Workshops
Monday, May 18, 2015
5:00 PM - 6:00 PM
WORKSHOPS - SESSION I
Health Policy Development Using Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon A, B, F (Level 5)

W1: PATIENT-CENTERED BENEFIT-RISK ANALYSIS: REGULATORY DEVELOPMENTS AND PROSPECTS

Discussion Leaders:

F. Reed Johnson, PhD, Senior Research Scholar, Duke Clinical Research Institute, Duke University, Durham, NC, USA

John F.P. Bridges, PhD, Associate Professor, Department of Health Policy and Management and International Health, John Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

Lou Garrison, PhD, Professor, Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy, University of Washington, Seattle, WA, USA bio

Bennett Levitan, MD, PhD, Senior Director, Benefit-risk Assessment, Epidemiology, Janssen Research & Development, LLC, Titusville, NJ, USA

PURPOSE:

This workshop will focus on recent efforts to engage patients in regulatory decision making and to incorporate patient preferences into regulatory benefit-risk analysis. Workshop participants will become familiar with alternative methods for eliciting and summarizing qualitative and quantitative data on patients’ benefit-risk tradeoff preferences and be able to critically evaluate how novel approaches to stakeholder involvement could influence the evolution of regulatory science. 

DESCRIPTION:

Participants will obtain an overview of recent developments at FDA related to patient engagement and use of qualitative and quantitative preference data. The workshop will review a) how benefit-risk preferences are considered by other Federal agencies, b) how regulators have responded to mandates to become more patient-centered, c) how industry uses patient risk-preference data in drug development and regulatory submissions, and d) what role patient advocacy has in regulatory reviews. Drs. Johnson, Bridges, and Levitan will describe recent efforts by FDA's Centers for Drugs and Devices to elicit patient-preference data, including advisory-committee testimony, therapeutic-area public meetings, public-comment dockets on FDA patient engagement, and draft guidance currently in development on including patient-preference data in regulatory submissions. Dr. Garrison will assess the practicality and usefulness of preference data for drug development.  Audience participation will include a survey of attendees’ perspectives on patient-centered benefit-risk analysis and participants will be encouraged to share their experience and perspectives during the workshop. This interactive and informative workshop will be valuable to researchers, clinicians, and industry analysts who are interested in understanding recent developments in patient-centered regulatory benefit-risk assessment.


5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon C, D, E (Level 5)

W2: MAKING BETTER USE OF COMPANY PHARMACOECONOMIC MODELS

Discussion Leaders:

Michael Drummond, MCom, DPhil, Professor of Health Economics, Centre for Health Economics, University of York, Heslington, York, UK bio

Laurie Fazio, BS, Vice President, Market Access Technologies, Dymaxium, Inc., Toronto, ON, Canada

John Watkins, PharmD, MPH, BCPS, Formulary Manager, Premera Blue Cross, Mountlake Terrace, WA, USA

H. Keri Yang, PhD, MPH, MS, Director, Center for Observational and Real World Evidence, Merck & Co., Inc., West Point, PA, USA

PURPOSE:

Pharmaceutical companies invest considerable resources in producing pharmacoeconomic (PE) models, in order to demonstrate the cost-effectiveness of their products. In jurisdictions with a well-defined structure and process for conducting technology assessments, the purpose of models, and their use by payers, is clear. However, in jurisdictions with multiple payers, such as in the US, the reimbursement landscape is much more complex and the precise nature of the use of PE models is less clear and quite variable. The purpose of this workshop is to explore the use of company models by  payers and to determine precisely how and when models can make an effective contribution to improving formulary decision-making.

DESCRIPTION:

The workshop discussion will be informed by the results of a recent survey, in which formulary decision-makers in the US were asked whether they used company models and the ways in which they found them to be useful. The workshop discussion leaders will both reflect on the survey results and share their own knowledge and expertise on the uses of models in the setting(s) they are familiar with. Members of the audience will also have an opportunity to share their own experiences of the use company PE models. The overall objective of the workshop will be to identify some principles of best practice, both in the development of models and their scrutiny by payers.

