CONTRIBUTED WORKSHOPS PRESENTATIONS

Adherence/Compliance Issues

W1: METHODS FOR MEDICATION COMPLIANCE STUDIES: AN OVERVIEW OF THE ISPOR MEDICATION COMPLIANCE SIG GUIDELINES
Peterson AM¹, Nau D², Cramer JA^3
1University of the Sciences in Philadelphia/Philadelphia College of Pharmacy, Philadelphia, PA, USA; ²University of Michigan, Ann Arbor, MI, USA; ³Yale University, West Haven, CT, USA

Workshop Purpose: Participants will gain an understanding of newly developed guidelines for evaluating and conducting medication compliance studies using retrospective databases. These guidelines can assist researchers in their development of appropriate methods for measuring and analyzing medication compliance data and assist readers of compliance literature in evaluating the data

Workshop Description: Compliance with medications has become an increasingly important area of research as policy makers have perceived the extent and effect of the problem. Unfortunately, the wide range of measures and analytic methods have led to difficulty in evaluating and comparing research in this field. The purpose of the workshop is to share guidelines developed by members of the ISPOR Medication Compliance SIG Analysis Standards Working Group that will assist researchers in conducting appropriate analyses and evaluating the literature related to retrospective analysis of medication compliance. The guidelines focus on methodological issues such as study design, compliance measurement (e.g., Medication Possession Ratio, gap analysis), statistical analyses and presentation/discussion of findings. Members of the Analysis Methods Working Group will present and facilitate the discussion on the guidelines, and workshop participants will be encouraged to offer their perspectives and recommendations.
 

Clinical Study Methodology Issues

W2: SELECTION BIAS AND USE OF PROPENSITY SCORES IN OBSERVATIONAL STUDIES
Nordyke RJ
Cerner Health Insights, Beverly Hills, CA, USA

Workshop Purpose: Participants will learn how to identify selection bias, perform propensity score analyses, and critically appraise outcomes studies that use propensity score methods.

Workshop Description: Selection bias occurs in outcomes studies when the assignment to treatment/control groups is not at random. Thus, selection bias must be considered in all observational studies and analyses of secondary data. There are a number of methods to address selection bias in regression analyses. One of the easiest to use and most popular methods is the use of propensity scores. Propensity scores were first developed to deal with non-random assignment to treatment groups in education research and has been readily adopted in health outcomes research. Essentially, this method attempts to quasi-randomize a dataset by determining the probability that a study participant is treated conditioned on each participants measured characteristics, or covariates. This probability, or propensity score, may then used in several ways to better estimate the impact of treatment on study outcomes. This session will review the problem of selection bias and the fundamentals of propensity scores with specific instructions on how to do this work in STATA. We will present a worked example using hospital data demonstrate the use of propensity scores to 1) match; 2) stratify; 3) adjust covariates; and 4) weight regression estimates. In addition, propensity score methods and results will be compared to those of treatment effects and instrumental variable approaches.

 
Cost Study Methodology Issues

W3: IMPROVING YOUR SAMPLE WITH BOOTSTRAP METHODOLOGIES IN HEALTH ECONOMIC STUDIES: WHEN, WHY AND HOW?
Lamure M¹, Dansette GY², Castanier JC^2
1Lyon 1 University, VILLEURBANNE, France; ²IMS Health, Lyon, France

Workshop Purpose: The participants of this workshop will learn principles of bootstrap methodologies, their application in the field of health economics and in what circumstances they can or they must be used: particularly in case of feasibility studies prior to a real life observational study

Workshop Description: In a first part, concepts of bootstrap are presented in the context of health economic studies: -Monte Carlo methods for simulation and random number generation techniques -Parameters estimates –Confidence Intervals building -Hypothesis testing In a second part, simulated examples as well as real world examples illustrate in what circumstances and how bootstrap methods can be used, and what their benefit is. A special focus will be given on calculation of predictive indicators such as: -Proportion of at risk subjects in a sample of patients -Proportion of failures or relapses after a first line treatment.

 
Formulary Development Research Issues

W4: DESIGNING AND DEVELOPING BUDGET IMPACT MODELS SUITED FOR GLOBAL ADAPTATION
Smith TW, Tierce JC, Cherry SB
ValueMedics Research, LLC, Falls Church, VA, USA

Workshop Purpose: Participants in this workshop will gain perspective on common barriers to the successful development and cross-national adaptation of budget impact models (BIM) and learn sound practices for addressing these challenges. The focus will be on issues associated with design, development, and implementation stages that may impact the ease with which BIMs can be readily adapted for use in multiple markets.

Workshop Description: BIMs are increasingly used by healthcare decision-makers to evaluate the financial impact of adding a new product to a formulary. As manufacturers take an increasingly global perspective on product marketing they are forced to deal with the complexities inherent with simultaneous or sequential presentation of budgetary impact information to multiple reimbursement authorities in different countries. This workshop is designed to give researchers and research sponsors undertaking global BIM development an understanding of methods and practices that will improve the chances of a successful development effort. Audience participation will be encouraged in defining possible measures of success when developing BIMs as well as in outlining various factors that may complicate BIM development within their organizations. Presenters will employ a case study approach involving various global implementations of BIMs and will illustrate key concepts with concrete examples drawn from the underlying models. Participants will gain an understanding of how modular design can improve development efficiency and model transparency; the importance of ensuring adequate processes for communication, feedback, and revision throughout the project lifecycle; and how the creation of a well-documented framework for model adaptation can facilitate the implementation process.

