Workshops Session I - Sunday, 21 October, 15:30 – 16:30

W1: OUTCOMES RESEARCH- ITS ROLE IN RISK MANAGEMENT
Discussion Leaders: John Parkinson, PhD, Head, GPRD, MHRA, London, UK; Tjeerd P Van Staa, MD, PhD, Head of Research, GPRD, London, UK; Frank De Vries, PhD, Lecturer, Utrecht University, Institute for Pharmaceutical Science, Utrecht, Holland, The Netherlands.
Workshop Purpose: Outcomes (effectiveness) research and drug safety research are often seen as separate and are undertaken by different groups or departments. They are however highly complementary if, as is becoming increasingly important, risk and benefit form part of RISK MANAGEMENT PLANS for new medications. The workshop will show how risk and benefit data can be integrated in a way that meets the needs of clinical care as well as public health.
Workshop Description: Risk Management Plans provided at the time of drug licence application are now required in EU countries for new products and for major changes to existing licences. A major part of such a plan relates to person level exposure and event data that can be used to demonstrate how the drug performs in real world patients; patients who often have a very different range of characteristics from those in the phase 3 clinical trials. Performance of a drug naturally embraces both safety and effectiveness and, although, Risk Management Plans are require by medicine regulators, with particular reference to safety/risk, there is the obvious trade-off between the risks and the benefit offered by the drug. A trade-off made at different levels depending upon the related morbidity and mortality of the treated disease. On this basis those involved with outcomes/ effectiveness research have a naturally part to play in such Risk Management Plans. The workshop will cover 1) how outcomes data can be incorporated 2) the trade-off equation 3) the need for “absolute” – excess risk and benefit outputs 4) managing the trade off when some events may be short term and others long term. To end the workshop participants will, in groups, work up an outline protocol for a Risk Management Plan covering both risk and benefit. The results from each group will be discussed and summarized.

W2: RISK FACTORS THAT CHANGE OVER TIME: HOW TO ESTIMATE THEIR IMPACT ON THE OCCURRENCE OF HEALTH OUTCOMES?
Discussion Leaders: K Jack Ishak, PhD, Director, Biostatistics, Caro Research Institute, Montreal, QC, Canada; Krista F. Huybrechts, MS, Senior Researcher, Caro Research Institute, Concord, MA, USA; J. Jaime Caro, MDCM, FRCPC, FAC, President & Scientific Director, Caro Research Institute, Concord, MA, USA.
Workshop Purpose: Many risk factors change over time, making quantification of their impact challenging. Rather than including only their baseline value, it is important to consider their time-varying nature, which requires proper definition of the period over which the variable is measured. In this workshop, we will describe factors that should be taken into account when defining a time-dependent variable in the context of time-to-event analyses, and illustrate potential biases that may occur when incorrect definitions are employed. The concepts will be illustrated using data from studies of the effect of compliance and persistence on adverse health outcomes.
Workshop Description: In analyses relating the effect of a time-dependent variable to the risk of outcomes, a decision must be made about the window over which the time-dependent variable should be defined at any given time. For example, in compliance analyses, compliance may be defined as cumulative usage from the start of therapy, or over a shorter window like the last 12 months. The length of the window should reflect the duration as well as any latency of the effect of the variable on the risk of the outcome. Thus, if a variable has a “long-lasting” effect, the window should be defined over a long enough span to include the earliest values that might have an effect on the current risk of the outcome. On the other hand, for variables with a very acute effect, only the most recent values should be included. Similarly, if the variable is “slow-acting”, the current risk of the outcome is unlikely to be affected by recent values, and so, the window should stop some time before the current point. Examples based on compliance and persistence analyses will be used to illustrate that misspecification of the window can lead to biased estimates of the effect of the variable due to the inclusion of values not related to current risk, and also distort inferences about the time-dependence (or proportionality) of the effect.

COST STUDY METHODOLOGY

W3: THE USE OF EPISODES OF CARE TO DEFINE AND MEASURE UNITS OF HEALTH CARE COSTS
Discussion Leaders: William Crown, PhD, President, i3 Innovus, Waltham, MA, USA; Steven Wickstrom, MS, Senior Director of Research and Development, Ingenix, Eden Prairie, MN, USA.
Workshop Purpose: Discuss the pros and cons for use of episode groupers to define units of medical services.
Workshop Description: Diagnosis-related groups (DRG) are widely used as a case mix adjustment method to classify inpatient stays based on the patient's condition and intensity of care. Costs and services associated with inpatient care are bundled into clinically meaningful episodes that are comparable across facilities and health plans. Unfortunately, DRGs can not be applied to outpatient care so current outcomes research does not generally group administrative data into episodes of care. The insurance industry has developed episode groupers for actuarial risk assessment. Groupers are also used for case mix adjustment to evaluate the effectiveness of health care delivery. Claims for a condition are bundled beginning with an anchor record until the condition is resolved as indicated by a clean period, a transition to another condition or yearly for chronic conditions. Both medical and pharmacy claims are bundled so total medical costs are aggregated.

Methods used to define episodes of care will be introduced. Handouts describing the coding hierarchy will be disseminated. Examples will illustrate potential advantages of defining costs using episodes of care over methods currently used by outcomes researchers. Workshop participants will discuss: 1) Is there a need to more uniformly define costs associated with patient conditions and severity of disease? 2) Do groupers address this need? 3) What are the current alternatives for defining episodes of care? 4) Barriers to further implementation and potential solutions 5) Further applications of groupers for outcomes research.

