ISPOR 15th Annual European Congress: Workshop Presentations
 
WORKSHOP PRESENTATIONS
 
WORKSHOPS - SESSION I
MONDAY, 5 NOVEMBER 2012: 16:45-17:45

Clinical Outcomes Research

W1: USING DECOMPOSITION METHODS TO ESTIMATE HETEROGENEITY OF TREATMENT EFFECTS IN RANDOMIZED TRIALS AND OBSERVATIONAL STUDIES
Discussion Leaders:

William Crown, PhD, Group President, HEOR and Late Phase Research, OptumInsight, Waltham, MA, USA; William H Olson, PhD, Leader, Research and Analysis Strategy, Janssen Pharmaceuticals, Inc., Titusville, NJ, USA

Purpose:

Analyses of RCTs focus upon the estimation of average treatment effects.  Heterogeneity of treatment response is generally conducted with ad hoc analyses of patient subgroups.  Using statistical decomposition methods, we will demonstrate that average treatment effects are a weighted average of the heterogeneity of subgroup mean treatment effects within and across the arms of the trial.  This framework can be used to examine the conditions under which randomized and observational studies can be expected to generate similar results.

Description: The workshop will begin with a presentation of decomposition methods originally developed in the labor economics literature to estimate gender and racial discrimination in wages.  The methods decompose total variation between groups into variation that is due to differences in sample characteristics and variation that is due to the relationships of each group’s characteristics to their outcomes.  Randomization is intended to balance the characteristics of comparison cohorts on both observed and unobserved variables.  Thus, application of decomposition methods to RCT data leads to the conclusion that average treatment effects are a weighted average of structural coefficients within and across each arm of the trial.  This will be illustrated with an empirical case study.  Comparison of RCT and observational studies within the decomposition analysis framework is also helpful for identifying the conditions under which estimates of treatment effect would be expected to be similar for the two types of designs. Challenges, such as unobserved confounders and sample size limitations will be discussed.  Finally, we will place these methods in the context of the traditional biostatistical literature  as well as accompanying developments in the economics literature .  Workshop attendees will be encouraged to discuss the pros and cons of using decomposition methods in their own research.
W2: META-ANALYSIS OF DIAGNOSTIC TEST ACCURACY AND EFFECTIVENESS DATA: ARE THEY REALLY THE SAME?
Discussion Leaders:

Sandrine Cure, MSc, Associate Director, OptumInsight, Uxbridge, Middlesex, UK; Keith Abrams, BSc, MSc, PhD, Professor of Medical Statistics, Department of Health Sciences, University of Leicester, Leicester, UK; Helene Cawston, MSc, Senior Research Analyst, OptumInsight, Nanterre, France

Purpose:

The purpose of the workshop is to critically discuss the challenges and methodological differences of carrying out meta-analyses of diagnostic test accuracy data compared to those on effectiveness data and assess the feasibility of performing indirect comparisons on these data.

Description: In recent years, new diagnostic and screening tests have been developed due to technological advances. In parallel, interest in evidence-based diagnosis has also been increasing as policy makers and clinicians consider the assessment of the reliability and performance of these new tests. Performing meta-analyses on diagnostic test data however is more difficult than to do them on effectiveness data. This workshop will endeavour to stress the methodological differences between both types of meta-analyses. We will describe the statistical challenges encountered in the diagnostic test meta-analyses because of the need to deal simultaneously with multiple measures of the test accuracy (i.e. sensitivity and specificity) instead of a single efficacy or safety outcome (e.g. treatment success) and to incorporate the threshold effects in the estimation. We will also present the several meta-analytic models used so far in the literature. They range from very simple models that synthesize independently both the sensitivity and specificity to more complex ones including a hierarchical summary receiver operating characteristic (SROC) curve. Using examples derived from the literature, we will outline the respective strengths and limitations of the different methods and assess the feasibility and challenges of performing indirect comparisons in this area. Workshop participants will be invited to share their thoughts and experiences.

Economic Outcomes Research

W3: FUNDAMENTALS OF MODEL CALIBRATION: THEORY & PRACTICE
Discussion Leaders:

Douglas Taylor, MBA, Associate Director, Health Economics & Outcomes Research, Ironwood Pharmaceuticals, Inc., Cambridge, MA, USA; Ankur Pandya, PhD, Assistant Professor, Department of Public Health, Weill Cornell Medical College, New York, NY, USA; David Thompson, PhD, Senior Vice President & Head of Emerging Business, Quintiles Outcome, Cambridge, MA, USA

Purpose:

Calibration of model inputs has been recommended by the ISPOR Task Force on Good Modeling Practices as an important step in the process of model validation.  However, awareness of calibration techniques is not widespread.  The purpose of this workshop is to provide an overview of theoretical and practical issues in model calibration.

Description:

Pharmacoeconomic models may require calibration when data directly informing model input parameters are lacking, or when model outputs do not accurately reflect known external benchmarks.  A variety of calibration methods exist, but the general process involves running the model, comparing the outputs to the benchmarks, adjusting the inputs, re-running the model, evaluting whether the outputs are converging on the benchmarks, and continuing this process until a predetermined stopping criterion is reached.  In this workshop the theoretical and practical advantages and disadvantages of different aspects of model calibration will be discussed, including:  (1) how to identify the calibration inputs and outputs; (2) how to specify the objective function and determine goodness of fit (using windows-based targets, Euclidean distance, mean error, and maximum likelihood criteria); (3) selecting and implementing parameter search processes, including manual, random, and algorithmic (e.g., Nelder-Mead “Downhill Simplex”, Latin Hypercube, and Simulated Annealing) approaches (4) incorporating calibration into sensitivity analyses and (5) adjusting for temporal changes in treatment patterns.  A hypothetical and an actual research model will be used to illustrate the effects of calibration choices on model results and sensitivity analyses.  The workshop should be of interest to applied researchers involved in the construction and estimation of pharmacoeconomic models and managers responsible for the critical appraisal of such models.  Workshop participants will be encouraged to provide discuss their own experiences with model calibration.

Health Care Policy Development Using Outcomes Research

W4: DRUG DEVELOPMENT TOOL (DDT) QUALIFICATION BY THE EMA AND FDA: PURPOSE, PROCEDURES, CHALLENGES, AND OPPORTUNITIES
Discussion Leaders:

Margaret Vernon, PhD, Research Scientist, United BioSource Corporation, London, UK; Nancy Kline Leidy, PhD, Senior Vice President of Scientific Affairs, Outcomes Research, United BioSource Corporation, Bethesda, MD, USA; Brigitta U. Monz, MD, PhD, Head Global Health Economics & Outcomes Research, Corporate MAPOR, Boehringer Ingelheim GmbH, Ingelheim, Germany

Purpose:

The objective of this workshop is to provide greater clarity around drug development tool (DDT) qualification offered by the EMA and FDA by discussing the purposes of qualification, procedures used by the two agencies, and the challenges and opportunities this process presents to the drug development process.

Description:

Research instruments intended for use in drug development may be submitted to the EMA and/or FDA for formal regulatory qualification review. Candidate instruments include clinical outcome assessments [patient-reported (PRO), clinician-reported (ClinRo) and observer-reported (ObsRO) outcome measures] and biomarkers, intended for use in sample selection, stratification, or as an efficacy endpoint.  Although qualification purposes and procedures were described by the EMA in 20091 and the FDA in 20102, experiences with the procedure have been limited.  As of April 2012, five PRO measures have undergone the process of qualification advice at the EMA, with two PRO measures under formal qualification review by the FDA.  No PRO measures have completed the entire process at either institution.  Given the lack of precedent, the qualification process is seen as both intriguing and daunting.  Workshop leaders represent a range of experiences and perspectives of the DDT process and will engage the audience in a discussion of (1) Determining when DDT qualification is and is not appropriate; (2) The procedures involved in DDT qualification submissions; (3) The regulatory review process for DDTs submitted for qualification; and (4) Qualification feasibility, utility, opportunities, and alternatives.  Case examples will be used to highlight key points and offer practical insights into the process.  Throughout the workshop, the audience will be invited to raise questions and issues related to the challenges and opportunities of DDT qualification, offering attendees an opportunity to contribute to and obtain further clarification around this important process in drug development.