 
Health Care Policy Development Issues

W5: STANDARDIZED REPORTING FORMAT FOR OUTCOMES RESEARCH OR ECONOMIC EVALUATION AND A SYSTEM TO MEASURE THE STRENGTH OF EVIDENCE OF STUDIES
Wong PK¹, Bakst A², Mutnick A³, Blackburn J^4
1Good Samaritan Hospital, Dayton, OH, USA; ²TAP Pharmaceutical Product Inc, Chicago, IL, USA; ³University of Virginia Health System, Charlottesville, VA, USA; ⁴University of Illinois at Chicago, Chicago, IL, USA

Workshop Purpose: The objectives of the workshop are to 1) discuss the proposed standardized format for reporting studies designed for clinicians; 2) explore the usefulness of a system to measure the strength of evidence of studies; 3) obtain opinions from ISPOR members for further refinement of the proposed standards; and 4) start building consensus among ISPOR members for the adoption of this format.

Workshop Description: In 2004, the International Society of Pharmacoeconomics and Outcomes Research (ISPOR) redesigned the Clinical Practice Special Interest Group (SIG) to focus on 4 areas of interest: 1) to develop clinician skills for interpreting, evaluating, and applying the economic analyses and outcomes research; 2) to translate research into practical recommendations; 3) to communicate research in a relevant, transparent, and comprehensible format for clinical decision-making; and 4) to promote outcomes research at other clinical meetings. Among the three SIG working groups the Standardized Reporting Working Group was charged with establishing a standardized format for reporting economic and/or outcomes research findings. Additionally, the group was to develop a ranking system that would delineate the strength of evidence for comparator studies. After a series of conference calls and follow-up discussion, the group drafted a proposal for review. In this workshop, the working group will present the proposed format for reporting research findings and will review a system who's purpose is to measure the strength of evidence of the studies. Participants will be divided into two groups, in order to provide comments and suggestions. This open forum will allow ISPOR members to provide their opinions while also serving as the beginning in a series of consensus processes for standardized reporting for clinicians. Comments and suggestions from members will be posted on the ISPOR website after the 10th Annual International Meeting.

 
QOL/PRO/Preference-based Studies/Methodology Issues

W6: ESTIMATING MINIMALLY IMPORTANT DIFFERENCE: IS IT REALLY MINIMAL?
Bala M¹, McLeod L^2
1Centocor, Inc, Malvern, PA, USA; ²RTI Health Solutions, Research Triangle Park, NC, USA

Workshop Purpose: The objective of this workshop is to discuss potential limitations of current approaches to estimating a minimally important difference (MID) for patient reported outcomes (PROs), and to propose a few new approaches

Workshop Description: Collection of PRO data in clinical studies is becoming increasingly common. However, there continues to be considerable debate regarding the interpretation of these data. Researchers have started using MID estimates to understand the clinical relevance of changes seen in PRO scores seen during clinical studies. In this workshop, we will provide an overview of several approaches commonly used to estimate MID, focusing primarily on those that are anchor-based. Currently used anchor-based MID estimation methods tend to ignore a large proportion of the data, and can lead to estimates that are not truly “minimal”. Further, these approaches often do not examine the variability of MID based on demographic or disease characteristics. We will discuss a few new anchor-based approaches to estimating MID that utilize all available data. Some of these approaches also allow us to compute multiple MID estimates based on baseline patient characteristics. We will also examine how to determine the proportion of subjects within a clinical study who have achieved a clinically meaningful improvement in a PRO, both from a deterministic as well as a probabilistic perspective. We will illustrate the approaches using data from a clinical trial. We will conclude with an interactive discussion regarding the merits of the different approaches to estimating MID.
 

Risk Assessment/Risk Management Issues

W7: BALANCING RISKS AND BENEFITS OF HEALTH CARE TREATMENTS IN THE 21ST CENTURY
Raisch DW¹, Liu LZ², Stang P³, Rahman A⁴, Lockett A⁵, Lipskiy N^6
1VA Cooperative Studies Program, Albuquerque, NM, USA; ²Pfizer Global Pharmaceuticals, New York, NY, USA; ³Galt Associates, Inc. and West Chester University, Blue Bell, PA, USA; ⁴Shenandoah University, Winchester, VA, USA; ⁵ICO, Leeds, W Yorkshire, United Kingdom; ⁶Surveillance Data Inc, Plymouth Meeting, PA, USA

Workshop Purpose: The participants of this workshop will be able to describe key risk management activities. The activities are: (1) risk/benefit identification, (2) implementation of interventions that mitigate risk, and (3) communication of risks and benefits to stakeholders. Participants will gain exposure to the experience of colleagues involved in risk management and be able to explain the state–of-the-art of these key activities.