Use of Real World Data

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon I (Level 5)

W3: SELECTION BIAS AND COST ALLOCATION IN PATIENTS WITH POLYCHRONIC DISEASE

Discussion Leaders:

Kathy L. Schulman, MS, Managing Director, Outcomes Research Solutions, Inc., Bolton, MA, USA

Lei Liu, PhD, Associate Professor, Preventive Medicine-Biostatistics, Northwestern University, Chicago, MA, USA

Robin Turpin, PhD, Director, Takeda Pharmaceuticals USA, Deerfield, IL, USA

PURPOSE:

This workshop will educate the audience on the challenges associated with evaluating economic outcomes in patients with polychronic disease (PCD) and will demonstrate the sensitivity of estimates to various methodologies and assumptions.

DESCRIPTION:

The clinical decision to treat any given patient population with drug A vs. drug B is complicated by the presence of PCD, which can limit the utility of standard therapies, increase the risk of drug-drug interactions, and confound outcome evaluation. Commonly used methods for dealing with selection or ascertainment bias (e.g., propensity score matching, multivariable regression) can be highly sensitive to model specification and are complicated by both the number of comorbidities and the varying levels of severity within each disease state. Disease-based estimates of health care costs are also complicated in this context because care delivered is not always easily assignable to a single procedure, drug, or diagnosis and disease-specific sequelae are not always clinically germane and/or quantifiable. A case study of patients with gout and chronic kidney disease will be used to illustrate the sensitivity of cost estimates to differences in study design and various statistical and nonstatistical assumptions. Additional comparisons will be drawn from published studies estimating the cost of gout therapies, and the advantages and disadvantages of each approach will be discussed. The following topics will be specifically reviewed: 1) Conceptualization and measurement of comorbidity 2) Definition of PCD 3) Options for addressing selection bias, including propensity score and multivariable regression‒based approaches and their impacts on study findings 4) Options for cost allocation (primary, secondary, proportional) and their implications for patient- and disease-based estimates 5) Considerations for multivariate modeling 6) Implications for payers and market access Workshop participants will be encouraged to offer their perspectives and recommendations.


5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon G (Level 5)

W4: IDENTIFYING PATIENTS WITH RARE DISORDERS USING ADMINISTRATIVE DATA

Discussion Leaders:

Daniel C. Malone, PhD, RPh, FAMCP, Professor of Pharmacy, College of Pharmacy and Associate Professor, Mel & Enid Zuckerman College, University of Arizona, Tucson, AZ, USA bio

Wei-Shi Yeh, PhD, Associate Director, Global Market Access, Biogen Idec, Cambridge, MA, USA

Edward P. Armstrong, PharmD, President, Strategic Therapeutics, LLC, Oro Valley, AZ, USA

Eric J. Bell, BS, Principal, One Tall Tree, LLC, Seattle, WA, USA

PURPOSE:

The workshop will demonstrate the challenges and solutions to conducting health outcomes studies involving persons with rare disorders.  Health outcomes studies for rare conditions can be challenging because diagnosis coding in administrative data is nonspecific, frequent use of rule-out diagnoses, lack of relevant disease specific variables in claims data, and few observations. This workshop will provide guidance on selecting an appropriate dataset, identifying the correct cohort, and use of data visualization tools to describe the population.

DESCRIPTION:

Conducting real world health outcomes studies for patients that have rare disorders can be challenging due to the unique combination of health care services used and the lack of sensitivity and specificity of ICD9 diagnosis codes. This session will provide an in depth discussion of issues and potential solutions. Dr. Malone will begin the workshop by providing an overview of the challenges associated with observational data with respect to diagnostic information and prescription drug coding. Dr. Yeh will highlight the need for such studies and how these studies are used as part of the development process for new treatments for rare disorders.  Dr. Armstrong will discuss the challenges and solutions used in conducting studies using observational data for two rare conditions, hemophilia and spinal muscular atrophy. Mr. Bell will demonstrate a data visual analysis tool that was used to display patient-level cost of care, enrollment, and utilization in relation to dates of diagnosis.

Clinical Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon J (Level 5)

W5: INDIRECT COMPARISONS FOR SINGLE-ARM TRIALS OR TRIALS WITHOUT COMMON COMPARATOR ARMS: WHAT METHODS ARE AVAILABLE, HOW HAVE THEY BEEN USED AND HOW CAN WE EVALUATE RESULTS?

