirements that such data entry systems must satisfy

Chairperson: Ruth McBride, University of Washington William Amoroso, University of Pittsburgh Bonnie Hermanson, University of Washington David Christiansen, University of North Carolina

Discussant: Herman M~tchell, New England Research

Workshop Vl ~fftat Make~ a Good Clinioai Center: sp.dr~ org~n~aUon~ ~

Harbour Ballroom C

The attributes of a good Clinical Center include specific characteristics for Principal Investigators, Clinic Coordinators, and patients. The Workshop will focus on these three aspects of clinical trials organization and management. Presenters will describe characteristics of a Principal Investigator including issues about attention to the clinical trial versus other demands for research time, appreciation of the protocol, liaison with the Coordinating Center, and Clinic Coordinator. The characteristics of a Clinic Coordinator include loyalty to the study with compulsive attention to details of data collection forms, edits, and patient follow-up, as well as diligence merged with charm in dealings with patients, investigator and Coordinating Center. The typical "good" patient who is sought for clinical trials is one who follows the protocol, takes his medication regularly, attends appointments as scheduled, and calls to discuss problems rather than taking unilateral action. Panel and audience discussion will address methods to select appropriate staff and patients and how to improve ~technical

Final Program 249

TUESDAY, MAY 8, lgg0 (Continued)

2:00 PM - 3:30 PM Wodc~ho~ VI (Continued) ~ Bldlmom C What Makes ,, Good Cl ink~ Gent=~. ,sp=dr= ~ s~n~h~

and interpersonal skills among these three essential facets of a good clinical site.

Chairpe~cm: Joyce Cramer, VA Medical Canter, West Haven Malcolm Anderson, Duke Medical Center Davida Kruger, Henry Ford Hospital Bernard Zinman, Mt. Sinai Hospital Maureen McGrath, Pacific Medical Center

3.30 PM - 4:00 PM Break Mebo East

4:00 PM - 6:00 PM Renaly Session IV (,Sp~c~ Pan~ O~K==aon)

~ Ballroom

Sludying Treatments for AIDS: New Cludlenges for Clinical Trials

The evaluation of treatments for AIDS and HIV-related disorders has become a large-scale effort in a very short period of time. While the fundamental principles which drive the design, conduct and analysis of clinical trials are as applicable to AIDS as to other diseases, there is no question that we have been confronted with unusually difficult challenges in studying therapeutic approaches in this disease area. The purpose of" this session is to present a variety of perspectives on the special problems faced by all those concerned with the development of effective treatment strategies for HIV-infected populations.

The issues will be discussed from the following perspectives: clinical investigator, statistical center, FDA, industry, community, and the NIH.

Chaiq)erson: Susan Ellenberg, National Institutes of Health Henry Sacks, Mount Sinai Medical Center Dianne Finkelstein, Dana-Father Cancer Institute Ellen Cooper, Food and Drug Administration Sandy Lehrman, Burroughs Wellcome Company Jim Eigo, ACT UP Dan Hoth, National Institute of Allergy and Infectious Disaasas

7:00 PM

WEDNESDAY, MAY 9, 1990

8:00 AM - 12:00 PM

Coddail Banquet Harbour Ballroom

Guest Speaker: Samuel Greenhouse, Ph.D.

'~iews of Clinical Trials - Rashomon With Missing Data"

250 Final Program

WEDNESDAY, MAY 9, 1990

8'00 AM - 12:00 PM

8'00 AM - 10.00 AM

8:00 AM 36

8:20 AM 37

8'40 AM 38

9.00 AM 39

9 20 AM 40

9 40 AM 41

(Continued)

Exhibit Program Metro East

Cor~bu t~ Paper Se=~oe m (four [4] simultaneous sessions)

Ses~on IliA: Study D~ign: ~ 2/3 (X~aJity of L ~ e / D i = ~ e~e~etioe

Chaiqoerson: Peter Peduzzi

Design Issues in the Measurement of Quality of l ife in a Trial of Two Prenatal Diagnostic Technologies: Genetic Amniocen- tesis (GA) and Chorionic Villi Sampling (CVS); Marie Townsend, McMaster University

Cost and Quality of Ufe Outcomes in the TAMI 5 Trial: The Primacy of Secondary Endpoints; James Melton, Duke University

Design of an RCT of Prevention of Non-A Non-B Post- Transfusion Hepatitis when Incidence is Decreasing; Shelley Bull, Samuel Lunenfeld Research Institute

Experimental Design of the Hutchinson Smoking Prevention Project; Arthur Peterson, Fred Hutchinson Cancer Research Center

Chemoprevention: Challenges in Design and Conduct of Cancer Clinical Trials; Victor Vogel, University of Texas, M D. Anderson Cancer Center

The Effect of Partial Non-comphance on the Power of a Cancer Prevention Clinical Trial; Laurence Freedman, National Cancer Institute