HEALTH CARE POLICY AND FORMULARY DEVELOPMENT USING OUTCOMES RESEARCH

W4: CONDITIONAL REIMBURSEMENT FOR HOSPITAL BASED DRUGS IN THE NETHERLANDS: POLICY MEASURE, COST-EFFECTIVENESS CRITERIUM, FUTURE OUTCOMES RESEARCH, METHODOLOGICAL AND CLINICAL CHALLENGES.
DISCUSSION LEADERS: Gepke Delwel, PhD, Policy Advisor, CVZ Healthcare Insurance Board, Diemen, The Netherlands; Johan L Severens, PhD, Professor of Health Technology Assessment, Academic Hospital Maastricht, Department of Clinical Epidemiology and Medical Technology Assessment, Maastricht, The Netherlands; Carin A Uyl-de Groot, PhD, Professor, iMTA/VUMC, Rotterdam, The Netherlands.
Workshop Purpose: Outline and discuss the methodological and clinical challenges and bottlenecks we face in the conduction of outcomes research for hospital based drugs. Results from outcomes research should enable us to determine the cost-effectiveness of the drug and the efficiency of prescribing the drug in daily clinical practice. Two criteria needed for the reimbursement decision.

Workshop Description: A policy measure for additional funding of expensive hospital based drugs has recently been issued. In order to get conditional reimbursement, applicants such as hospitals/ physicians have to compile data demonstrating the therapeutic added value of the drug, it's predicted costs and its predicted cost-effectiveness. In addition, a proposal for outcomes research is obligatory. Based on outcomes research, i.e. the collection of data from daily clinical practice during 3 years, applicants should assess the cost-effectiveness of the drug in clinical practice and the efficiency of prescribing the drug in hospitals. After 3 years of conditional reimbursement, a re-appraisal of the therapeutic added value, the real costs, and the cost-effectiveness of the drug takes places. Currently, 10-15 drugs are conditionally reimbursed. Outcomes research proposals have been appraised and outcomes research has started for these drugs. In a working party we are addressing issues that are relevant and pragmatic to investigate in outcomes research. We identified cost, clinical and patient-reported outcomes, patient characteristics; data collection sources, and the methods of data analysis, with a strong focus on the modelling approach. Since outcomes research is a new era, we started two pilot studies in order to get insight in real life clinical and methodological issues and the feasibility of different approaches. In the workshop, we will share and discuss our experiences on this topic with the audience.

W5: VALUE AND COST-EFFECTIVENESS: US AND EUROPEAN PERSPECTIVES
DISCUSSION LEADERS: Sean D. Sullivan, PhD, Professor, University of Washington, Department of Pharmacy, Seattle, WA, USA; Brian W. Bresnahan, PhD, Senior Fellow, University of Washington, Department of Pharmacy, PORPP, Seattle, WA, USA; Peter J. Neumann, ScD, Professor, Tufts-New England Medical Center, Institute for Clinical Research and Health Policy Studies, Boston, MA, USA; Louis P. Garrison, PhD, Professor of Pharmacy, University of Washington, Department of Pharmacy, Seattle, WA, USA.
Workshop Purpose: To explore if there are any lessons from the payer experience in the U.S. for European payers and health systems?
Workshop Description: The faculty will present evidence from i) a recently completed structured, qualitative telephone survey of U.S. health plans and related stakeholders focused on value perceptions and evidence needs, and ii) a quantitative assessment comparing literature-based cost-effectiveness estimates with actual health plan coverage decisions for pharmaceutical products.
Telephone survey interviews were used to collect data from fifteen private U.S. payer representatives/experts. Key findings suggest that health plans are most interested in clinical safety, efficacy, and effectiveness evidence. Payers' demands are increasing for more clinical utility and clinical performance studies providing comparative information for medical products. Cost and cost-effectiveness were identified as important but secondary considerations for coverage and reimbursement decisions.

The Tufts-NEMC cost-effectiveness registry contains 1587 cost-utility (CU) ratios from 1998-2003, from 555 separate cost-analyses. The data analyzed for this study included 627 CU ratios pertaining to pharmaceuticals from 221 drug-related CU assessments and representing 144 distinct drugs. For purposes of comparison, we examined publicly available formularies of nine health care providers. In numerous cases, drugs with favorable data from a value or cost-utility standpoint (i.e., drugs associated with a CU ratio under $50,000/QALY) were placed on unfavorable tiers or non-preferred list. There were also numerous instances of drugs with unfavorable cost-effectiveness that received favorable formulary placement.
The discussion leaders will present these study results, identify potential lessons (30 minutes), and engage the workshop participants in a discussion of their potential relevance in Europe. Participants representing three country perspectives (France, Germany, and the UK) will be invited to make five minute commentaries on the relevance of the US experience for their settings, and there will be approximately 10-15 minutes for general audience discussion.

PATIENT-REPORTED OUTCOMES STUDY METHODOLOGY (INCLUDING PREFERENCE-BASED & COMPLIANCE STUDIES)

W6: INTERVENTIONS THAT ENHANCE MEDICATION COMPLIANCE AND PERSISTENCE: THEORETICAL GROUNDING, AND ECONOMIC ASSESSMENT
DISCUSSION LEADERS: Dyfrig Hughes, PhD, Senior Research Fellow, University of Wales, Bangor, Centre for Economics and Policy in Health, Bangor, UK; Femida Gwadry Sridhar, PhD, RPH, MSc, Assistant Professor, University of Western Ontario & McMaster University, ON, Canada; Rachel Elliott, PhD, Professor, University of Nottingham, Division of Social Research in Medicines and Health School of Pharmacy, Nottingham, UK; Warren Cowell, MSc, Health Economist, Roche Products Ltd, Healthcare Management, Welwyn Garden City, UK; Sangeeta Budhia, PhD, Consultant, Heron Evidence Development, Letchworth Garden City, Hertfordshire, UK.
Workshop Purpose: Participants will learn about: 1. The theoretical, including economic, basis for the development of interventions aimed at enhancing patient compliance and persistence. 2. How to assess which interventions are clinically- and cost-effective on the basis of their impact on compliance and persistence.
Workshop Description: Clinicians and researchers should understand the determinants of non-compliance in order to effectively target interventions that impact these determinants. Planning interventions requires a theoretical framework. Demonstrating the efficacy of interventions requires robust and unbiased data on compliance and persistence. However, demonstrating the clinical- and cost-effectiveness of interventions requires evidence their impact on health outcomes. Presenters in this workshop will provide an overview of important questions from a methodological and economic perspective to address the following areas: 1. Do interventions that are reproducible and theoretically grounded have greater effect on compliance? 2. Is there evidence that interventions vary based on treatment of a chronic or acute condition? 3. Do interventions result in improved outcomes by improving compliance? 4. Do the improvement in outcomes justify the costs? 5. What economic considerations should future studies consider when implementing interventions? 6. Is there evidence on how often should interventions be delivered? (i.e. Does increased frequency result in better outcomes?) The audience will be encouraged to participate in discussion of each issue.