Workshop Description: ISPOR's Risk Management Special Interest Group includes 3 sub-groups devoted to developing white papers on risk/benefit identification, risk mitigation interventions, and communication of risk/benefit. Important perspectives include those of health care providers, health payers, the pharmaceutical and device industries, and government. In the first part of the workshop, chairs of each sub-group will present overviews of the current literature and identify key past and present issues. Identification involves discernment, assessment, and balance of benefits and risks of health care treatments. Intervention incorporates the implementation of action plans which set the standards for acceptable risk, influence patient and provider behavior to mitigate risk and maximize benefit, and capture key follow-up information. Communication involves the identification of stakeholders and selection of optimal mechanisms for dissemination of risk/benefit information. These overviews will serve as a backdrop to an active participant discussion regarding successes and failures of risk management within these three key areas. This interactive session will revolve around an exchange of information and opinion about the nature of risk and management of risk. Discussion of the participant's experiences/case studies will be summarized and integrated.

SESSION II (7 workshops, 7 categories)
Monday, May 16, 2005, 3:15PM-4:15PM

Adherence/Compliance Issues

W8: MEDICATION COMPLIANCE (ADHERENCE): ISSUES AND DEFINITIONS CONSENSUS WORKSHOP 2005
Wong PK¹, Burrell A², Ollendorf DA³, Cramer JA⁴, Fuldeore M⁵, Fairchild C^6
1Good Samaritan Hospital, Dayton, OH, USA; ²Aventis Pharma, Bridgewater, NJ, USA; ³PharMetrics, Watertown, MA, USA; ⁴Yale University, West Haven, CT, USA; ⁵Walgreens Health Initiatives, Deerfield, IL, USA; ⁶Alcon Laboratories, Fort Worth, TX, USA

Workshop Purpose: The objectives of the workshop are 1) to evaluate the applicability of the terms that are accompanied with the consented definitions in various practice settings; 2) to identify issues related to these proposed terms and methods in conducting and evaluating research; and 3) to gather opinions for the further refinement of these definitions.

Workshop Description: The “Issues and Definitions Working Group” of the Medication Compliance Special Interest Group will use this workshop to provide ISPOR members with a forum for discussing consistency of definitions that should be used for medication compliance/adherence and persistence analyses, and implications for health outcomes. The ISPOR website provided members an opportunity to comment and vote on definitions in December 2004. Although consensus was reached for the compliance and persistence definitions, many key issues related to these definitions require resolution. At this workshop, The Working Group will update participants on the definitions for terms and conditions such as methods in calculating medication compliance, treatment holidays or other gaps in therapy, etc. After an introduction to the issues, participants will work in small group discussions to focus on specific items related to operationalizing compliance, adherence, and persistence definitions; illustrative examples of the operational challenges faced for each set of definitions will be provided. Each group will present their recommendations to all the participants for discussion. Key issues raised from this workshop will be addressed and recommendations generated from this workshop will be published at the ISPOR SIG website.

 
Clinical Study Methodology Issues

W9: PROVIDING INDIVIDUAL ESTIMATES OF LIFE-EXPECTANCY FOR ECONOMIC ANALYSES OF CLINICAL TRIALS
Ishak KJ¹, Proskorovsky I², Caro JJ^3
1Caro Research, Dorval, QC, Canada; ²Caro Research Institute, Dorval, QC, Canada; ³Caro Research, Concord, MA, USA

Workshop Purpose: Participants will learn 1) statistical techniques to extrapolate survival beyond the time window of available data, 2) how to individualize estimates to account for patients’ characteristics in the target clinical trial, including the course observed in the trial, and 3) how to use the estimates in a an economic evaluation.

Workshop Description: Economic evaluations often incorporate the impact of medications on the life expectancy of patients, either through life-years or QALY gained. Life-expectancy is the area under the full cumulative survival curve but, in practice, survival is hardly ever observed completely due to losses to follow-up and restricted study duration with consequent incomplete follow-up for most of the population. Few trials last until all participants have died. It is therefore necessary to extrapolate survival beyond the available time. Standard parametric survival analysis methods are often not appropriate for this purpose because variation in the underlying death hazards over time is not adequately described by any single parametric distribution. In this workshop, we will describe two approaches: the first uses a piecewise parametric method to model observed death hazards in a suitably defined time windows; the second employs fractional polynomials, a flexible regression technique that fits a single model to the hazards over time. To individualize the estimates, we show how to take into account patient characteristics and course during the trial. Cox regression models provide the effect of the characteristics of interest. To account for the non-proportionality of these effects we demonstrate two methods: for the piecewise approach, a separate Cox equation is fit for each time window; for the fractional polynomial approach, time-dependent effects are incorporated. We then show how the resulting equations can be used to predict the complete survival curve, which is then integrated numerically to estimate life-expectancy; and these estimates are then used in the economic evaluation. These methods are illustrated with actual analyses estimating life-expectancy in trials of prevention of atherothrombotic cardiovascular events.