Discussion Leaders:

Elyse Swallow, MA, MPP, Manager, Health Economics and Outcomes Research, Analysis Group, Inc., Boston, MA, USA

James Signorovitch, PhD, Vice President, Health Economics and Outcomes Research, Analysis Group, Inc., Boston, MA, USA

Anupama Kalsekar, MS, Director, World Wide Health Economics and Outcomes Research, Bristol-Myers Squibb, Princeton, NJ, USA

Yong Yuan, PhD, Director, World Wide Health Economics and Outcomes Research, Bristol-Myers Squibb Pharmaceuticals Ltd, Princeton, NJ, USA

PURPOSE:

Single-arm trials may be used for regulatory evaluations of efficacy and safety, especially for breakthrough therapies, in rare diseases or in diseases with high unmet medical need. In these settings, single-arm trials can avoid practical and ethical challenges presented by comparator arms. However, once a product is approved, single-arm trials present important challenges for economic evaluations and health technology assessments. In particular, single-arm trials cannot be directly included in traditional indirect comparisons of efficacy and safety that inform economic comparisons. Similar challenges arise when two treatments of interest have only been trialed against different comparators, which often occurs in rapidly-evolving therapeutic areas. This workshop will survey different statistical methods for conducting indirect comparisons across single-arm trials and across trials that do not share a common comparator.

DESCRIPTION:

A number of methods have been used for indirect comparisons without common comparators, including benchmarking based on historical controls, simulated treatment comparisons (also called regression-prediction), network meta-analyses with multiple indirect links, and matching-adjusted indirect comparisons. We will summarize the practical requirements, advantages and limitations of the approaches, and will illustrate the methods through real-world applications. A framework will be provided for answering the following key questions: Given the available trial data, which method(s) should be used? How do the assumptions and limitations differ from those of traditional anchor- or network-based indirect comparisons? How can the appropriateness, quality and reliability of an analysis be assessed? Examples of publically available health technology assessment submissions that included each method will also be reviewed and discussed. The audience will be invited to participate in the discussion of the methods and real-world applications.

Economic Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon H (Level 5)

W6: APPLYING DYNAMIC SIMULATION MODELING METHODS IN HEALTH CARE DELIVERY RESEARCH – THE SIMULATE CHECKLIST AND EMERGING GOOD PRACTICES: REPORTS OF THE ISPOR SIMULATION MODELING EMERGING GOOD PRACTICES TASK FORCE

Discussion Leaders:

Deborah A. Marshall, PhD, Canada Research Chair, Health Services and Systems Research & Associate Professor, Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Alberta Bone and Joint Health Institute, Calgary, AB, Canada

Mitchell Higashi, PhD, Chief Economist, GE Healthcare, Barrington, IL, USA

Peter Wong, PhD, MS, MBA, BSC, RPh, Vice President & Chief Performance Improvement Officer, Illinois Divisions and Medical Group, HSHS, Belleville, IL, USA

Kalyan S. Pasupathy, PhD, Faculty, Healthcare Policy & Research, Mayo Clinic, Rochester, MN, USA

PURPOSE:

To assist researchers and decision makers in deciding whether dynamic simulation methods are appropriate to address specific health systems problems, recommendations from the task force’s two recently published reports will be presented, including the eight-point SIMULATE (System, Interactions, Multilevel, Understanding, Loops, Agents, Time, Emergence) Checklist.

DESCRIPTION:

Health care delivery system interventions are complex, multidimensional interventions among the interconnected elements of a system - beyond a specific drug or technology - including people, processes, and technology.  They are adaptive to changes in the local environment and behave in a non-linear fashion. For health care planning, interventions need to incorporate the dynamics and complexities of the health care system context in which the intervention is delivered. Conventional evaluations of health care interventions are often limited because they neglect the wider health system impacts that could be critical for achieving desired health goals. We present a comparison of three dynamic simulation modeling methods that can be suitable for simulating complex health care delivery and system interventions: agent-based modeling (ABM); discrete event simulation (DES); and system dynamics (SD) modeling.  Simulation results from “what if” scenarios can inform decisions considering intended and unintended consequences of an intervention from multiple dimensions. The specific selection of the appropriate simulation modeling method depends on a number of factors, such as whether the problem is specific to individuals or groups, the level of the problem (strategic, operational or tactical) and whether stochastic or deterministic solutions are sought. We suggest criteria for selecting the best method amongst the alternative dynamic simulation modeling methods for a particular problem, and provide guidance for best practices for dynamic simulation modeling. The audience will be invited to provide examples of problems they face for interactive discussion about the advantages and disadvantages of these alternative modeling methods to solve these problems in that context.