W7: COMPLIANCE, ADHERENCE, AND PERSISTENCE IN HEALTH OUTCOMES RESEARCH: WHY, WHEN, AND WHAT'S THE RIGHT CHOICE
DISCUSSION LEADERS: B Tang, MD, PhD, Associate Director, Centocor, Inc, Horsham, PA, USA; M I Rahman, MD, MPH, Senior Director of Pharmacoepidemiology, Centocor, Inc, Horsham, PA, USA
Workshop Purpose: To explore the importance, methodology, and limitations of assessing treatment compliance, adherence, and persistence in health outcomes and clinical studies.
Workshop Description: Treatment compliance, adherence, and persistence are important measures of treatment effectiveness in real world practice, and have been used in many health outcomes studies. However, the terminologies of compliance, adherence, and persistence are not consistent in many studies. Some researchers have advocated differentiating among these terms.
Compliance measures the degree of constancy and accuracy with which a patient follows a prescribed regimen, as distinguished from adherence or maintenance. Adherence is a term used to describe how well a patient or client is sticking to the recommendations made to them by their health professional. While persistence measures the duration of therapy or continuation of a therapy, including time metrics in the calculation, other methods, such as medication possession ratio, gaps, discontinuation, and switching are also used in some studies.
This workshop will discuss the importance of treatment patterns, and how to use appropriate methodology in real world clinical studies. The methods of assessing therapy compliance, adherence, and persistence in prospective and retrospective studies will be compared and the pros and cons will be discussed. The choice of using the correct measure(s) depends on the nature of the disease, objectives of the study, and the therapies studied.
The appropriate statistical techniques to be used and limitations in observational studies will be discussed. Examples of different measures of treatment patterns in chronic diseases (such as hypertension, cardiovascular diseases, and diabetes) and different data (randomized clinical trials versus observational studies, and retrospective databases) will be provided. At the end of the workshop, attendees should be able to understand how to apply the appropriate methodology to specific outcomes studies and are encouraged to bring their own research questions for discussion.

CLINICAL OUTCOMES STUDY METHODOLOGY
W8: USE OF ELECTRONIC MEDICAL RECORD (EMR) DATA IN OUTCOMES RESEARCH
DISCUSSION LEADERS: Gary M Oderda, PharmD, MPH, Professor, University of Utah, College of Pharmacy, Department of Pharmacotherapy, Salt Lake City, UT, USA; Michael Lieberman, MD, MS, Informatics Director, GE Medical Systems, Hillsboro, OR, USA; Carl V. Asche, PhD, Research Associate Professor, University of Utah, College of Pharmacy, Outcomes Research Center, Department of Pharmacotherapy, Salt Lake City, UT, USA; Diana Brixner, RPh, PhD, Associate Professor and Chair, Department of Pharmacotherapy, Executive Director Pharmacotherapy Outcomes Research Center, University of Utah, Salt Lake City, UT, USA
Workshop Purpose: To introduce and expand the knowledge and experience of attendees in using Electronic Medical Records (EMR) data in outcomes research. This will include discussion of data available in an EMR, comparison of EMR and claims data, and strengths and weaknesses of using EMR and claims data in outcomes research
Workshop Description: The technology and sophistication of health care utilization databases have expanded over the last decade to include results of lab tests, vital signs, and other clinical information. The richness and depth of information on “real world” use of health services for large population-based patient cohorts at relatively low cost render such databases an attractive resource for outcomes research. These data may be used to explore variation in health care provision and assess the impact of variation on clinical and economic outcomes.
EMR data includes clinical measures not generally available from claims data including clinical outcomes, laboratory values, weight, BMI, non-prescription drug use and vital signs. Drug orders are present but information on prescription fills is not. Data is available from all payers within days of entry into the medical record, and is generally available for a longer period of time than most claims data. This makes using EMR data an excellent choice to answer some questions but not others. The workshop will discuss these important differences in using EMR data including limitations and strategies to maximize its value, utilizing an interactive format with interactive examples. This will be done primarily through presentation and discussion of contemporary examples of studies done in a national US EMR. Examples will be presented and discussed with the audience in a variety of therapeutic areas including diabetes and hypertension.

W9: THROUGH A GLASS DARKLY: TRANSPARENCY IN DECISION-ANALYTIC MODELS
DISCUSSION LEADERS: Nancy Neil, PhD, Senior Director, Health Economics, ICON Clinical Research, Lifecycle Sciences Group, San Francisco, CA, USA; Michael Chambers, MA, MSc, Director of Health Economics, GlaxoSmithKline, Global Health Outcomes, Uxbridge, Middlesex, UK.
Workshop Purpose: The purpose of this workshop is to explore, understand and discuss the role of transparency in decision-analytic models and to identify specific modeling approaches, features and techniques that can influence a model’s clarity.
Workshop Description: The health care industry is increasingly embracing decision-analytic models as an aid to decision making. Although the need for, and value of transparency in these models is widely recognized, researchers often have trouble finding practical guidance in the literature about what it is that can make or break a model's clarity. The purpose of this workshop is to discuss how the term “transparent” is used in the context of decision-analytic models, and to consider what people are looking for when they call for transparency. Using screen shots from finished models and drawing upon principles from cognitive science, interaction design and software user-interface research we will consider how, from a practical perspective, we might collectively begin to raise the standards for transparency in models used for health care decision making. By more clearly defining the nature of transparency and initiating a conversation about practical ways of achieving it, we will be helping our community not only to build better models, but to build the kind of models that are more likely to be used to support decision making.