Cost Study Methodology Issues

W10: ANALYSIS OF COSTS USING PATIENT LEVEL DATA FROM RANDOMIZED DESIGNS
Doshi JA, Glick HA, Polsky D
University of Pennsylvania, Philadelphia, PA, USA

Workshop Purpose: The objectives of this workshop are to familiarize participants with issues and common mistakes in the analysis of costs using patient-level data identified from our systematic review of 115 studies on randomized-trial based economic evaluations published in 2003. The workshop will illustrate the common mistakes and preferred analytic approaches specifically for the (1) analysis of patient-level cost data; (2) addressing censored and missing cost data; and (3) assessing the stochastic uncertainty around the cost-effectiveness estimates.

Workshop Description: Prospective economic evaluation of clinical trials is an increasingly important component of the clinical development program for new clinical therapies (e.g., treatments, behavioral interventions, and drugs). However, analysis of economic data from such studies has raised serious analytic issues. In this workshop, the participants will explore common mistakes and issues faced by analysts in the economic evaluation of medical interventions in randomized trials, including analytic approaches for the (1) analysis of patient-level cost data; (2) addressing censored and missing cost data; and (3) assessing the stochastic uncertainty around the cost-effectiveness estimates. We will summarize the common fallacies identified from our systematic review of 115 studies on randomized-trial based economic evaluations published in 2003 and will outline preferred analytic approaches to address these areas. The workshop will be practical in orientation and will routinely provide examples to illustrate the "do's" and “don't's”. Throughout the workshop, participants will be given the opportunities to review actual study examples, to identify mistakes, and to offer suggestions on the appropriate methods to deal with the issue.

Formulary Development Research Issues

W11: INTEGRATING AND IMPLEMENTING COST-EFFECTIVENESS ANALYSIS IN A FORMULARY MANAGEMENT SYSTEM AT A COMPREHENSIVE CANCER CENTER: PEARLS AND PITFALLS
Arbuckle RB, King KM, Adamus AT
The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA

Workshop Purpose: The participants of this workshop will have the opportunity to review the formulary submission and evaluation process instituted at a comprehensive cancer center and discuss the implications for incorporation of cost-effectiveness data in the decision making process.

Workshop Description: The pharmacy formulary is a preferred list of pharmaceutical products available for patient treatment within a healthcare institution or system. The purpose of a formulary is to maximize product availability for those drugs that provide benefit over similar drugs or reduce cost through market advantage. Historically, the decision for drug incorporation on to the formulary has been based on safety and efficacy data, contract pricing, or a combination of both. Recently, including cost-effectiveness analysis (CEA) information into the formulary development process has been gaining interest. Several health systems and organizations throughout the world have developed models for this process. This workshop will present the process used to incorporate CEA into the Formulary Management System (FMS) at a major cancer center. The discussion will include many of the successes and pitfalls encountered and plans for future revision of the process.

Health Care Policy Development Issues

W12: USING HEALTH OUTCOMES TO MAXIMIZE MARKET ACCESS
Tolley K, Ryan J
Mapi Values, Bollington, Cheshire, United Kingdom

Workshop Purpose: The participants of this workshop will gain an understanding of the challenges faced in achieving market access for a new drug in one of the largest European markets, the UK. In the absence of a National Institute for Clinical Excellence (NICE) recommendation at drug launch, insight regarding different strategies for using health outcomes to maximize market access will be presented.

Workshop Description: Recognizing that achieving reimbursement status is only one hurdle in market access, the workshop will discuss how health outcomes research can be used to maximize market access for a new drug. In particular, it will focus on how to use health outcomes research throughout drug development, launch and then post launch. A key aspect of the workshop will be how to use different communication platforms and messages to promote your health outcomes research to key customers, opinion leaders and decision-makers. A summary of key health outcome requirements in leading countries will be provided, before strategies for achieving market access are discussed. Using England and Wales as a case study, the workshop will use a real world example of a drug class perceived to be expensive but which has not undergone assessment by the National Institute for Clinical Excellence. Participants will have an opportunity to discuss the strategies proposed, as well as those strategies that they have used, with an emphasis on those that they have found successful.

 
QOL/PRO/Preference-based Studies/Methodology Issues

W13: USING CONJOINT ANALYSIS TO UNDERSTAND PATIENT PREFERENCES IN OUTCOMES RESEARCH – A PRACTICAL HOW TO GUIDE
Bridges JFP¹, Terris DD^2
1University of Heidelberg, Heidelberg, Baden-Wuerttembu, Germany; ²University of Heidelberg, Heidelberg, Baden Wuerttembe, Germany

Workshop Purpose: The participants of this workshop will develop a sound understanding of the benefits of using theory driven conjoint analysis and practical experience of using the methods to understand patient preferences for health care interventions.

Workshop Description: Governments and regulators are becoming increasingly concerned with patient preferences, yet the recent developments in this area are not commonly understood or utilized by health outcomes researchers. This workshop will demonstrate conjoint analysis, a tool that has a sound theoretical foundation in economics, by designing, piloting, conducting and analyzing a simple conjoint analysis. This will truly be a hands on experience that will allow participants to better understand the stages involved in a conjoint analysis and the economic concepts that underpin it.