W5: UNPLANNED EVIDENCE: IMPLICATIONS OF INVESTIGATOR-INITIATED TRIALS (IIT)
Discussion Leaders:

Rachel Beckerman, PhD, Principal, Value Demonstration, CBPartners, New York City, NY, USA; Corinna Sorenson, MPH, MHSA, Research Fellow, European Health Policy, Health & Social Care, London School of Economics and European Health Technology Institute for Socio-Economic Research, London, UK; Mónica Martín de Bustamante, AB, BE, Managing Director, Europe, Middle East, and Africa, CBPartners, Basel, Switzerland

Purpose: The purpose of this interactive workshop is to explore how unplanned evidence from investigator-initiated trials (IITs) has shaped the outcome of HTAs for specialty therapies, in order to better understand how a product’s reimbursement can be optimised in such situations.
Description: Particularly in specialty arenas such as oncology and autoimmune diseases, investigator-led trials are commonplace, especially to explore clinical scenarios that may be considered too risky by the manufacturer (e.g. head to head studies, outcomes research). To this end, in order to better inform pharmaceutical decision-making , many government stakeholders in the EU (such as AIFA in Italy and the NHS in the UK) are providing funding for IITs to address such questions. Manufacturers vary in whether IITs are included alongside industry-sponsored research within the products’ dossier submissions to reimbursement authorities: although on the one hand, the provision of additional safety data may be attractive, IIT data tend to be collected from a less stringently chosen patient population, potentially affecting the desired efficacy outcomes.  Likewise, HTA bodies vary in whether they conduct an independent review to consider IIT-generated evidence that may have been omitted from the MFG dossier.  The implications of a manufacturer’s omission of IITs within its HTA submission typically manifest themselves via reimbursement authorities’ concerns over appropriate comparator selection or acceptable therapy positioning, potentially impacting the product’s pricing and reimbursement outcome. This workshop will explore three examples of dossier submissions that considered unplanned evidence from IITs, and the consequent impact on the product’s reimbursement success or failure in the major HTA markets.  Participants will be asked to contribute to an interactive discussion around how best to optimise HTA strategy in the face of unplanned IIT evidence.

Patient-Reported Outcomes & Patient Preference Research

W6: THE ART AND SCIENCE OF EXPERIMENTAL DESIGN - APPLYING THE RECOMMENDATIONS OF THE ISPOR CONJOINT ANALYSIS EXPERIMENTAL DESIGN TASK FORCE
Discussion Leaders:

F. Reed Johnson, PhD, Distinguished Fellow and Principal Economist, Health Preference Assessment, RTI Health Solutions, Research Triangle Park, NC, USA; Axel C. Mühlbacher, PhD, Professor for Health Economics and Health Care Management, IGM Institute, Hochschule Neubrandenburg, Neubrandenburg, Germany; Deborah Marshall, PhD, MHSA, Canada Research Chair, Health Services & Systems Research, Associate Professor, Department of Community Health Sciences, Faculty of Medicine and Director, HTA, Alberta Bone & Joint Health Institute, University of Calgary, Calgary, AB, Canada; John FP Bridges, PhD, AssociateProfessor, Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

Purpose: Discrete choice experiments (DCE) are established stated-preference methods for the assessment of the preferences of patients and other stakeholders. Increasingly, regulatory and health technology assessment agencies in Europe are examining DCE as a means to identify and value patient-centered outcomes and outcomes researchers in the public and private sector now routinely use DCE to either design or evaluate in health care interventions and services. The ISPOR Task Force on Good Research Practices for Conjoint Analysis has produced a checklist outlining good research practices for the application of DCE, the complexity and rapid advancements in the experimental design arena required a more detailed examination. The ISPOR Conjoint Analysis Experimental Design Task Force was established to identify and examine alternative strategies for experimental design and to provide guidance for good research practices on this critical aspect of DCE.
Description: This workshop will summarize the key aspects of the Experimental Design Task Force report, and discuss practical applications of its findings. We will present a conceptual framework for assessing the properties of designs and discuss the limitations of several experimental design approaches. We will demonstrate practical solutions to ensure that an unbiased and efficient estimation of the parameters of the resulting choice model. This workshop will provide a detailed and practical introduction to the art and science of experimental design. As such, it will provide a valuable educational experience for those who want to commission, conduct, review, or interpret applications of conjoint analysis and DCE in health. It is targeted to outcomes researchers, health economists, policy makers and other stakeholders that seek to learn more about the practicalities of experimental design. In addition to formal presentations, this interactive workshop will offer an opportunity for participants to comment on the Task Force report and to share their personal experiences with experimental design.
WORKSHOPS - SESSION II
TUESDAY, 6 NOVEMBER 2012: 15:00-16:00

Clinical Outcomes Research

W7: INNOVATIVE USES OF PHASE III DATA TO PLAN PHASE IV RESEARCH
Discussion Leaders:

David Thompson, PhD, Senior Vice President & Head of Emerging Business, Quintiles Outcome, Cambridge, MA, USA; Sandrine Cure, MSc, Associate Director, OptumInsight, Uxbridge, Middlesex, UK; Raquel Cabo, MSc, Global Health Economics Manager, Health Economics and Reimbursement, GE Healthcare, Chalfont St. Giles, Bucks, UK

Purpose:

Recent years have witnessed heightened focus on Phase IV research, including pragmatic clinical trials and other real-world data collection activities.  Manufacturers often use their Phase III clinical trial data for internal planning of Phase IV studies, but such use is generally informal and not at the level of rigor inherent in scientific research for external dissemination.  The purpose of this workshop is to highlight opportunities and challenges for more rigorous use of Phase III data to plan Phase IV research.

Description: Guidelines for the conduct of relative effectiveness assessments in Europe and comparative effectiveness research in the US emphasize the importance of head-to-head data to evaluate treatment alternatives.  However, because registration trials in Phase III remain focused on demonstrating safety and efficacy (often versus placebo), collection of head-to-head treatment effectiveness data typically does not occur until Phase IV.  This workshop will discuss how Phase III research can be leveraged to improve the information base on which Phase IV studies are designed and executed.  Attention will be focused on two specific methodologies:  (1) indirect/multiple treatment comparison (ITC/MTC) and (2) stochastic trial simulation modeling.  Evidence synthesis via ITC/MTC permits head-to-head treatment comparisons by linking data from trials involving a common comparator, such as placebo.  Stochastic trial simulation modeling provides an analytic apparatus to assist in the design of a Phase IV trial and addresses a variety of questions of interest, including expected differences in study endpoints between treatment groups, the likelihood that the trial outcome will favor a given intervention, and statistical power for various sample sizes.  Alternate-scenario (“what-if?”) analyses can be performed to assess the potential impact of modifications to the trial design.  Workshop participants will be engaged in discussion of the opportunities and challenges that arise when these techniques are applied to Phase III data to assist in the planning of Phase IV research.
W8: IT'S TIME TO REASSESS TRADITIONAL TIME-DEPENDENT REGRESSION METHODS
Discussion Leaders:

Christopher M. Blanchette, PhD, Research Associate Professor, Public Health Sciences, University of North Carolina, Charlotte, NC, USA and Principal, Health Economics & Outcomes Research, IMS Health, Alexandria, VA, USA; Alex Exuzides, PhD, Director, ICON Late Phase & Outcomes Research, San Francisco, CA, USA; Roger Luo, PhD, Director, Advanced Analytics, IMS Health, Plymouth Meeting, PA, USA

Purpose:

Time-dependent regression models are a mainstay in the analysis of patient level data in the search for exposure – outcome associations while allowing for the adjustment of both baseline as well as time-dependent covariates. However, these traditional models which are usually expressed as either a Cox-proportional hazards or a Weibull model, have been challenged recently. One of the new approaches is the marginal structural model (MSM) originally proposed by James Robins. The MSM is a causal model for the estimation of causal effect of time-dependent covariates that may be simultaneously confounders and intermediate variables. It offers unbiased and consistent estimates compared with other models. This workshop will discuss how innovative approaches seek to improve traditional time-dependent regression methods through case studies and audience participation.

Description: Two topics will be presented and simulated through both theoretical and technical detail.  Topic 1 will describe a case study comparing traditional Cox-proportional hazards regression models with marginal structural models to adjust for time-dependent fluctuations in risk mediated by treatment exposure.  Topic 2 will discuss an innovative approach in predicting the chance of mortality in the presence of time-dependent biomarkers; a method that combines the estimation of a Cox-proportional hazards model, in the presence of time-dependent covariates, while simultaneously fitting a parametric accelerated failure time model. Both case studies will illustrate the methodological approach through interactive SAS tutorials. In this workshop we will discuss each method by exploring the benefits and limitations compared to more traditional time-dependent approaches while also providing a tutorial using statistical software.  Attendees will leave with a better understanding of how these methods are applied and how they can be used to overcome various limitations.