8 00 AM 42

8 20 AM 43

8:40 AM 44

9'00 AM 45

~ lllEk Dala ~ a~d I~izal~

Hadaour Ballroo~ A

Chairperson: Kenneth Chu

Randomization in Clinical Trials Through a Computerized Network: Does it Improve the Data Quality?; Josette Renard, EORTC Data Center

PSM/PRS- Micro to Mainframe System for Protocol Selection, Eligibility Checking and Randomization; Thomas Zicarelli, Frontier Science and Technology Research Foundation

Evaluating the Need for Sending Paper Forms to the Data Coordinating Center (DCC) in a Clinical Trial with Distributed Data Entry; Jennifer Gassman, The Cleveland Clinic Foundation

Automating Database Changes Following Remote Data Entry: The BARI Automated Database Update System; Mary Ann Carr, University of Pittsburgh

Final Program 251

WEDNESDAY, MAY 9, 1990 (Continued)

8:00 AM - 10:00 AM CoeMbute¢l Paper Semaion III (Continued) (four [4] simultaneoue ~essione)

9:40 AM 59 Using State of Death to Identify False Positive National Death Index Matches; Gregory Evans, Bowman Gray School of Medicine

10:00 AM - 10:30 AM

10:30 AM - 12:00 PM Plenary Ses~on V Front=mac Ballroom

From Clinical Tdais to Clinical Practice: Generalizing from Parlicipant to Patient

Clinical trials, properly analyzed, have clear internal validity. The potential practical importance of a trial that shows therapeutic benefit of some treatment, however, lies not in its internal validity, but rather in the extent to which its results can be and are applied to larger populations. Recent discussions in the scientific and lay press concerning the applicability of the results of clinical trials studying cholesterol have sensitized many of us who work in clinical trials to the interface between trials and practice. This panel discussion will address several issues pertinent to the applicability of results of trials.

To whom can the results of a trial be generalized?

Do we typically generalize the results of trial too broadly or too narrowly?

What kind of evidence doesthe practicing M.D. need to apply the results of trials to practice?

What kind of evidence is compelling to people who run clinical trials?

What kind of evidence is convincing to the medical and scientific community not involved in clinical trials?

A~e there methods of designing and reporting trials that will encourage their appropriate applicability?.

Chairpemon: Janet Wittes, VA Medical Center, West Haven Hanna Rubins, VA Medical Center, Boston Charles Cowan, Opinion Research Corporation C.E. Davis, University of North Carolina Kent Bailey, Mayo Clinic

12.00 PM Adjournment

252 Final Program

WEDNESDAY, MAY 9, 1990 (Continued)

8:00 AM - 10:00 AM

9.20 AM 46

9 40 AM 47

8.00 AM 48

8:20 AM 49

8:40 AM 50

9:00 AM 51

9:20 AM 52

9:40 AM 53

ConldbulKI Pap~ ~ Ill (Continued) (four [4] simultaneous sessions)

Tracking, Accessing, and Reporting on Edit Queries; Paul Connor, Johns Hopkins University

Were the Results of CAST Due to Drug Packaging Error?; Clair Haakenson, VA Medical Canter, Albuquerque

~ Ii~: I ~ I ~ i t o r i r ~ II Hmtx~r Ballroom C

C~airpemon: Marian Fisher

Monitoring and Reviewing Drug-Safety Issues in a Blinded- Trial: A Sponsor's Perspective; Janis Goodlow, Bristol-Myers Squibb Company

A Survey of Data Monitoring Committees in NIH-Sponsored Multicenter Trials; Barbara Hawkins, Johns Hopkins University

Benefit of Early Identification of Ancillary Hypotheses; Mae Gordon, Washington University Medical School

The Early Termination, for Harm, of a Major Clinical Trial; Joel Verter, National Heart, Lung and Blood Institute

The Timing of Important Events Surrounding the Emergent (4/19/89) Alteration of a Clinical Trial: The Cardiac Arrhythmia Suppression Trial (CAST); James Myles, University of Washington

Continuation of the Trial of Anti-Hypertensive Interventions & Management: Design Features; Berry Davis, University of Texas School of Public Health

8:00 AM 54

8"20 AM 55

8:40 AM 58

9:00 AM 57

9'20 AM 58

~__o, ess!on__" IIID: Statistical Topics II Piers 4/5

Chaiq~e~o~: Blossom Patterson

A Comparison of Methods for Calculating a Stratified Kappa; William Barlow, University of Southern California

Analysis of Interrater Agreement on an Ordered Categorical Scale; Michele Melia, Johns Hopkins University

Application of New Two-Sample Tests to Data from a Randomized, Placebo-controlled Heart Failure Trial; P.K. Tandon; Sterling Research Group

Crossover Trialsin Chronic Disease; Stephen Walter, McMaster University

Methodological Issues in a Randomized Trial to Evaluate the Effects of Omega-3 Polyunsaturated Fatty Acids and Low Molecular Weight Heparin on Restenosis Following Percutaneous Transluminal Coronary Angioplasty; Lorrie Costantini, McMaster University