 
Risk Assessment/Risk Management Issues

W14: ISPOR MEMBER PERSPECTIVES: HOW DO INDIVIDUALS PERCEIVE RISKS AND BENEFITS OF PHARMACEUTICALS IN COMPARISON WITH OTHER NON-MEDICAL ACTIVITIES?
Raisch DW¹, Lockett A², Stang P³, Rahman A⁴, Liu LZ⁵, Lipskiy N^6
1VA Cooperative Studies Program, Albuquerque, NM, USA; ²ICO, Leeds, W Yorkshire, United Kingdom; ³Galt Associates, Inc and West Chester University, Blue Bell, PA, USA; ⁴Shenandoah University, Winchester, VA, USA; ⁵Pfizer Global Pharmaceuticals, New York, NY, USA; ⁶Surveillance Data Inc, Plymouth Meeting, PA, USA

Workshop Purpose: The participants of this workshop will be able to describe: (1) how perceptions of pharmaceutical risks and benefits can be measured, (2) the relative importance of different types of information sources regarding risks and benefits, and (3) perceptions of ISPOR members regarding risks and benefits of pharmaceuticals, medical, and non-medical activities. Participants will provide examples from practice during the discussion.

Workshop Description: ISPOR's Risk Management Special Interest Group (RM-SIG) conducted a web-based survey of the ISPOR membership to determine the perceptions of risks and benefits of pharmaceuticals in comparison with other medical treatments and public activities. First, we will present initial results from the survey. The survey measured general perceptions of risks and benefits of pharmaceuticals, marketed and in-development. Specific types of pharmaceuticals, medical interventions, and non-medical activities were assessed. The information sources regarding risks and benefits were also compared. The second part of the workshop will be an open, active discussion of the implications of the survey. Participants will be asked to describe other important initiatives surrounding risk and benefit perception, quantification and communication. Third, we will propose alternative RM-SIG activities and/or ISPOR policy recommendations that integrate the results of the survey. This portion the workshop will incorporate significant issues from the discussion and give participants opportunities to describe examples from practice. In addition to clarifying important issues regarding risk/benefit perceptions, the participants will have opportunity to provide guidance for future activities of the RM-SIG.

 

SESSION III (6 workshops, 6 categories)
Tuesday, May 17, 2005, 1:30PM-2:30PM

Adherence/Compliance Issues

W15: METHODS FOR MEDICATION COMPLIANCE STUDIES: THE IMPORTANCE OF STATISTICAL DISTRIBUTIONS
Nichol MB¹, Gwadry-Sridhar F², Benner J³, Cramer JA^4
1University of Southern California, School of Pharmacy, Los Angeles, CA, USA; ²University of Western Ontario, London, ON, Canada; ³ValueMedics, Arlington, VA, USA; ⁴Yale University, West Haven, CT, USA

Workshop Purpose: This workshop will demonstrate the statistical properties of commonly used measures of medication compliance, and discuss implications for future research in this field.

Workshop Description: Compliance with medications has become an increasingly important area of research as policy makers have perceived the extent of noncompliance and the effect of this problem on important patient outcomes. One of the areas that can benefit from clarity is the methodology involved in measuring and defining compliance. Differences in distributional characteristics of the measures contribute to the difficulty researcher's encounter in compliance analysis. This workshop will illustrate the distributional characteristics of compliance measures commonly used in the literature, explore the implications of these characteristics, and recommend specific descriptive analyses that should be included in articles that report measures of compliance or persistence. Examples will be drawn from published and ongoing projects measuring compliance in a variety of medications and diseases. Workshop participants will be encouraged to offer their perspectives and methodological recommendations regarding future studies of medication compliance.
 

Clinical Study Methodology Issues

W16: DEVELOPING AND IMPLEMENTING AN INTERNATIONAL REGISTRY: GLOBAL CHALLENGES AND STRATEGIES
Gordon MJ, Vreeland MG, Noe L, Larson LR
Ovation Research Group, Highland Park, IL, USA

Workshop Purpose: Participants will gain an understanding of: 1) marketplace trends and regulatory requirements; 2) the logistical considerations unique to international programs; and 3) design principles and strategic opportunities for integrating outcomes measures.

Workshop Description: Growing demands from regulators – both domestic and international, as well as from global commercial markets, to better document the real-world effects of new and marketed products are contributing to the necessity for – and proliferation of – international patient registries. Although some tactics associated with US-based registries are applicable to international programs, there are numerous logistical components (i.e., patient confidentiality, data ownership and transfer, challenges associated with patient-reported data, etc.) unique to each country. From a project team / project management perspective, effectively managing a program with international scope requires considerable coordination and structure from the earliest stages of registry development. It is during this design stage that a unique opportunity exists to determine the scope of outcomes and pharmacoeconomic measures as well as strategic, global marketing initiatives.
Those attending this interactive workshop will leave with a better understanding of the intricacies specific to developing and implementing an international registry. In addition, case studies of international registries will be presented.
 