Economic Outcomes Research

W9: EARLY MODELLING IN MEDICAL PRODUCT DEVELOPMENT AND MARKET ACCESS
Discussion Leaders:

Maarten J. IJzerman, PhD, Professor & Chair, Department of Health Technology & Services Research, University of Twente, Enschede, The Netherlands; Mark J. Sculpher, MSc, PhD, Professor of Health Economics, Centre for Health Economics, University of York, Heslington, York, UK; Pierre Sagnier, MD, DEA, MPH, Head, Global Market Access, HEOR, General Medicine, Bayer HealthCare, Berlin, Germany

Purpose: Early modelling aims to objectively inform industry regarding technology development decisions as well as the health system about upcoming technologies. The purpose of this workshop is to introduce the different perspectives in early modelling, the specific requirements and challenges and finally the relevance and use of early models.
Description: Medical technology innovation is critical to improvements in patient care, but costly and uncertain. Early modelling is increasingly promoted as a way to increase the likelihood of successful product development and, hence, rapid patient access to valuable new technologies. Stage gate models qualitatively describe the product development path including key decisions to be made. In at least two distinct stages, early modelling may be used, i.e. (1) early modelling to inform about the future health economic benefit of medical products to inform health policy and (2) early modelling to support decisions to further invest or disinvest in R&D and to manage product portfolios. Where health economic models do have many advantages, there are a number of issues to consider. For instance, how should we deal with developmental uncertainty and uncertain regulatory requirements. This workshop will cover four topics: (1) How early modelling can help with technology development decisions (i.e. commercial decisions by manufacturers); (2) How early modelling can help health systems (i.e. system perspective); (3) Methodological challenges (and how to resolve them) with this type of modelling; (4) Why models are not widely used, how we can increase their use. The workshop will also include a 15 minute audience participation on experiences and future directions. In particular, the audience is requested to join the discussion on good research practices for early modeling.

Health Care Policy Develpment Using Outcomes Research

W10: THE IMPLEMENTATION OF DISINVESTMENT INITIATIVES: A EUROPEAN PERSPECTIVE ON PROGRESS TO DATE
Discussion Leaders:

Christian A Gericke, MD, PhD, MPH, MSc, Professor and Deputy Director, PenCLAHRC, Plymouth, Devon, UK; Sarah Garner, PhD, BPharm, Associate Director, Research and Development, National Institute of Health and Clinical Excellence (NICE), London, UK; Francois Meyer, MD, Advisor to the President and Director for International Affairs, Haute Autorité de Santé (HAS), Saint-Denis, France

Purpose:

As countries across Europe grapple with acute budgetary pressures, a general consensus has been made that health budgets should not be used to fund ineffective or low-value services. Yet the challenge of achieving this extends beyond the appropriate assessment and allocation of new resources. In the current economic climate, it is increasingly important that the issue of active disinvestment of health technologies is implemented.

Description: Although the notion of disinvestment (withdrawing health services from an existing healthcare service that is considered to deliver little or no health gain for its cost) has an obvious logic, the practical implementation of disinvestment decisions has proved to be both controversial and problematic. While most countries have recognised the need to prioritise services and remove those which are either ineffective or of low-clinical value, very few formal mechanisms to guide this process currently exist. Consequently, the uptake and diffusion of disinvestment decisions have a propensity to be influenced by a range of social, financial, professional and institutional factors rather than clinical evidence. Arguably, this fails to produce optimum levels of health outcomes or efficient use of scarce resources. Accordingly, international experts from both health service and research settings who have led disinvestment initiatives will discuss practical implementation strategies and share their experiences to date. Discussion leaders will consider questions such as: (1) What is the best way to identify candidates for disinvestment?, (2) Is there a lack of mechanisms to identify and prioritise technologies and/or practices for disinvestment?, (3) What are the political, clinical and social challenges faced when removing an established technology?, (4) Is there a lack of dedicated resources to build and support disinvestment policy mechanisms?, (5) How are procedures/drugs with multiple indications handled?, (6) How rigorous are the assessments of clinical and economic value of identified procedures for disinvestment?
W11: ADHERENCE AS A DIFFERENTIATING FACTOR IN THE APPRAISAL OF NEW TREATMENTS: SHOULD ECONOMIC EVALUATION CONSIDER WHETHER PATIENTS TAKE THEIR TREATMENT APPROPRIATELY?
Discussion Leaders:

Antoine Regnault, PhD, Research Director, MAPI Consultancy, Lyon, France; Gert Bergman, PhD, Associate Director, MAPI Consultancy, Houten, The Netherlands; Chloe Brown, PHD, Director, RJW & Partners, Royston, Hertfordshire, UK

Purpose: To explore opportunities to demonstrate the economic value of enhanced medication-taking behaviors in the appraisal of new treatments
Description:

Medication-taking behaviors (adherence, compliance, persistence, concordance) have been the object of an increasing body of research in recent years, drawing on a fairly solid theoretical background. The importance of integrating these behavioral aspects into economic evaluations has been recognized however, to date, they are not systematically nor consistently considered in economic models. The position of HTA agencies on such issues is often unclear. Furthermore, comparative data concerning adherence or data linking adherence to clinical outcomes are in many instances not available or of poor quality. This makes their integration into economic models hazardous. This workshop will outline the theoretical background on medication-taking behaviors, with a particular focus on the relevance of the different existing concepts, theories and measurement approaches from the economic evaluation perspective.  It will demonstrate how economic models may incorporate parameters relating to medication-taking behavior, thus allowing for the rigorous integration of such behavioral aspects into the appraisal of new treatments. It will present the critical challenges for designing and running such models, focusing particularly on the type of data required. Finally, the workshop will examine adherence-related questions in evaluations already completed by HTA agencies. To this end, examples from reviews by NICE, SMC, and HAS will be presented and discussed. This will be an interactional workshop and participants’ contributions will be encouraged throughout. This session is directed at individuals who interact with HTA agencies or are interested in economic modeling.

Patient-Reported Outcomes & Patient Preference Research

W12: PAYER AND HTA PERSPECTIVES ON CLINICAL OUTCOME ASSESSMENTS (COAS)
Discussion Leaders:

Erin Tomaszewski, MPH, Clinical Outcomes Research Scientist, Quintiles Outcome, Pittsburgh, PA, USA; Peter Black, MS, Senior Scientist, invivodata Consulting, Pittsburgh, PA, USA; Marta Andreykiv, PharmD, MSc, Senior Consultant, Consulting, Quintiles, Hoofddorp, The Netherlands; Stefan Holmstrom, MSc, Director, HEOR, Astellas Pharma Global Development, Leiderdorp, The Netherlands

Purpose: The purpose of this workshop is to evaluate payer perspectives of use of COAs in comparison to clinical application of COAs. Examples of application of COAs and methodological considerations for use of COAs (including PROs, ClinROs, and ObsROs) will be discussed in the context of current regulatory trends.
Description: Applying more methodological rigor to COAs has been driven primarily by communication from regulatory agencies (e.g. EMA HRQOL Reflection Paper, 2006; and FDA PRO Guidance, 2009). While this trend has primarily focused on increasing precision of COA instruments to assess clinical effectiveness, there is increasing awareness to develop and apply COAs to support reimbursement and payer strategies.  Some of the European HTA agencies are explicit in their guidance about preference for certain COAs necessary to support the reimbursement of a new product.  In other cases, the historical decisions have to be reviewed to understand payer perceptions and preferences.  Discrepancies exist between which COAs are preferred and recommended by the payers and which COAs are applied in clinical trials.  Insight into these differences will be discussed and examples given for the payer perspective, application in clinical trials, and COA regulatory considerations. This workshop will present evidence of what payers are recommending for COAs, as well as payer reactions and potential criticism of selection and application of COAs in clinical trials.  Examples of application of COAs in clinical trials will be presented. These COAs will be further evaluated in light of reimbursement decisions for treatments, and any existing trends in relationship to payer perspective and COA methodology will be discussed.  The workshop will conclude with an audience interactive session in which a basic framework will be presented for the development of a COA endpoint strategy to inform reimbursement and payer considerations and strategies in drug development programs.
WORKSHOPS - SESSION III
TUESDAY, 6 NOVEMBER 2012: 16:15-17:15