W10: ADHERENCE TO CLINICAL GUIDELINES: A KEY CONCEPT IN CLINICAL AND ECONOMIC MODELS FOR DIAGNOSTIC DEVICES
DISCUSSION LEADERS: Won Chan Lee, PhD, Director, Abt Associates Inc, Health Economic Research and Quality of Life Evaluation Services, Bethesda, MD, USA; Yu-Chen Yeh, MS, Lead Pharmacoeconomist, Abt Associates Inc, Health Economic Research & Quality of Life Evaluation Services, Lexington, MA, USA; Mary Jo Williams, BS, Principal, Health Economics, Medtronic, Gastroenterology and Urology, Shoreview, MN, USA.
Workshop Purpose: The primary goal of this workshop is to help participants understand the importance and value of integrating adherence to clinical guidelines into the conceptual framework of such models, and how this concept can be successfully weaved into model features unique or particularly relevant to diagnostic devices.
Workshop Description: Participants will actively react to case studies, thereby facilitating discussion and debate. Using primarily case-study discussions and debate among workshop participants led by workshop leaders, this session will teach: (1) why it is important to incorporate adherence to diagnostic guidelines into models of diagnostic devices; (2) how to successfully do it; (3) what obstacles to avoid; and (4) what impact it can have. Significantly, these discussions will be moderated to bring in important aspects related to the value of diagnostic devices, such as (1) sensitivity and specificity; (2) impacts of false positives and false negatives on clinical prognostics and associated economic costs; (3) trade-offs between cost savings related to treatments and cost increases in diagnostic tests; (4) timing of tests and other procedures; (5) threshold analysis; (6) patient compliance variation; and (7) patient and provider preferences. Examples will be taken from gastrointestinal and other disease areas.

HEALTH CARE POLICY AND FORMULARY DEVELOPMENT USING OUTCOMES RESEARCH

W11: FACTORS TO CONSIDER IN SELECTING A STUDY DESIGN FOR POST-LAUNCH RESEARCH – STRATEGIC, METHODOLOGIC AND PRACTICAL CONSIDERATIONS
DISCUSSION LEADERS: Deborah Marshall, PhD , Vice President, Global Health Economics and Outcomes, i3 Innovus, Burlington, ON, Canada; Monique Martin, MBA , Health Economist, i3 Innovus, Health Outcomes, Uxbridge, Middlesex, UK; Mike F Drummond, PhD , Professor of Health Economics, Centre for Health Economics, University of York, York, UK.
Workshop Purpose: To define and apply a framework to inform the selection of the optimal study design for post-launch studies involving health economics that includes the methodological, operational and strategic perspective.
Workshop Description: There is increased interest in post-launch economic studies as more jurisdictions require economic data for the formal decision process of pricing and reimbursement of drugs. Much of the data requested, such as provider and patient compliance, cannot be obtained in regulatory trials. For example, in a review of NICE pharmaceutical appraisals 47 of 48 contained one or more recommendation(s) for post-launch research. The primary motivation for industry to undertake these studies is to strengthen the market position of their product by demonstrating product benefits over the competitors or justifying the price by demonstrating ‘value for money'.
There are numerous study designs, including pragmatic clinical trials and prospective observational cohort studies (registries) for collecting health economic data in post-launch studies. But what criteria should be used to decide the optimal approach? Although pragmatic clinical trials with randomization had been advocated as having a good balance between external and internal validity, opportunities for undertaking them are limited, and from a practical perspective, they require considerable resources. Registry studies have inherent limitations of observational studies and care must be taken in their analysis and interpretation.
In this workshop, we first discuss the motivation for conducting these studies, and the benefits and drawbacks of different study designs. Then we present a framework including strategic, methodological and practical considerations in selecting an optimal study design to answer a specific research question. Issues will be highlighted using case examples. Discussion and questions will be solicited from the audience throughout the presentation with brief exercises. The last component of the session will be an interactive exercise for the audience to apply the contents of the workshop to decide on and develop a post-launch real world study design for a specific example.

PATIENT-REPORTED OUTCOMES STUDY METHODOLOGY (INCLUDING PREFERENCE-BASED & COMPLIANCE STUDIES)

W12: THE IMPLICATIONS OF VALUE BASED PRICING FOR THE PRICING AND REIMBURSEMENT OF PHARMACEUTICALS.
DISCUSSION LEADERS: Keith Tolley, MPhil, Research Director, Mapi Values, Market Access Unit, Bollington, Cheshire, UK; Francis Pang, MSc, Phil, DM, European Health Economics Manager, Takeda Pharmaceutical Company Limited, Global Research and Development Centre (Europe), London, UK; Ulrich Oron, MSc, RPh, MBA, Associate Director Market Access, Mapi Values Netherlands, Houten, The Netherlands.
Workshop Purpose: To present the core principles of value based pricing of pharmaceuticals and to explore the implications for its application in European countries.
Workshop Description: The recent 2007 report by the UK Office of Fair Trading (OFT) into the UK profit regulating scheme, the PPRS, concluded this was a blunt tool for pharmaceutical price control. The report recommended that cost-effectiveness criteria, in particular incremental cost per QALY gained, as the metric for assessing the value of pharmaceuticals in order to set an efficient market price. In this system, existing HTA bodies (NICE, Scottish Medicines Consortium) would evaluate cost-effectiveness and a separate body would set the reimbursed price based on this evidence. Premium prices could be secured if a company demonstrates a new product is innovative, in particular meets an unmet health need. This value based pricing (VBP) system is controversial as the UK is a ‘free pricing' country – a concern is that such a system could at best slow down acceptance of innovative pharmaceuticals in the UK. The workshop will describe the main features of VBP and explore the rationale behind the main recommendations in the OFT report, and the potential impact on market access in the UK. Key questions will be explored, for example, if cost per QALY is used to help set prices, could this lead to threshold pricing? Other European countries already operate types of VBP system as part of new product pricing and reimbursement negotiations with pharmaceutical companies, but with focus on therapeutic value rather than cost-effectiveness (e.g. France, Italy). The future implications for VBP utilising cost-effectiveness evidence for countries other than UK will be explored in the workshop. A debate with audience participation will be held in order to assess costs and benefits of VBP from health care decision maker versus pharmaceutical company perspectives.