Cost Study Methodology Issues  

W17: CONDUCTING COST-EFFECTIVENESS ANALYSES ALONGSIDE CLINICAL TRIALS: LESSONS FROM THE TRENCHES
Sullivan S¹, Ramsey S², Hollingworth W³, Buxton M^4
1Univ of Washington, Pharmaceutical Outcomes Research and Policy Program, Seattle, WA, USA; ²Fred Hutchinson Cancer Research Center, Seattle, WA, USA; ³University of Washington, Seattle, WA, USA; ⁴Brunel University, Uxbridge, Middlesex, United Kingdom

Workshop Purpose: The purpose of this session is to explore many aspects of conducting cost-effectiveness evaluations of pharmaceuticals, surgical interventions and diagnostic strategies alongside clinical trials, ranging from technical challenges to practical issues.

Workshop Description: Three large, randomized clinical trials of surgical, diagnostic and pharmaceutical interventions that included pre-specified cost-effectiveness analyses will be presented and discussed. Audience participation will be incorporated into each of three presentations. We share lessons learned "outside the methodology box" for those conducting joint clinical and economic trials. First, we present the National Emphysema Treatment Trial, a multicenter trial of lung volume reduction surgery vs. medical therapy for Medicare enrollees with severe emphysema. We describe the design and results of the trial and discuss pragmatic issues related to beyond trial modeling and simultaneous publication. Then, we present the Seattle Lumbar Imaging Project - a multi-site randomized controlled trial comparing a rapid MR imaging with conventional x-rays for primary care patients with low back pain. The primary outcome was a back pain specific measure of physical function, secondary outcomes included resource utilization, costs, and health state preference measures which were combined in a cost-effectiveness analysis. Myriad issues related to CEA of diagnostic tests will be discussed. Finally, we present data from a multi-country randomized trial assessing regular use of inhaled corticosteroids in mild asthma. Because the trial was conducted in 33 different countries, issues were raised as to the relevance of the information to other countries. An analysis of trial data using unit cost estimates from 8 countries showed marked differences in cost-effectiveness ratios, particularly when productivity loss estimates were included. We conclude by engaging the audience to share current concerns with trials that involve cost-effectiveness analysis.
 

Formulary Development Research Issues

W18: FORMULARY DECISIONS: ASSESSING HARM, SHOWING BENEFIT, PROVING VALUE, FROM MANAGED CARE DATA
Shaya FT¹, Mullins CD², Wong W^3
1University of Maryland, Baltimore, MD, USA; ²Pharmaceutical Health Services Research, Baltimore, MD, USA; ³CareFirst BlueCross BlueShield, Baltimore, MD, USA

Workshop Purpose: This workshop will outline the value of managed care claims databases used for formulary development. The aim of the workshop is to provide practical tools to use claims data and interpret results used for safety, effectiveness, cost-effectiveness, value and clinical claims considered by formulary committees.

Workshop Description: Decisions made by P&T committees can affect access to therapies and outcomes in large populations. They are based on safety and efficacy information from clinical trials as well cost-effectiveness information from observational studies. There has been a growing interest in looking at managed care administrative claims databases to assess how actual utilization by specific populations affects safety and effectiveness evaluations. This hands-on workshop will examine practical issues and characteristics of managed care databases. We will outline their strengths and mention their limitations in answering research questions and supporting safety, effectiveness and value claims, with an emphasis on NSAIDs and COX-2 inhibitors. With active participation from the audience, we will discuss methods to develop medical and prescription data queries, build patient cohort, specify endpoints and choose analytical methods. Participants will be guided in methods to merge and match medical and prescription claims, and will identify validity issues related to properly specifying the data fields (e.g. primary, secondary and tertiary diagnoses, time of diagnosis, or billed charges versus paid charges etc ..). We will outline bias and confounding concerns, and demonstrate tools to handle them in retrospective claims data, as compared to prospective data. Participants will use and critique different methods, to show how they can lead to conflicting results. A case study will be used to interpret the answers to questions of value statement, cost-effectiveness, risk projection, compliance modeling and budget impact. We will assess bias, comparing and contrasting with the clinical trial design. By the end of the workshop, the participants will be able to apply the tools for research geared at formulary development.

Health Care Policy Development Issues

W19: DISCRETE EVENT SIMULATION FOR INFECTIOUS DISEASES: CONCEPTS AND APPLICATION
Möller J¹, Getsios D², Caro JJ^1
1Caro Research Institute, Concord, MA, USA; ²Caro Research Institute, Hammonds Plains, NS, Canada

Workshop Purpose: Workshop participants will gain an understanding of discrete event simulation (DES), which has been gaining prominence in epidemiologic and health economic modeling. Application of DES to infectious diseases and immunization policy will be highlighted.

Workshop Description: Given their potentially widespread use, assessing the impact of new and emerging vaccines for infectious diseases is critical. DES represents a novel and powerful approach to the modeling of infectious diseases. The workshop will describe key concepts in DES, outline its past use in modeling these illnesses, and demonstrate its application via a model of invasive meningococcal disease. The strengths and limitations of DES, especially in comparison with other modeling techniques, will be discussed. Potential applications to other infectious diseases, and specific issues that might need to be addressed in model development will also be discussed.