Clinical Outcomes Research

W13: RETHINKING ANALYSIS OF OUTCOME MEASURES FOR DEMONSTRATING VALUE OF PHARMACEUTICAL PRODUCTS: IMPLICATIONS FOR STUDY DESIGN, PERSONALIZED MEDICINE, AND COMPARATIVE EFFECTIVENESS RESEARCH
Discussion Leaders:

Donald E. Stull, PhD, Director Retrospective Data Analysis, RTI Health Solutions, Didsbury, Manchester, UK; Katherine Houghton, BSc, Research Health Outcomes Scientist, RTI Health Solutions, Didsbury, Manchester, UK; Jennifer Petrillo, PhD, Associate Director, Global Health Economics & Outcomes Research, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA

Purpose: This workshop will introduce the concept of heterogeneity in responses in clinical trials and observational studies; describe why it is important from the perspective of patients, providers, manufacturers, and payers; present novel methods to identify groups of differential responders to treatment; compare results from standard methods with those of innovative methods; and present examples from recent research comparing these methods.  This workshop is designed for those familiar with analyses of clinical trial and observational data, but want to learn more about innovative classification methodology.
Description: Heterogeneity in treatment response – i.e. patients responding differently to the same to treatment –is often encountered in research.  Some of this is anticipated through pre-defined subgroups, such as treatment versus placebo; some is a consequence of factors not known beforehand.  In many instances this heterogeneity can minimise differences between treatment arms.  Identifying these patients efficiently can play a key role in study design and comparative effectiveness research.  Analytic methods are now available to leverage patient-reported outcomes data to identify subsets of individuals who show different responses on outcome measures but who are not necessarily pre-defined.  Once identified, the characteristics of these individuals can be examined to see if they shed light on their differential response to treatment.  These methods are a good choice when not all individuals will show the same response to interventions and analysts would like to identify differential responders efficiently.  Comparisons with standard analytic methods will demonstrate the added insight these newer methods can provide.  The potential value of these analyses for drug development and patient care will be discussed along with results from recent studies.  Following the presentation, the panel will engage the audience in a discussion of the benefits of these methods for current drug development programs and comparative effectiveness research.

Economic Outcomes Research

W14: MODELING OF INFECTIOUS DISEASES FOR PREVENTION AND MITIGATION: HOW MODELING HAS INFORMED PUBLIC HEALTH POLICY
Discussion Leaders:

Maarten J. Postma, PhD, Professor, Department of Pharmacy, University of Groningen, Groningen, The Netherlands; Kamal Desai, PhD, Research Scientist, Health Economic Modeling and Simulation, United BioSource Corporation, London, UK; Deirdre Hollingsworth, PhD, Research Fellow, MRC Centre for Outbreak Analysis and Modelling, Imperial College London, London, UK; Ruth Chapman, PhD, Research Assistant, Health Economic Modelling and Simulation, United BioSource Corporation, London, UK

Purpose: In order to inform policy, public health agencies increasingly depend on epidemiological transmission models to make predictions on the impact of emerging diseases as well as the effectiveness of prevention or control measures. For instance, agencies face questions related to mitigation of disease outbreaks like flu or SARS, immunization policy for novel or newer generation vaccines that replace older ones, and approaches for reduction of person to person infection as for HIV. Models developed to address these questions are unlike those encountered in non-infectious diseases since they must incorporate the particular nature of infection and prevention to capture herd immunity, vaccine-induced serotype replacement, shifting age of infection or other characteristics. The aim of this workshop is to discuss recent topics faced by public health agencies, illustrate the range of model types which have provided answers, and discuss how they have supported policy decisions.
Description:

The workshop will begin with a background on the differences between the two traditions of infectious disease and health economics in terms of disease concepts, public health questions, and distinctive modeling techniques. Recent examples of case-studies drawn from the presenters’ experiences working with the Health Protection Agency (UK) and other agencies will be presented. Examples will be : Real-time analysis of disease outbreaks for rapid assessment of basic reproduction rate and case-fatality (SARS and H1N1 influenza), analysis of immunization strategies for new vaccines (Pertussis and HPV), and infection control approaches from recent clinical trials (HIV). In each example, participants will be provided with basic information regarding the disease or prevention approach and invited to identify the critical public health questions and solutions through interactive discussions and demonstrations. We will then discuss how modeling contributed to informed policy decisions.  This session is aimed at those interested in infectious diseases or modeling; however no modeling experience is required.

W15: WHAT IS VERSUS WHAT COULD BE: INCORPORATING OPERATIONS RESEARCH METHODS IN THE HEOR TOOLKIT
Discussion Leaders:

William Crown, PhD, Group President, HEOR and Late Phase Research, OptumInsight, Waltham, MA, USA; Deborah Marshall, PhD, MHSA, Canada Research Chair, Health Services & Systems Research, Associate Professor, Department of Community Health Sciences, Faculty of Medicine and Director, HTA, Alberta Bone & Joint Health Institute, University of Calgary, Calgary, AB, Canada

Purpose:

Virtually all health economics and outcomes research studies utilize methods from health economic modeling, biostatistics, epidemiology, health services research, or econometrics.  The conceptual framework of optimization is rarely used, even though the ultimate goal of .much research is to inform policies that will encourage the optimal allocation of scarce resources to achieve a given objective (e.g., maximize public health).  To this end, the workshop will consider the potential advantages and disadvantages of using operations research methods in HEOR research.

Description:

Optimization is rarely used as the conceptual framework for HEOR studies.  For example, although it is common to estimate health economic models of the treatment of disease there are virtually no examples of dynamic programming analogues where patient treatment pathways are optimized subject to a series of constraints such as proximity to treatment centers.  Applications of operations research methods in health care are mainly to be found in managing health care operations--scheduling, transportation, and queuing problems—rather than outcomes research.  This workshop will begin with a very brief introduction to several operations research methodologies such as linear and dynamic programming.  We will then illustrate how a health economics model can be recast as a dynamic programming model and will discuss the interpretation of differences in the results stemming from the two approaches.  We will discuss the advantages and disadvantages of incorporating optimization methods in broader systems dynamics models designed to guide policy development—particularly, the issue of describing the current state of a health system versus its optimal state.. Workshop participants will be encouraged to discuss how operations research methods could be incorporated into their own research.

Health Care Policy Development Using Outcomes Research

W16: DON'T FORGET ABOUT THE PATIENT! LISTENING TO PATIENTS AND INVOLVING THEM IN RESEARCH
Discussion Leaders:

Elisa Cascade, MBA, Vice President, MediGuard/Digital Patient Unit, Quintiles, Rockville, MD, USA; Derek C Stewart, BA, Associate Director for Patient & Public Involvement, Clinical Research Networks, National Institute for Health Research, Leeds, UK; Dean Summerfield, MA, DPhil, Vice President, Europe, Quintiles Consulting, Reading, UK

Purpose: The purpose of this workshop is to hear from patients about research, better understand how patient feedback can shape research designs, and provide case examples of how patients can influence study design and participate in research.
Description: In late 2011, the UK presented vision to become the global hub for life sciences that brings together researchers, clinicians, and patients to translate research into clinical use for medical innovation.  Around the same time, the US Congress created the Patient-Centered Outcomes Research Institute (PCORI) to “conduct research to provide information about the best available evidence to help patients and their health care providers make more informed decisions” and the FDA recently announced the plan to include feedback from patients with 20 different conditions under PDUFA V.  Taken together, these initiatives underscore a growing recognition of the importance of involving patients in research, but questions remain on how to collect and act upon patient feedback.  Outside of the traditional patient association route, the growth in digitally enabled patient pools combined with advances in medical internet research provide a unique opportunity for capturing patient opinion on research design and endpoints; involving patients in research execution; and monitoring disease trends prospectively in observational studies including not only patient-reported information, but also medical record and laboratory/genetic data.  In this session, we will present feedback from the patient perspective both at the individual level and in aggregate.  In addition, we will summarize learnings related to how patients perceive value, risk, and study design.  Finally, we will present case studies on approaches to communicating with patients regarding research design and participation. The session will conclude with a facilitated discussion of the role of the patient in research and how researchers can involve patients in research design and execution.
W17: ADDITIONAL PATIENT RELATED BENEFITS ARE THE KEY TO PRICE NEGOTIATION IN GERMANY – PRACTICAL EXPERIENCE WITH BENEFIT DOSSIERS AND THE ASSESSMENT PROCESS
Discussion Leaders:

Olaf Pirk, MD, PhD, Director, Olaf Pirk Consult, Nuremberg, Germany; Meriem Hind Bouslouk, PhD, Official Adviser, Drug Department, Federal Joint Committee (G-BA), Berlin, Germany; Frank-Ulrich Fricke, PhD, Professor, Deputy Member of the Arbitration Panel for Drugs, Georg-Simon-Ohm University of Applied Sciences, Nuremberg, Germany

Purpose: The workshop’s objective is guiding through the German assessment process for drugs with new chemical entities (NCE) by the respective representatives involved in the process. The whole process starts with an advisory meeting at the Federal Joint Committee (G-BA) on the appropriate comparator (AC), followed by the production of the benefit dossier and its assessment by the G-BA. The price negotiation process with the National Association of Statutory Health Insurance Funds (GKV-SV) and the arbitration process will be elucidated as well.
Description: Since beginning of the process, more than 30 new drugs have been assessed by G-BA supported by Institute for Quality and Efficiency in Health Care (IQWiG). When launching drugs with NCEs manufacturers must submit a benefit dossier comparing the new drug against the AC to G-BA. G-BA supported by IQWiG will assess this dossier. Based on the assessment’s result G-BA decides on the new drug’s additional benefit. Drugs with no additional benefit will get a reference price immediately if there is a reference-pricing group; the beneficial drug is allowed to keep its price for one year. During that time, manufacturers must negotiate a discount with the GKV-SV. If the negotiation fails, an arbitration body will determine the discount for the drug. Overall, workshop attendees will learn about the characteristics of the German regulatory framework and develop a notion on how to cope with the process at an early stage within product development and in later stages from a national and a global perspective. Detailed information is given about the contents of the dossier (e.g. comparators in trials, patient relevant endpoints). In a simulation of the hearing process at G-BA to give relevant stakeholders the opportunity to comment on the assessment participants will actively take over roles of stake holders to learn more about the decision making process.

Patient-Reported Outcomes & Patient Preference Research

W18: CHOICE DEFINES VALUE: NEW APPROACHES TO ESTIMATING QALYS, HYES, AND EFFICIENCY FRONTIERS
Discussion Leaders:

Benjamin M Craig, PhD, Assistant Faculty Member, Health Outcomes & Behavior, Moffitt Cancer Center, Tampa, FL, USA; Juan Marcos Gonzalez, PhD, Research Economist, Health Preference Assessment, RTI Health Solutions, Research Triangle Park, NC, USA; Axel C. Mühlbacher, PhD, Professor for Health Economics and Health Care Management, IGM Institute, Hochschule Neubrandenburg, Neubrandenburg, Germany

Purpose: To describe innovative advances in the application of discrete choice experiments (DCEs) for health valuation, which have yet to become mainstream.
Description:

The workshop presenters have conducted DCEs for the National Institutes of Health (NIH), Center of Prevention and Disease Control (CDC), the Pharmaceutical Research and Manufacturers of America (PhRMA), and Food and Drug Administration (FDA) in the United States, as well as Institute for Quality and Efficiency in Health Care (IQWiG) in Germany, and an array of pharmaceutical companies. This work in health valuation has shown that the integration of DCE evidence into well-known decision frameworks—specifically quality adjusted life years (QALYs), healthy year equivalents (HYEs), and efficiency frontiers—has the potential to revolutionize health care decision-making at the individual, clinic, system, and national level. Yet, this innovation risks upsetting the fragile balance struck between scientists and pragmatists to achieve the appearance of methodological convention marketed to health policy makers (e.g., National Institute for Health and Clinical Excellence; NICE). While DCEs have been used in psychophysics and economics for over a century and have been applied progressively in health, only recently health regulatory agencies and industry associations have incorporated and endorsed DCE evidence for health valuation. Loyalties to prohibitively complicated scale-based techniques, such as the time trade-off (TTO), remain pervasive, even after DCEs have been shown to be methodologically superior. This workshop will discuss methodological advantages and shortcomings of DCEs compared to traditional methods (such as TTO), challenging common perceptions. Published examples will be presented to illustrate the potential of DCEs, showing how any health outcome or medical endpoint can be valued to inform medical decision-making without using bounded scales or personal interviewing (e.g., TTO). Using hands-on exercises, the workshop will instruct in non-technical terms how choice defines values, specifically how to estimate QALYs, HYEs, and efficiency frontiers using DCE.

WORKSHOPS - SESSION IV
WEDNESDAY, 7 NOVEMBER 2012: 8:45-9:45

Clinical Outcomes Research

W19: PREDICTIVE MODELLING OF REAL-WORLD OUTCOMES: HOW USEFUL FOR HTA EVALUATIONS?
Discussion Leaders:

Billy Amzal, PhD, Senior Scientific Vice President, LA-SER Analytica, London, UK; Venkat Timmaraju, PhD, Senior Statistical Consultant, LA-SER Analytica, London, UK; Helene Karcher, PhD, Expert Modeler, Novartis Pharma AG, Basel, Switzerland; Adam Lowy, MB, ChB, MSc, Expert Global Epidemiologist, Novartis Pharma AG, Basel, Switzerland

Purpose: This workshop's purpose is to illustrate how advanced modelling and simulation tools can serve HTA evaluations via real-world outcomes predictions, provided a transparent and robust assessment of uncertainty.
Description: HTA evaluations typically require the synthesis of all relevant evidence in relation to effectiveness and risks of the use of a drug. Although evidence synthesis methods such as network meta-analyses have been increasingly used to elaborate or defend pricing and reimbursement strategies, these methods have mainly been limited to deriving quantitative summaries of evidence from randomized controlled trials (RCT). Though even more relevant to HTA assessments, the projections of such summaries to real-world settings are more rarely used as they require the integration of more heterogeneous information sources (e.g. from RCTs, observational studies, surveys) and often more complex modelling to adjust for the actual drug use observed in real-life. Using one example in HIV and one in the cardio-vascular area, this session will exemplify situations were such predictive modelling can actually be used to simulate real-world outcomes and quantify the current knowledge relevant for health technology assessments. In each case, model assumptions, inputs and outputs will be discussed, and the added value discussed from different angles. Finally, the use of such predictive models for the design of prospective real-world studies will also be presented and discussed.  A particular focus will be laid on how to ensure transparency and robustness of such model-based approaches. A broad audience composed of modellers, clinicians and decision makers will be expected to share their experiences and views.
W20: META-ANALYSIS OF RARE EVENTS: SUGGESTING A PRACTICAL GUIDANCE
Discussion Leaders:

Julie Roiz, MSc, Project Leader, HEOR, OptumInsight, Nanterre, France; B. Schweikert, PhD, Senior Lead Analyst, Life Sciences, OptumInsight, Munich, Germany; Keith Abrams, MSc, PhD, Professor of Medical Statistics, Department of Health Sciences, University of Leicester, Leicester, UK

Purpose: To review the methods and guidelines on meta-analysis of rare events, and to propose practical guidance to select appropriate methods.
Description:

Meta-analysis involves the statistical synthesis of results from two or more studies, to arrive at a single estimate often using clinical data.  By combining several (clinical) studies while maintaining randomization, meta-analyses increase the power of the estimation and improve precision. Many guidelines exist on how to conduct meta-analyses and correct for any bias. Most guidelines also address the issue of how to deal with sparse data, especially in the case of rare events. However, none of these guidelines provide practical guidance in terms of precisely what to do and when. At this workshop, we will present the different methods for statistical pooling, and the methods suggested in diverse guidelines on how to deal with rare events such as imputation, transformation, or the choice of specific models. We will also present evidence from the published literature that is helping to shape which method might lead to the least biased estimate and under which circumstances. We will also propose a practical guidance using examples from the literature. This guidance will offer straightforward advice and will propose a decision-aid, accounting for many issues, such as the reason for the rarity of data, the type of outcome considered the programming abilities, and the main objective of the meta-analysis. The audience will be encouraged to share their own experience with sparse data to improve on the suggested guidance.