W13: SPECIFYING THE CONCEPTUAL FRAMEWORK OF AN INSTRUMENT: TWO TOOLS TO MAKE YOUR LIFE EASIER
DISCUSSION LEADERS: Pennifer Erickson, PhD, Co-founder, OLGA, State College, PA, USA; Richard J Willke, PhD, Senior Director/Group Leader, Pfizer Inc, Worldwide Outcomes Research, Bridgewater, NJ, USA.
Workshop Purpose: Workshop Purpose: The 2006 FDA draft guidance on patient reported outcomes (PRO) measures states that one of the first steps in the drug approval process is to specify the conceptual framework for each PRO instrument to be used to support a statement about treatment benefit. This workshop will introduce two tools to aid in the development of the conceptual framework, namely, the Conceptual Taxonomy and the Endpoint Hierarchy. The workshop aims to help participants: 1) use the Conceptual Taxonomy to identify the structure of an instrument in terms of concepts and measurement approach; and 2) understand the relationship of this structure to an intended statement of treatment benefit.
Workshop Description: The Conceptual Taxonomy is a graphical presentation of concepts of health status and well-being within one or more families. Singular and low-level singular concepts are the most fundamental units in the taxonomy. A compound concept includes at least two singular concepts. Aggregate concepts include multiple families that can be explicitly measured, e.g., health-related quality of life. Although there can be numerous conceptual frameworks depending on the measurement goal of each instrument, endpoints fall into nine categories that can be arranged hierarchically from single items based on event counts to complex patient reports involving three or more families. Following a discussion of these two tools, participants will be presented with case studies based on existing product labels and asked to develop conceptual frameworks for the PROs in each label using the Conceptual Taxonomy and to use this information to place the instrument within the Endpoint Hierarchy to evaluate the instrument's ability to support an intended claim.

W14: FROM MINIMAL IMPORTANT DIFFERENCE TO CUMULATIVE RESPONSE CURVES: KEYS TO THE INTERPRETATION OF PRO RESULTS
DISCUSSION LEADERS: Benoit Arnould, MSc, MA, Director, Patient Scales for Clinical Practice, Mapi Values France, Patient Scales for Clinical Practice, Lyon, France; Muriel Viala, MSc, Senior Research Analyst, Mapi Values France, Lyon, France; Gilles Berdeaux, MD, Associate Director, Alcon France, Rueil-Malmaison, France; Aude Roborel de Climens, PhD, Research Associate, Mapi Values France, Lyon, France.
Workshop Purpose: To discuss how best to present PRO data for interpretation of clinical meaningfulness and to present the use of cumulative response curves as an alternative method to comparison with minimal important differences.
Workshop Description: The presentation and interpretation of PRO results is critical in convincing clinicians, health authorities and payers of their relevance in the assessment of treatment effects. As statistical significance is generally considered insufficient when interpreting PRO data from clinical trials, clinical significance of differences in PRO measures needs to be demonstrated and discussed. Since the seminal definition of Jaeschke (1989), a variety of terms and approaches have been used to define and establish a minimal important difference (MID). However, it appears that no single method is considered sufficiently robust to meet the needs of clinicians and health authorities for a definition of a clinically meaningful effect. Indeed, the FDA recommends the use of several methods to establish a range of MIDs rather than a single value. The diversity of published methods for the interpretation of the clinical significance of PRO data will be described according to the perspective taken (patient, clinician or society), the nature of the comparison made (within patient or between groups) and their applicability to longitudinal or cross-sectional assessments. Cumulative response curves (CRCs) were proposed as an alternative to MIDs in the conclusion of the presentation of the FDA guidance at the Mayo Clinic. CRCs allow for the complete pattern of response to treatment to be shown and for a statistical comparison to be made between treatments. The great potential of CRCs for the interpretation of PRO results will be presented and illustrated using two examples; one in respiratory medicine and one in ophthalmology. Participants will be asked to join in an interactive debate with their own experience in establishing MIDs.

W15: MIXED TREATMENT COMPARISON: A PRACTICAL APPROACH
DISCUSSION LEADERS: Sibilia Quilici, MSc, Senior Analyst, i3 Innovus, Health Outcomes, Uxbridge, Middlesex, UK; Keith Abrams, PhD, Professor of Medical Statistics, University of Leicester, Centre for Biostatistics & Genetic Epidemiology, Leicester, UK.
Workshop Purpose: To present the principles of mixed treatment comparison (MTC), its application and implication for medical decision makers.
Workshop Description: There is a growing interest in evidence synthesis and more specifically in indirect and mixed treatment comparison in the absence of head-to-head trials. While the indirect comparison becomes quickly limited in the presence of too many pair-wise comparisons, MTC provides a clear answer to the question ″what is the best available intervention for a specific indication?″. This workshop will start with the general principles of MTC and illustrate through several applications how to compare multiple treatments simultaneously and how to implement MTC depending on the type of data (binary, continuous and survival). We will present what can be achieved depending on which package is used and we will direct our attention more specifically on the advantages of WinBUGS and the different issues to consider. The presentation will provide information in relation to MTC and the size of the network of evidence as well as when to use MTC and how MTC should be reported. The workshop will conclude with the use of MTC and its implication for health authorities, using current NICE requirements as an example.