  
Risk Assessment/Risk Management Issues

W20: USE AND MISUSE OF FRAMINGHAM HEART STUDY EQUATIONS FOR CARDIOVASCULAR DISEASE RISK ASSESSMENT
Russell MW, Miller JD
Boston Health Economics, Inc, Waltham, MA, USA

Workshop Purpose: The purpose of this workshop is to introduce participants to the technical aspects of using Framingham Heart Study (FHS) risk equations used to predict incidence and outcomes of cardiovascular disease (CVD).

Workshop Description: FHS has contributed importantly to understanding the epidemiology of CVD and has played a vital role in defining and quantifying how risk factors factor into incidence and outcomes of the disease. In the past 15 years, numerous multivariate risk prediction equations have been developed from Framingham data that account for the complex and synergistic nature of CVD risk factors, which include high blood pressure, dyslipidemia, smoking, obesity, diabetes, as well as age and sex. These equations allow time-dependent prediction probabilities to be estimated that are generally valid over a range of 4–12 years for individuals aged 30–74 years. The equations continually appear in published literature and are now available on the internet as risk assessment “tools.” Despite their obvious utility and versatility, FHS equations are not necessarily easy to use. Moreover, they have inherent limitations that potentially could lead to their misuse. In this workshop, the technical aspects of using FHS CVD risk equations will be explored. Topics to be covered include (1) overview of available equations, (2) pros and cons of using various equations, (3) tutorial of how typical equations work, (4) limitations and potential for misuse of the equations, and (5) application of the equations to healthcare research and decision making. The workshop will provide a case study from current research as well as an audience-directed case study.

SESSION IV (6 workshops, 6 categories)
Tuesday, May 17, 2005, 2:45PM-3:45PM

Adherence/Compliance Issues

W21: UTILIZATION OF LATENT CLASSIFICATION ANALYSIS (LCA) IN PATIENTS WITH SCHIZOPHRENIA: IS IT A USEFUL TOLL FOR ADHERENCE CASEMIX?
Ahn J¹, McCombs JS¹, Shi L^2
1University of Southern California, Los Angeles, CA, USA; ²Eli Lilly and Company, Indianapolis, IN, USA

Workshop Purpose: The participants of this workshop will gain an understanding of the challenges faced in collecting and analyzing antipsychotic drug adherence among patients with schizophrenia. Insight will be shared regarding the value of latent classification analysis in terms of assigning antipsychotic adherence casemix memberships.

Workshop Description: Poor outcomes have been associated with non-adherence to antipsychotic medications, which may precipitate relapse and can result in significant increases in total costs of care. LCA is a versatile methodology, which can be used to identify non-adherence, partial adherence, and adherence casemix classes. This approach assumes a latent index, which is an unobservable variable used to distinguish latent classes, and multiple observable indicators can be used to estimate the index. Once the latent classes are identified by LCA estimation, the profiles of each latent class can be examined by cross-tabulations with various covariates such as demographics, clinical outcomes, and utilization measures. The concept of adherence casemix is emerging as an aid to understand population characteristics, as relate to medication adherence. This could potentially contribute to greater utility of these casemix classes to optimize medical intervention. But is it another attention-grabber or a concept that holds genuine promise? What pitfalls should be cautioned to ensure better assessment of patient's adherence to antipsychotic agents and to further ensure better use of adherence casemix in the clinical care of patients?

Clinical Study Methodology Issues

W22: TIME-VARYING EFFECTS IN SURVIVAL ANALYSIS: TREATMENT EFFECTS IN DISGUISE
Nitz NM¹, Shetty S¹, Harley C¹, Crown WH^2
1i3 Magnifi, Eden Prairie, MN, USA; ²i3 Magnifi, Auburndale, MA, USA

Workshop Purpose: Participants in this workshop will better understand the issues related to the application of survival analysis in outcomes research, with an emphasis on time-varying techniques in isolating treatment effects.

Workshop Description: The Cox Proportional Hazards model is one of the most commonly used methods in time to event analysis. The usefulness of this technique depends upon whether researchers fully examine the proportionality assumption. The average treatment effect estimated under the proportionality assumption may not accurately capture differences between treatment cohorts that change over time. Failure to consider the possibility of non-proportionality therefore may lead to erroneous conclusions. Using specific examples of health outcomes research, this workshop will identify potential factors that may cause differences in treatment effects and will review techniques to address non-proportionality. Participants will be engaged in a critical comparison of each technique and its impact on study results through the use of an applied example. Discussion and questions will be solicited throughout the presentation (rather than limited to a question and answer period at the end of the session). The health outcomes that will be assessed in the workshop will include treatment adherence, hospitalization and health events.

Cost Study Methodology Issues

W23: APPROACHES TO ESTIMATING MAXIMUM STUDY POWER FOR ECONOMIC ENDPOINTS IN FIXED SAMPLE SIZE DESIGNS.
Simpson KN, Chumney EC, Dismuke CL, Mazyck-Brown JF
Medical University of South Carolina, Charleston, SC, USA

Workshop Purpose: To introduce participants to the process of performing a power analysis for the economic endpoints when sample size has been fixed by clinical or programmatic constraints.