Economic Outcomes Research

W21: WHY, WHEN AND HOW TO CONDUCT ECONOMIC EVALUATIONS USING COMPREHENSIVE DECISION ANALYTICAL MODELING?
Discussion Leaders:

Samuel Aballea, MSc, Director, HEOR, Creativ-Ceutical, Paris, France; Benjamin Briquet, BSc, Student, Paris Institute of Statistics (ISUP), Paris, France; Clément François, PhD, Divisional Director, Global Outcomes Research, Lundbeck SAS, Issy-les-Moulineaux, France; Anne-Lise Vataire, MSc, Senior Analyst, Creativ-Ceutical, Paris, France

Purpose:

The principle of comprehensive decision analytical modelling (CDAM) is to combine the different steps of an economic evaluation (meta-analysis, estimation of parameters, model evaluation and sensitivity analyses) into one coherent model. In this workshop, we will review and discuss the pros and cons of CDAM vs. conventional approaches, using an evaluation of antidepressant drugs as example.

Description: The workshop will start with an overview of the utilisation of CDAM in published evaluations. A key advantage of this approach is to directly propagate uncertainty around results of meta-analyses to the cost-effectiveness estimates. We will present an overview of the implementation of an evaluation of antidepressants using CDAM, programmed in WinBUGS, and compare the results with conventional modelling approach. Based on this example, the assumptions and model characteristics under which the results based on CDAM and conventional approach differ will be investigated. Most examples presented in the literature are based on relatively simple models. Issues arising with more complex models will be discussed, such as how to incorporate correlations between different endpoints evaluated in the same studies. We will show how CDAM offers potential to reduce the variability and bias around cost-effectiveness estimates by borrowing strength across endpoints and broadening the scope of evidence considered. The workshop will cover practical issues as well as theoretical issues associated with the use of CDAM. The audience will be invited to share their experience with CDAM following presentations, and the workshop will conclude with a discussion about the applicability and usefulness of this approach in real situations.

Health Care Policy Development Using Outcomes Research

W22: USING TIME DEPENDENT ENDPOINTS TO INFORM REIMBURSEMENT DECISION OF CANCER DRUGS IN THE ABSENCE OF MATURE OVERALL SURVIVAL DATA
Discussion Leaders:

Jorge Félix, MSc, Director, Exigo Consultores, Alhos Vedros, Setúbal, Portugal; Anant Murthy, PhD, Executive Director, Global Pricing and Market Access, Celgene Corporation, Boudry, Switzerland; Stefan Michiels, PhD, Biostatistician, Breast Cancer Translational Research Laboratory, Institut Jules Bordet, Brussels, Belgium

Purpose: Supporting health care sector decisions using time-dependent endpoints (TDE) such as time to progression (TTP), progression-free survival (PFS), and event-free survival (EFS) in the absence of mature survival data remains controversial. This workshop will review key aspects of predicting OS from TDE in order to set the discussion about the merits of using TDE to support reimbursement decisions of cancer drugs in the absence of mature overall survival data.
Description: Time dependent endpoints like TTP, PFS and EFS are commonly accepted by the European Medicines Agency and the US Food and Drug Agency in the approval of new cancer drugs. Once in the market, the pressure to use this new cancer drugs overcomes the time needed to have a long term vision of OS, frequently leaving decision makers relying on the results of methods to estimate OS from TDE. There some evidence suggesting that TDEs such as PFS, TTP, and EFS are appropriate surrogate endpoints for OS in several types of cancer. However, some conflicting evidence and some methodological, regulatory, and conceptual/practical arguments fuel the on-going discussion about surrogate endpoints in the cancer literature and challenge the establishment of TDE in oncology financing decisions. This workshop aims to discuss these controversial issues from different perspectives.
W23: UNIFYING COVERAGE AND RESEARCH DECISIONS: HOW CAN QUANTITATIVE ANALYSIS INFORM THE ASSESSMENTS REQUIRED?
Discussion Leaders:

Karl Claxton, PhD, Professor of Health Economics, Centre for Health Economics, University of York, Heslington, York, UK; Adrian Towse, MA, MPhil, Director, Office of Health Economics, London, UK; Claire McKenna, PhD, Research Fellow, Centre for Health Economics, University of York, Heslington, York, UK; Marta O Soares, MSc, Research Fellow, University of York, Heslington, York, UK

Purpose: To identify the sequence of assessments required to make coverage and research decisions and understand how each of these assessments can be practically informed by quantitative analysis.
Description: The value of access to a technology and the value of additional evidence about its performance are central to the question of whether coverage should be withheld until research findings are available or granted while the research is being conducted. The workshop will set out the context, principles and methods by which this policy choice can be more accountably informed through interactive presentations. Recent research developed a general framework and algorithm which identifies the sequence of assessments required.  These include: i) expected cost-effectiveness and population net health effects over time; ii) the value of different types of evidence; iii) whether the type of research required to provide this evidence can be conducted once coverage is granted; iv) whether other sources of uncertainty will resolve over time; and v) whether there are significant irrecoverable (opportunity) costs.  The sequence of assessments can be summarised as seven point check list which is applied to two case studies (from NICE Technology Appraisal), to demonstrate practically how they can be informed by the type of quantitative analysis that is currently available. The case studies allow a practical demonstration of how methods of analysis can be applied when a range of interesting characteristics are present, including: the significance of irrecoverable opportunity costs; the presence of different sources of uncertainty; time to research reporting; and complexity associated with multiple alternatives.  They also demonstrate the general insights from this research; that cost-effectiveness is a necessary but not sufficient condition for coverage, since restricting coverage may be appropriate even if a technology is expected to be cost-effective, i.e., it provides the link between price, uncertainty and the need for evidence.
WORKSHOPS - SESSION V
WEDNESDAY, 7 NOVEMBER 2012: 13:45 - 14:45

Clinical Outcomes Research

W24: SURVIVAL ANALYSIS IN HTA: IS CURRENT PRACTICE BEST PRACTICE?
Discussion Leaders:

John William Stevens, PhD, Statistician, School of Health & Related Research (ScHARR), University of Sheffield, Sheffield, UK; Noemi Muszbek, MSc, Research Scientist, United BioSource Corporation, Budapest, Hungary; Martin William Hoyle, MA, PhD, Senior Research Fellow, PenTAG, University of Exeter, Exeter, UK; Edit Remak, MSc, Research Scientist, United BioSource Corporation, Budapest, Hungary

Purpose: Predicting survival beyond the time horizon of clinical trials is a challenge.  Extrapolating survival patterns observed in clinical trials using parametric methods maximizes the internal validity of economic evaluations.  However, the choice of survival distribution can have a significant impact on the estimated overall costs and health benefits of a technology; therefore, careful selection of the appropriate survival distribution is vital. The aim of the workshop is to challenge commonly used methods, specifically to consider extrapolation as more than simply an exercise in curve fitting.    
Description: The workshop will begin with a brief introduction of the need for survival analysis (SA) in cost-effectiveness analysis, and its potential uses and pitfalls.   Next, the different approaches, criteria for the choice of methods, together with their justification will be reviewed.  Common practice in this area will be challenged both in terms of methodology and choice of distribution. This will be followed by a series of case studies from oncology. Issues covered include: considerations of the underlying disease process; heterogeneity of the patient populations; the assumptions behind candidate distributions and their clinical appropriateness; examples on the use of registries, retrospective databases, and the published literature to provide additional information to support or reject assumptions based on clinical trials; and statistical/modelling issues related to representing uncertainty in extrapolation.   Audience participation will be encouraged as attendees will be asked to rank the criteria for selection of survival distributions at the start of the workshop. The workshop will conclude with a discussion of the results of the survey of attendees in the light of the workshop presentations. This workshop is appropriate for health economic modelers and more broadly those responsible for assessing economic evaluations involving extrapolation of survival.  In addition, it will provide valuable insight to those less familiar with the technical details of SA. 