COST STUDY METHODOLOGY

W16: PROGRAMME BUDGETING AND MARGINAL ANALYSIS – ITS PLACE IN HEALTH ECONOMICS
DISCUSSION LEADERS: Monique Martin, MSc, MBA, Vice-President UK Operations, i3 Innovus, Uxbridge, Middlesex, UK; Alistair McGuire, PhD, Chair in Health Economics, London School of Economics, Health and Social Care, London, UK.
Workshop Purpose: To explore the uses and limitations of Programme Budgeting and Marginal Analysis in the evaluations of new technologies.
Workshop Description: Programme budgeting and marginal analysis (PBMA) has traditionally been used by the UK National Health Service (NHS) to evaluate the cost of specific programme changes by comparing past resource allocation in programmes with future allocations and appraising the benefits and costs of the extension of programmes or new programmes. This methodology, although of intrinsic benefit, is rarely used in other areas of health economics. The aim of Programme Budgeting, particularly when used with Marginal Analysis, is to outline current or past patterns of resource use as means of identifying the most effective and efficient means of investing resources in future services. These are disaggregated into appropriate programmes to simulate the aggregate budget outcomes and patient flows that would occur with growth in particular demands and shifts in both the demands and service responses. Marginal Analysis appraises the additional cost and benefits attributed to a proposed change or investment in service provision. Through application of a perspective that takes account of the whole treatment programme, rather than focusing on parts of care in isolation, PBMA helps overcome issues of silo budgeting and assessment of the aggregate impact of introducing new treatments. In this way PBMA is comprehensive and inclusive in the consideration of changes in investment strategy within a disease area. It is normally relatively quick and simple, but maintains sufficient robustness to be credible and informative. In this workshop we will show interactively, by encouraging attendee participation, how PBMA can be used to supplement other cost accounting tools, such as budget impact models, through identifying the impact of service provision on various service providers within the overall programme. In addition we will demonstrate its use to highlight the opportunity cost of a given programme, through providing information on aggregate resource use that is not traditionally provided by cost-effectiveness models.

HEALTH CARE POLICY AND FORMULARY DEVELOPMENT USING OUTCOMES RESEARCH

W17: THE COST-EFFECTIVENESS THRESHOLD: A HURDLE FOR INNOVATION?
DISCUSSION LEADERS: Mark JC Nuijten, MD, PhD, MBA, Consultant, Ars Accessus Medica/Erasmus University Rotterdam, Amsterdam, The Netherlands; Rogier M Klok, PharmD, PhD, Health Outcomes Officer, Wyeth Pharmaceuticals, Hoofdorp, The Netherlands; John Hutton, PhD, Professor of Health Economics, Centre for Health Economics, University of York, York, UK
Workshop Purpose: The aim of this workshop is to explore the potential negative impact of the use of a cost-effectiveness threshold on innovation, when taken into account a long-term societal perspective.
Workshop Description: The impact of cost-effectiveness data in the health care decision-making process has been increasing over the last 5 years leading already to formal requirements in a number of countries. In their decisions on paying or not for new drugs and technologies, decision makers apply - either explicitly or implicitly – a threshold for an acceptable incremental cost-effectiveness ratio, often the cost per QALY. Therefore a clinically innovative drug may not be reimbursed because of an unfavourable assessment of its cost-effectiveness. As a consequence cost-effectiveness outcome is a decision criterion that affects patients' lives and medical communities' capacity to deliver optimal health care. However, the lack of sales of the new drug may also have a negative impact on R&D investment by the pharmaceutical industry. As a consequence this reduction of innovation will reduce the development of future drugs, which may have a substantial superior clinical benefit and cost-effectiveness compared with the current treatments. In this workshop we illustrate that the use of cost-effectiveness outcomes for the reimbursement decision, is ignoring the potential future societal values resulting from market access of new drugs, which is not captured in the incremental cost-effectiveness ratio. The concept is illustrated for biologicals. Would biologicals have been available for all current indications, if the cost-effectiveness ratio would have been a formal requirement for reimbursement at time of launch for the first indication?

W18: TRANSLATING CLINICAL TRIAL DATA TO LOCAL ENVIRONMENTS: THE VALIDITY OF EXTRAPOLATING DATA FOR HTA SUBMISSIONS
DISCUSSION LEADERS: Phil McEwan, PhD, Director, Cardiff Research Consortium, Cardiff, South Glamorgan, UK; Phil James, BSc, Director, CHKS, Alcester, Warwickshire, UK; Peter Sharplin, MSocSc, Head of Pharmaceutical Development, Cardiff Research Consortium, Cardiff, South Glamorgan, UK; Mike Baldwin, MBA, Project Manager, Cardiff Research Consortium, Cardiff, South Glamorgan, UK.
Workshop Purpose: To explore the extent to which routine data can be used to characterise how generalisable clinical trial data is to local environments.
Workshop Description: Health economic evaluations for health technology assessments (HTA) require the synthesis of evidence from many sources. Cost effectiveness models use data from a variety of sources such as randomised clinical trials (RCTs) to describe efficacy and observational data to describe costs and resource use. It is increasingly the case that models developed outside the UK require modification before submission to either the SMC or NICE.
Multi-centre RCTs robustly describe the treatment effect of a new technologies, however translating these results to local environments can be difficult. The main challenge is the choice of a local comparator and the question of how this local comparator compares to the “control” arm of the study. Only when these uncertainties are understood can HTA bodies see how a new technology compares to current standard of care in their local area. An analysis of observational data in this setting could “bridge” this gap.
In the workshop, participants will be asked to discuss examples where technology assessments have questioned the validity of translating trial data from countries outside the UK. Results from a number of gold standard trials across a variety of disease areas and countries will be presented. The attendees will discuss how these new translational tools could reduce uncertainties and potentially improve HTA decision making.

PATIENT-REPORTED OUTCOMES STUDY METHODOLOGY (INCLUDING PREFERENCE-BASED & COMPLIANCE STUDIES)