Workshop Description: Many clinical investigators would like to include economic outcomes as secondary endpoints in their proposals. These types of economic studies are often constrained to a maximum sample size defined by the clinical measures, and/or by design decisions related to the size of a demonstration program. In these cases the researcher responsible for the design of the economic parts of the study must perform a power analysis to ensure that it is possible to find meaningful and statistically significant differences for the economic outcome measures which will be used. Power analysis is especially important for the development of protocols for piggy-back economic studies in clinical trials and for proposals for program evaluation studies. The power estimation is complicated by the economic analysis methods which may be required. Transformation of cost data, variance reduction based on baseline difference in economic risk factors, and/or adjustment for sample selection bias in resource use measures complicate the estimation of study power. We will demonstrate the power analysis process using real examples from clinical trials and program evaluations, and work through the conceptualization and data requirement specification of a power analysis problem using an example from workshop participants. The effects on study power imposed by specific underlying distributions in economic data, effects of study exclusion criteria, and impacts of baseline collection of predictors of heterogeneity in economic risk factors, as well as the effects of the statistical methods used to reduce variance or bias will be illustrated.

Health Care Policy Development Issues

W24: WHY DECISION MAKERS COULD PREFER BAYESIAN NETWORKS AND INFLUENCE DIAGRAMS OVER DECISION TREES?
Baio G¹, Jansen JP^2
1MIT-Harvard Health Science & Technology Division, Cambridge, MA, USA; ²MAPI VALUES, Houten, Netherlands

Workshop Purpose: Participants will be introduced to Decision Theoretic Networks. Particular focus is on the applications of such extended decision models for pharmacoeconomic evaluation and decision-making. The workshop is intended to provide insight in the advantages of the use of such models in terms of understanding the problem, simplicity of representation, and power of analysis.

Workshop Description: In this workshop Bayesian Networks and Influence Diagrams are introduced as an alternative to classical decision making tools. These Decision Theoretic Networks consist of a graphical structure and a set of conditional probability distributions. The graphic side of Decision Theoretic Networks provides an intuitively appealing interface by which highly-interacting sets of variables can be modeled making the problem under study explicit for decision makers. Moreover, Decision Theoretic Networks have the advantage of allowing a more compact representation than Decision Trees. Another advantage is that decision makers can combine their expert prior opinion or information with empirical evidence in a very direct way. The workshop examines the versatility of the instrument as an extremely powerful modeling tool by presenting basic theory and applications in different settings, including clinical practice and economic evaluation. Participants will be able to access expert opinion on Bayesian Networks methodology and will be encouraged to share their own assessment of its practical utility.

QOL/PRO/Preference-based Studies/Methodology Issues

W25: DATA ANALYSIS WORKSHOP: BETTER WAYS TO EXPLORE AND ANALYZE OBSERVATIONAL DATA
Potter L
Ovation Research Group, San Francisco, CA, USA

Workshop Purpose: This workshop describes and demonstrates three powerful statistical techniques for analyzing observational data. A case study approach is used to demonstrate better ways to explore the structure of questionnaire data, evaluate cost-effectiveness, and perform group comparisons in non-randomized studies. Participants will gain an understanding of how and when to apply these approaches in the context of real world research.

Workshop Description: Using a hypothetical case study constructed from several actual observational research studies, this workshop presents three examples of innovative approaches to analyzing data. The first section demonstrates how variable clustering can be used to construct simple composite scores from numerous scales and items. In the second section, bootstrap sampling allows the estimation of confidence intervals around cost-effectiveness ratios. The final section shows how propensity scoring can be used to control for group differences in non-randomized comparisons between treatments. This workshop highlights specific data analytic techniques and addresses the relative strengths and weaknesses of each approach. Audience participation is encouraged through guided discussion for each section of the case study and its corresponding analytic technique.

Risk Assessment/Risk Management Issues

W26: RISK-BENEFIT ASSESSMENT AND COST-EFFECTIVENESS ANALYSIS: DIFFERENCES, SIMILARITIES, AND SYNERGIES
Menzin J¹, Neumann P², Cohen JT^2
1Boston Health Economics, Waltham, MA, USA; ²Harvard University, Boston, MA, USA

Workshop Purpose: To provide an overview and discussion concerning how risk-benefit assessment and cost-effectiveness analysis relate to one another and to identify possible synergies.

Workshop Description: In recent years, concerns about the risks associated with certain pharmaceutical products have continued to mount. Quantifying these risks poses numerous challenges. Moreover, a careful weighing of risks and benefits is required for sound decision making. The field of risk-benefit assessment offers useful tools for this purpose. It is also clear, however, that the risks associated with the use of drug therapies need to be incorporated into formal cost-effectiveness analyses. A major question is how analysts can identify drug risks and quantify them in both health and monetary terms. In this workshop, we will provide an overview of risk and risk perceptions in the context of health and non-health interventions. We will then compare and contrast the objectives and general methods behind both risk-benefit assessment and cost-effectiveness analysis. We will identify areas of overlap and possible synergies associated with drawing from tools available from both disciplines. Special attention will be paid to examples in which the use of risk-benefit techniques could strengthen cost-effectiveness analysis.