Economic Outcomes Research

W25: ISSUES IN MODELING THE PROGNOSIS AND TREATMENT OF NON-COMMUNICABLE DISEASES USING PATIENT-LEVEL SIMULATION: DO THE GUIDELINES HELP US?
Discussion Leaders:

Talitha L. Feenstra, PhD, Health Economist, Epidemiology, University Medical Centre Groningen and RIVM, Groningen, The Netherlands; Christian Asseburg, PhD, MSci, Technical Director, ESiOR Oy, Kuopio, Finland; Annemieke Leunis, MSci, Researcher, Institute for Medical Techonology Assessment (iMTA), Rotterdam, The Netherlands; Paul F. M. Krabbe, PhD, Associate Professor, Department of Epidemiology, University of Groningen, Groningen, The Netherlands

Purpose: To discuss the use of discrete event modeling and other patient-level simulation techniques to evaluate non-communicable diseases (NCD) treatment. Examples of approaches will be presented, followed by a discussion of alternatives, also in relation to recent ISPOR/SMDM guidelines. Issues to be addressed include modeling of long term complications alongside main treatment effects; dealing with competing risks, and internal and external validation. Furthermore the added value of using patient-level simulation will be discussed. 
Description: One thing that all of these topics have in common is that various approaches can be seen in the literature and that they vary in complexity. Each topic will be introduced by presenting a concrete application, illustrating one alternative approach. We will then contrast the approach used with the advice found in the ISPOR/SMDM guidelines, if any, to open up discussion with the audience concerning the strengths and weaknesses of different approaches. We will also discuss whether a uniform “best” approach should be strictly followed or whether the choice should be left to depend on the particular disease and intervention being modeled.   The discussion leaders share experience in disease modeling of NCDs, while their background varies over economics, statistics and epidemiology. They will bring examples from their own recent modeling work on cancer to illustrate the methodological topics pointed out above. Target audience: researchers interested in patient-level simulation models. 
W26: HETEROGENEITY IN COST EFFECTIVENESS ANALYSIS: IMPLEMENTING METHODS TO REALISE ITS VALUE
Discussion Leaders: Pedro Saramago, MSc, Research Fellow, Centre for Health Economics, University of York, York, Heslington, UK; Manuel Antonio Espinoza, MD, MSc, PhD Student, Centre for Health Economics, University of York, Heslington, York, UK; J. Grutters, PhD, Postdoctoral Researcher, School for Public Health and Primary Care, Maastricht University, Maastricht, Limburg, The Netherlands; JL (Hans) Severens, PhD, Professor of Evaluation in Health Care, Institute of Health Policy and Management, Erasmus University Rotterdam, Rotterdam, The Netherlands
Purpose: This workshop aims to present and discuss recent methodological contributions to the analysis of heterogeneity in cost-effectiveness analysis.
Description: Decisions about which medical technologies to reimburse/fund based on average cost effectiveness estimates may disguise sources of heterogeneity. Making decisions what formally consider between-patient heterogeneity has been proved consistent with an efficient use of limited resources. This workshop presents and discusses the current available methods to analyse, characterize and inform cost effectiveness analysis when heterogeneity has been taken into account. This includes analytical techniques to express the value of subgroup cost effectiveness analysis and individualised care. It highlights that heterogeneity has two dimensions of value, first, the higher expected benefits derived from different decisions under current information, and second, the potential additional value of further research conditional to that heterogeneity. This is illustrated with a trial based cardiovascular cost-effectiveness study. Moreover, it discusses the potential importance of access to individual level evidence, compared to the use of aggregate data, in guiding and in quantifying the value of further research in the absence and in the presence of subgroups. A Public Health case study is used to support these elements. Finally, the workshop discusses the role and merits of the expected value of individualised care (EVIC) framework in providing opportunities to improve efficiencies by collecting individual level information. The feasibility of EVIC is demonstrated in the context of a glaucoma case study. The workshop is structured in order to acknowledge that a continuum exists from decisions made at the mean-, at the subgroup- and at the individual-level. Participants will improve their knowledge of how heterogeneity can be dealt within health care and are encouraged to appraise and/or propose extensions to existing frameworks.

Health Care Policy Development Using Outcomes Research

W27: ASSESSING THE REIMBURSABILITY OF NEW PRODUCTS: A STRUCTURED APPROACH
Discussion Leaders:

Joseph DiCesare, MPH, Global Health Economics & Outcomes Research, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA; Lou Garrison, PhD, Professor, University of Washington School of Pharmacy, Seattle, WA, USA; Gerry Oster, PhD, Vice President, Policy Analysis Inc. (PAI), Brookline, MA, USA; Patricia Sacco, MPH, RPh, Director, HEOR, Novartis Pharmaceuticals Corporation, Hanover, NJ, USA

Purpose: To explain and discuss the development and initial pilot testing of a new multi-criteria tool designed to assess the potential reimbursability of new products.
Description: Health economics & outcomes research (HE&OR) professionals in pharma and biotech companies often are asked to provide input into portfolio management and commercial planning decision making to increase the likelihood that new products will achieve favorable pricing and reimbursement.  Needed input should take account of numerous factors that can affect payer decision making, including safety, efficacy, budgetary impact, cost-effectiveness, and a host of other potential factors.  In practice, the input actually provided for these early decisions is often incomplete, variable, and inconsistent, and does not take into account all relevant considerations.  In this workshop, we will present and discuss a new multi-criteria tool that we have developed to examine, in a more structured fashion, factors that drive stakeholder decision making. We will explain how this tool was developed and its underlying conceptual framework. We will also present findings from initial pilot testing of this tool.  We will engage the workshop participants with a hypothetical product profile that they can use to practice with the tool, assessing the strengths and weaknesses of a new product with regard to likelihood of securing broad market access and reimbursement.  Finally, we will identify the key challenges for assessing reimbursability both with current approaches and this new tool, and suggest possible solutions. Participants in this workshop will interact directly with presenters and discussion leaders and will be engaged via a case study and the discussion of our pilot study results.
W28: THE TRADE-OFF BETWEEN QUALITY AND COSTS OF PRIMARY HEALTH CARE: THEIR IMPACT IN THE PROCESS OF DECISION-MAKING
Discussion Leaders:

Antonio Sarría, MD, MPH, Director, Health Technology Assessment Agency, Instituto de Salud Carlos III, Madrid, Spain; Simo Kokko, MD, PhD, Senior Professor & Researcher, National Institute for Health and Welfare, Helsinki, Finland; Marje Oona, MD, PhD, Assistant Professor, Senior Researcher, Department of Family Medicine, University of Tartu, Tartu, Estonia; Kadri Suija, MD, PhD, Researcher, Department of Family Medicine, University of Tartu, Tartu, Estonia

Purpose:

The purpose of this workshop is to present a series of methods to measure quality and cost of primary care services in Europe, and to discuss the link between cost and quality in different primary care models existing in Europe.

Description:

While the literature reviewing the normative characteristics of health systems is plentiful, a common framework to describe primary health care models in Europe is not available. In addition, a trans-national consensus on how to define quality of care does not currently exist and costs of primary care are not well identified in national accounting systems. Research tends to focus on secondary care due to higher budgets and technology intensive procedures. This workshop will be based on the contribution of the EUPRIMECARE consortium to measuring quality and costs in Europe. EUPRIMECARE is a project funded by 7th EU Framework Programme with partners from seven countries (Italy, Spain, Finland, Estonia, Lithuania, Hungary, Germany). The consortium has defined a framework to characterize primary care models based on 5 dimensions: financing, regulation, payment, organization and organizational behaviour. The methodology proposed to measuring costs of primary care services includes a macro- and micro-costing approaches. Micro-costing has been done through clinical vignettes used in order to assess how specific primary care services are provided in the partner countries. The time-driven activity-based costing was then used to cost each primary care activity in each country. The assessment of quality was based on several dimensions of quality (access, equity, appropriateness and satisfaction) through an audit of clinical records as well as surveys to the population and professionals in all the partner countries mentioned before. The main results of the project will be presented in the workshop, and their policy implications will be discussed.