W19: IS THERE A CONSENSUS ON BEST-PRACTICE METHODS FOR STATED-PREFERENCE STUDIES?
DISCUSSION LEADERS: Semra Ozdemir, MS, Research Economist, RTI International, Research Triangle Park, NC, USA; F. Reed Johnson, PhD, Senior Fellow and Principal Economist, RTI International, Research Triangle Park, NC, USA; A. Brett Hauber, PhD, Senior Economist and Head, RTI International, Research Triangle Park, NC, USA
Workshop Purpose: To compare various methods currently used in stated preference (SP) studies in health care, including differences in question formats, survey development and administration, data analysis, and interpretation of results. To evaluate the degree of current consensus on best practices based on a recent survey of active SP researchers.
Workshop Description: SP methods include both contingent-valuation (willingness-to-pay) and discrete-choice experiment (DCE) approaches. These methods have been used extensively and tested for validity in market research, and environmental and transport economics. Health economists are increasingly employing SP methods to elicit patients' and physicians' preferences. However, the quality of data collected and the validity of results are sensitive to a number of researcher decisions. Because health applications of these methods are relatively new, best-practice standards may still be evolving. This workshop will help participants appreciate the challenges of designing and conducting a high-quality SP study and help them to assess the state of the practice in preference research. Participants will complete actual surveys using contingent-valuation and discrete-choice experiment formats to elicit preferences for pain treatments. Presenters will analyze participants' data, compare results, and evaluate the reliability and validity of the two approaches. Participants will obtain an overview of current methods used in SP studies. Specific topics will include a) the conceptual basis for WTP measures, b) the relevance of willingness-to-pay estimates in health economics, c) applications of SP methods to quantify patient satisfaction and risk tolerance, d) the design, administration, analysis, and interpretation of SP studies, e) decisions researchers face when designing SP surveys, f) presentation of results from a recent SP "meta survey" on best-practice SP survey methods administered to active health and environmental preference researchers, g) discussion of research standards as reflected in peer-reviewed published studies, and h) a comparison of preference-research standards in health economics and environmental economics.

CLINICAL OUTCOMES STUDY METHODOLOGY

W20: ASSESSING THE EFFECTIVENESS OF COUNTER MATCHING AS A TOOL FOR IMPROVING THE EFFICIENCY OF THE NESTED CASE CONTROL DESIGN IN EFFECTIVENESS STUDIES
DISCUSSION LEADERS: Victor A Kiri, MSc, PhD, Director of Pharmacoepidemiology, PAREXEL International, Peri Approval Clinical Excellence (PACE), Uxbridge, London, Middlesex, UK; Camille Maringe, MSc, Epidemiology Fellow, GlaxoSmithKline R&D, Worldwide Epidemiology, London, Middlesex, UK.
Workshop Purpose: The participants of this workshop will gain an understanding of the counter matching methodology in nested case control design as we (1) describe its features (including the necessary adjustments to the resulting partial log likelihood function) and (2) use several simulated data to demonstrate its benefits over the conventional simple matching.
Workshop Description: When controls are drawn by simple random sampling of the risk sets in nested case control design, such can produce a very uneven split of exposed and unexposed subjects and lead to inefficiency. To optimise efficiency, statistical theory suggests counter matching cases and controls on exposure status based on the information in the full cohort. Participants will gain an understanding of the method as we (1) describe its features (including the adjustments required on the resulting partial log likelihood function) and (2) use several simulated data to demonstrate its benefits.
The nested design offers a simple method for avoiding unreasonable assumptions in the evaluation of time-dependent treatment effect. Its results are easy to interpret. Although its strength rests largely on the appropriateness of the controls matched to the cases, not much effort is generally made to evaluate its efficiency. Matching controls to cases on a confounding variable can improve the precision of the comparison of exposure groups although the effect of that variable cannot then be estimated. Whilst we do not match on the exposure variable, we still cannot estimate exposure effect, if in the matched data, all the cases are of the same exposure status and same will be true if all the controls are of the same exposure status or the status is the same on each matched pair. In other words, effect estimation is based entirely on the off-diagonal data of the resulting stratum-based 2x2 tables in the conditional logistic regression. Consequently in theory, the more such matched pairs our sampling can generate the more improvement we can expect on efficiency. Audience participation will be encouraged through an interactive discussion of at least 15 minutes with the presenters.

COST STUDY METHODOLOGY

W21: ANALYSIS OF SECOND ORDER UNCERTAINTY FOR PATIENT-LEVEL SIMULATION MODELS USING META-MODELS
DISCUSSION LEADERS: Jeremy VM Chancellor, MSc, Director, i3 Innovus, Health Economics and Outcomes, Uxbridge, Middlesex, UK; Patrick Mollon, MD, MBA, MSc, Senior Associate Director, Pfizer Ltd, Outcomes Research, Sandwich, Kent, UK; Samuel Aballéa, MSc, Lead Research Analyst, i3 Innovus, Health Economics and Outcomes, Uxbridge, Middlesex, UK.
Workshop Purpose: Conducting probabilistic sensitivity analysis (PSA) based on complex models, particularly patient-level simulation (PLS) models, may be impractical due to prohibitive running time. The aim of this workshop is to present approaches to circumvent this problem, and especially the use of meta-models (or emulators).
Workshop Description: We will conduct an audience poll of the respective priorities attached to conducting PSA and using PLS. Reasons generally mentioned for using PLS are that transition probabilities or pay-offs for some states depend on previous events or time. In some situations, those dependencies may be accounted for using a cohort-level simulation or a simplified model may provide an acceptable approximation. Where PLS is deemed necessary, questions to address before undertaking PSA are whether there is an easily predictable relationship between parameters and results, how many patients should be simulated per run and how many runs will be required. A possible approach to reduce the number of runs is to use a meta-model. The model is run a limited number of times, for different sets of parameters, and a statistical model predicting results of the original model as a function of parameters is then fitted, often using Gaussian processes. This meta-model can generate large numbers of approximated results, for parameters drawn at random, in minutes. The total running time is typically reduced from several months or more to hours. We will demonstrate this technique on a HIV model in which disease progression is related to CD4 cell count, viral load set point, treatment line, treatment response, age, and incidence of opportunistic infections. The model tracks the evolution of these variables to trigger antiretroviral switching on treatment failure and to modify progression probabilities accordingly. Software for implementing meta-models will be presented. The issue of whether the additional investment in analyst's time is justified will be discussed.