WORKSHOPS - SESSION VI
WEDNESDAY, 7 NOVEMBER 2012: 15:00 - 16:00

Clinical Outcomes Research

W29: SAMPLE SIZE ESTIMATION FOR OBSERVATIONAL STUDIES
Discussion Leaders:

Terry Alan Cox, MD, PhD, Director, Biostatistics, Real World & Late Phase Research, Quintiles Outcome, Rockville, MD, USA; Eric Gemmen, MA, Senior Director, Biostatistics & Outcomes Research, Real World & Late Phase Research, Quintiles Outcome, Rockville, MD, USA; Mark Nixon, PhD, Director, Biostatistics, Real World & Late Phase Research, Quintiles Outcome, Reading, Berkshire, UK; Pablo Mallaina, MD, MPH, PhD, Senior Medical Manager - Champix, Primary Care BU Europe Canada Australia & NZ (PECANZ), Pfizer, Inc., Madrid, Spain

Purpose: To highlight the challenges and offer potential solutions to sample size estimation for observational studies, through the use of numerous case examples and live calculations.
Description:

Unlike randomized clinical trials (RCTs), prospective observational studies and patient registries typically address objectives rather than test specific hypotheses. Nevertheless, estimation of sample size is an important part of the planning process for these studies.  A minimum sample size is required to allow for adequate exploration of the objectives and to ensure sufficient generalizability.  Sample size estimation for observational studies is more complex than sample size calculation for RCTs; subgroup analyses and modeling are to be expected in observational studies, and these analysis methods require more assumptions and larger sample sizes. On the other hand, sample sizes must not be so large as to raise concern that the study includes an unnecessarily high number of sites and patients.  This workshop will provide examples/case studies of sample size estimations performed for a variety of observational studies with an array of objectives, including burden of illness, comparative effectiveness, comparative safety and personalized medicine.  We will focus on sample size estimation methods for observational studies that take the following analysis techniques into account: (1) Outcome comparisons against historical comparators. (2) Outcome comparisons where patients serve as their own control (i.e., historical control). (3) Investigation of factors that influence outcomes within subgroups. (4) Propensity score matching to support comparisons of cohorts or subgroups. (5) Time-to-event analyses including a) the stabilization of laboratory values within an acceptable range, b) disease remission, c) major adverse cardiac event, etc. (6) Multiple comparison adjustments to support comparisons between multiple study sites and multiple patient types over multiple timepoints. (7) Re-estimation of sample size based on interim results. Audience input based on experience will be encouraged throughout the session, which will include live use of sample size estimation software.

Economic Outcomes Research

W30: DEMONSTRATING THE BENEFITS OF ONCOLOGY TREATMENTS: MINIMISING UNCERTAINTY AND BIAS
Discussion Leaders:

Nicholas Latimer, MSc, Research Fellow in Health Economics, School of Health and Related Research (ScHARR), University of Sheffield, Sheffield, South Yorkshire, UK; Neil Hawkins, PhD, Vice President, Health Economics, Oxford Outcomes, Oxford, UK; Mike Spencer, MSc, Senior Director, Health Economics, Market Access and Reimbursement, Janssen EMEA, Janssen-Cilag Limited, High Wycombe, Bucks, UK

Purpose:

Any assessment of the comparative value of treatments requires estimates of the expected outcomes for patients were they to receive each of the potential alternatives. The primary source of information for these estimates, especially for newly developed treatments will be randomised clinical trials.  However, estimates made directly from trial data may be biased (the estimate is incorrect) or highly uncertain (the estimate is not useful). Together, bias and uncertainty lead to the risk of incorrect decisions. With respect to oncology trials, estimates of overall survival may sometimes be highly uncertain, essentially due to the limited number of deaths observed within a trial. In addition crossover of patients from placebo to active treatment upon progression can lead to biased estimates of the difference in overall survival between placebo and active treatment.

Description: During this workshop we will review and discuss the practical application of statistical methods that seek to reduce the bias caused by treatment crossover and methods of evidence synthesis that combine trial and external evidence to increase certainty in our estimates (particularly with respect to overall survival). The audience will be challenged to play the role of a decision maker evaluating the value of a new product using these methods.

Health Care Policy Development Using Outcomes Research

W31: FROM DECISION POINT TO DECISION WINDOW: READINESS FOR A CHANGE OF PARADIGM
Discussion Leaders:

Mondher Toumi, MD, PhD, Professor, Market Access, University Claude Bernard Lyon 1, Lyon, France; Thomas Mueller, MD, Head, Pharmaceuticals Department, Federal Joint Committee (G-BA), Berlin, Germany

Purpose:

In several European countries, pricing and reimbursement decisions are no longer taken at a fixed point time, but over a period of time over which pharmaceutical companies are expected to provide additional evidence, to reduce uncertainty around outcomes of a new product. This workshop will present the implications of these changes when pharmaceutical companies develop market access strategies.

Description:

Historically reimbursement decisions used to happen at a point in time and were usually difficult to reverse unless a serious safety issue emerged. Today regulations tend to offer payers and regulators opportunities to review their decision over time. Different countries used different processes to control uncertainty. Regulators instituted conditional approval as well as risk management plans.  Payers refer more and more to coverage with evidence development (CED), which may be completed with an escrow agreement. In Germany the price decisions happen now after a long period that could last over one year. In France CED is becoming more and more frequent.  As pharmaceutical companies increasingly become accountable for the opportunity costs associated with uncertainty around the value of a product, the amount and complexity of information that they need to generate increases. The decisions that companies must take internally are increasingly complex, especially as payer requirements vary between countries. Modeling will become an essential tool to inform these decisions.  The initial experience of the new regulation in Germany will be shared. The audience will be invited to share their own experience and provide input on how to optimally address payers and regulators expectation for uncertainty management over the decision window. This change of paradigm is associated with an increase communication and collaboration between payers and regulators.

W32: LISTENING TO THE PATIENT- DEVELOPING STRATEGIES FOR ENHANCING THE USE OF DATA RELEVANT TO PATIENTS IN HEALTHCARE DECISION MAKING
Discussion Leaders:

Michelle Mocarski, MPH, Manager, Health Economics and Outcomes Research, Forest Research Institute, Inc., Jersey City, NJ, USA; Asha Hareendran, PhD, MA, Senior Research Scientist, United BioSource Corporation, London, UK; Keith Tolley, Mphil, BA, Director, Tolley Health Economics, Buxton, UK

Purpose:

The workshop will explore and review the many ways that patient-reported outcome measures can be used to guide decision making by various stakeholders—using Chronic Obstructive Pulmonary Disease (COPD) as an example.

Description:

In recent years, interest in understanding patients’ perspectives on treatment has grown substantially. There is, however, some lack of understanding of the value of data collected in clinical trials beyond regulatory approval. COPD, in particular, has been an area where patient-reported outcomes (PROs) are increasingly being used to inform stakeholder decision making. Using COPD as an example, this workshop will enable the audience to explore methods by which data from patients can be used for decision making beyond product registration and discuss innovative ways for capturing the patients’ perspective for evaluating new products. The workshop will begin with a brief introduction of COPD, including vignettes of patients’ experiences of COPD. The audience will be involved in an exercise to discuss patient-relevant outcome endpoints that could be used for making decisions about treatments for the example cases.  The panel will present specific examples of PRO data and tools that are being used in Europe for evaluating treatments for COPD, by various stakeholders —clinicians, registration and reimbursement agencies. The workshop will conclude with a discussion of emerging opportunities for engaging patients’ views for decision making, including an economist’s view about how patient experiences in COPD can be used for making decisions about the cost-effectiveness of  treatments.

Patient-Reported Outcomes & Patient Preference Research

W33: INCREASED STATISTICAL POWER FOR PRO OUTCOMES – USING ITEM RESPONSE THEORY METHODS TO DEVELOP COMPOSITE SCALES
Discussion Leaders:

Jakob Bue Bjorner, PhD, Chief Science Officer, QualityMetric, OptumInsight Life Sciences, Lincoln, RI, USA; Mark Kosinski, MA, Senior Scientist & Vice President, Outcomes Insight Consulting, QualityMetric, OptumInsight Life Sciences, Lincoln, RI, USA; Matthias Rose, MD, PhD, Professor, Department of Psychosomatic Medicine, Charité - University of Medicine, Berlin, Germany

Purpose:

Participants will: 1) learn methods for combining scales from different PRO instruments in order to increase power in statistical analyses; 2) understand the measurement requirements for such aggregation; and, 3) learn methods for expressing the results in the metrics of the original scales.

Description:

When patient reported outcomes are used in medical research, the same domain is often assessed by two or more scales from different questionnaires. In these instances, a more coherent analysis with greater statistical power can be achieved by combining scales measuring the same domain into a composite scale. Composite scales can be particularly useful when the original scales differ in their floor and ceiling effects. Item response theory (IRT) provides methods for evaluating measurement requirements, establishing the optimal way to combine information from different scales into an overall score, and developing cross-calibration procedures to present the results in the metric of the original scales. The workshop will consist of theoretical lectures outlining the methods to carry out the IRT analyses and practical examples of analyses of clinical trials. Particular attention will be given to checks of the statistical assumptions behind the approach and to methods for revising the IRT model. Finally, the workshop will provide examples of projects to develop cross-calibration between different scales.