W22: THE IMPACT OF DIFFERENT RESOURCE COST DEFINITIONS ON RESULTS, INTERPRETATIONS, AND COMPARABILITY OF HEALTH ECONOMIC STUDIES
DISCUSSION LEADERS: Sandra L. Tunis, PhD, Senior Research Scientist, IMS Health Consulting, Health Economics and Outcomes Research, Noblesville, IN, USA; Elise M. Pelletier, MS, Director, IMS Health Consulting, Health Economics and Outcomes Research, Watertown, MA, USA; Meaghan St. Charles, MS, Research Scientist, IMS Health Consulting, Health Economics and Outcomes Research, Noblesville, IN, USA; Phil McEwan, PhD, Director, Cardiff Research Consortium, Cardiff, South Glamorgan, UK.
Workshop Purpose: To produce scientific discourse on the impact of different cost definitions on results, interpretations, and comparability of cost-effectiveness studies, and to provide an illustration through modeling the cost-effectiveness of a hypothetical pharmacological intervention for type 2 diabetes in the US and in Europe.
Workshop Description: Health economic studies vary greatly in design, methodological assumptions, and analytic approaches. A particularly important but often overlooked source of variation stems from the lack of a uniform proxy for defining direct costs of treatment. The extent to which increased expenditures for a medication or other treatment can be offset by reductions in subsequent service costs can be directly related to the choice of cost definitions. For example, one major cost-driver for type 2 diabetes is hospitalization for an acute myocardial infarction. Unit costs based on US private payer (IMS Pharmetrics data) mean charged, allowed, or reimbursed amount would be “costed” at $71,277, $22,448, or $15,183, respectively. In the US, reimbursements for services are commonly used to reflect a single payer perspective. When coupled with intervention (e.g., medication) costs that are not analogous (e.g., average wholesale price) however, the ratio of medication to other expenditures can be biased. Moreover, costs extracted from the literature can represent a combination of definitions, requiring interpretation of biases and of the generalizability of results to various US and European payer settings. Audience members will be invited to share opinions on the strengths and limitations of different cost definitions. An illustration of the impact of cost definitions will be provided with a simulation to model the cost-effectiveness, in the US and Europe, of a hypothetical medication for type 2 diabetes. Finally, discussion leaders will summarize how explicitly addressing the cost definition issue can help provide meaningful cost-effectiveness data to payers for policy development and management of health care expenditures.

HEALTH CARE POLICY AND FORMULARY DEVELOPMENT USING OUTCOMES RESEARCH

W23: OPPORTUNITIES AND BARRIERS FOR ASSESSMENT OF EMERGING HEALTH TECHNOLOGIES: THE CASE OF PHARMACOGENETICS
DISCUSSION LEADERS: Katherine Payne, PhD, Health Economist, Manchester University, Academic Unit of Medical Genetics, Manchester, UK; Hans-Peter Dauben, MD, Head, International Affairs, German Institute for Medical Documentation and Information, Cologne, Germany; Emma Gutierrez de Mesa, PhD, Scientific Officer, European Commission, DG-JRC, IPTS, Seville, Spain.
Workshop Purpose: To explore and discuss the opportunities and barriers affecting the development and application of an appropriate clinical and economic evidence-base for emerging health technologies such as pharmacogenetic technologies
Workshop Description: For new health technologies, such as medicines or diagnostics, the point of entry into national health care systems is regulated by licensing bodies using evidence of safety, quality and efficacy. In some countries, and for selected cases, they are also assessed for evidence of effectiveness, clinical utility and cost-effectiveness by an authority, like the National Institute for Health and Clinical Excellence (NICE) in the UK (UK) or Haute Authorité de Santé (HAS) in France, before being approved and recommended for its use in the health care system. Pharmacogenetic technologies offer the potential patient benefits of more effective and timely treatment but may have a significant cost impact on healthcare resources. Pharmacogenetic and genetic tests are sometimes required to meet external quality assessment standards (for example in UK through the National External Quality Assessment Service) but in many countries such tests do not go through the same licensing process as medicines and may emerge into clinical practice without the wanted level of evidence on clinical utility or cost-effectiveness. To date there is no robust evidence to support the effectiveness, in terms of clear patient benefits, or cost-effectiveness of pharmacogenetic tests. We suggest that an iterative approach to providing clinical utility and economic data, moving from retrospective modelling to prospective studies is necessary in order to provide timely information for decision-makers. Decision-makers can use robust evidence to inform when and how best to move pharmacogenetic tests from laboratories to clinical practice and develop appropriate therapeutic guidelines. Conclusions from a previous European Commission's workshop on this issue that used panel of experts will be used to raise key issues for discussion with the participants. In the workshop, participants, based on their empirical studies and experiences, will state the main opportunities, barriers and challenges facing the scientific based evaluation of such technologies.

PATIENT-REPORTED OUTCOMES STUDY METHODOLOGY (INCLUDING PREFERENCE-BASED & COMPLIANCE STUDIES)

W24: INTRODUCTION TO PROBABILISTIC CHOICE MODELS FOR THE MEASUREMENT OF HEALTH STATES
DISCUSSION LEADERS: Paul Krabbe, PhD, Senior Researcher, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
Workshop Purpose: Introduction of the basic conceptual ideas behind choice-based measurement models and their relationships to measurement theory and to the measurement of health states (utilities).
Workshop Description: This workshop presents an introduction to a class of measurement models, namely scaling models or probabilistic choice models (PCM). PCMs are developed to establish the relative merit of subjective phenomena and are based on specific judgmental tasks (e.g., ranking, paired comparisons) combined with associated data analysis. These models are very attractive to measure a subjective phenomenon like health-related quality of life or health states. PCMs include Thurstone scaling, the Rasch model, conjoint analysis, and the modern work of discrete choice modeling. The attractiveness of these models in quantifying health-related quality of life (i.e., utilities or values) is based on uncomplicated and cognitively simple judgment tasks that guarantee response data of good quality. So far, PCMs are scarcely applied in the field of medicine, with exception of the Rasch model. In the area of health economics discrete choice models have been recently used mainly to evaluate treatment preferences, although recent attempts have been made for the measurement of health states. Outline of the workshop: I) Introduction of the conceptual ideas behind the most significant PCMs and their relationships to measurement theory and to the measurement of health outcomes and health-related quality of life. II) Demonstration of the Thurstone model and discrete choice models. III) Hands on experience with the most basic of all PCMs; Thurstone scaling. IV) Results from recent empirical studies will be presented.