Patient-Reported Outcomes & Patient Preference Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon K-L (Level 5)

W7: ASSESSING PERFORMANCE OUTCOME MEASURES FOR REGULATORY REVIEW: CONCEPTUAL AND METHODOLOGICAL CHALLENGES WITH REAL WORLD EXAMPLES

Discussion Leaders:

Rachel Simone Ballinger, PhD, Senior Outcomes Researcher, Patient-Reported Outcomes, ICON, Oxford, UK

Elizabeth Nicki Bush, MHS, Research Scientist, Global Patient Outcomes and Real World Evidence, Eli Lilly, Indianapolis, IN, USA

Ashley Slagle, PhD, Clinical Outcome Assessment Qualification Scientific Coordinator & Study Endpoints Reviewer, U.S. Food and Drug Administration, Silver Spring, MD, USA

Diana Rofail, PhD, CPsychol, Global Head of Patient-Centered Outcomes Research, Neuroscience and Metabolism, Roche Products Limited, Welwyn Garden City, UK

PURPOSE:

This workshop will explore specific challenges associated with development and use of performance outcome measures (PerfOs) in clinical trials, such as considerations for applying and adapting regulatory guidance originally developed for patient-reported outcome measures (PROs), when appropriate.

DESCRIPTION:

Methods and standards for PerfO development and use are currently evolving through outcomes research studies supporting ongoing drug development programmes and regulatory feedback on Industry plans.  There is, however, absence of specific regulatory guidance specific to PerfOs, with guidance on the use of PRO measures providing a useful framework (FDA/SEALD Guidance for Industry 2009). The establishment of  measurement properties for PerfOs, nevertheless presents specific challenges to ensure that regulatory standards outlined for PRO and patient-focused outcome assessment can be met. This workshop draws on experiences from evaluating PerfOs in two therapeutic areas, both with a strong history of performance measurement: orthopaedics and neurology.  Conceptual and logistical challenges of qualitative and quantitative study design in studies assessing the measurement properties of PerfOs will be discussed, informed by experiences from these studies, including meaning of content validity of PerfOs in the regulatory context, assessment of data saturation, standardisation of PerfO test administration across multiple sites and appropriate thresholds for evidence to support measurement properties. The workshop will include Industry, Regulatory, and Outcomes Research speakers to consider issues and challenges of PerfO validation from multiple perspectives. The workshop will finish with a chaired discussion in which the audience and speakers are invited to comment on the approaches presented and issues raised.

Tuesday, May 19, 2015
3:45 PM - 4:45 PM
WORKSHOPS - SESSION II
Health Policy Development Using Outcomes Research

3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon I (Level 5)

W8: DESIGN OF BUNDLED PAYMENT IN THE AMBULATORY SETTING OF CARE

Discussion Leaders:

Mike Ciarametaro, MBA, Director of Research, National Pharmaceutical Council, Washington, DC, USA

Joshua T. Cohen, PhD, Associate Professor, Center for the Evaluation of Value and Risk in Health,Tufts Medical Center, Boston, MA, USA

Lili Brillstein, MPH, Director, Episodes of Care, Horizon Healthcare Innovations, Newark, NJ, USA

Michael del Aguila, PhD, Vice President, HEOR US Medical, Bristol-Myers Squibb, New York, NY, USA

PURPOSE:

As US health care evolves towards bundled reimbursement, there are complexities which this workshop will explore.  Using specific case examples, we will a) identify when bundled payment can both constrain costs and enable high quality of care and b) when appropriate, identify key considerations that drive selection of services and reimbursement. Panelists from the payer, manufacturer, and bundled payment research communities will address bundled payment appropriateness and bundle design for each case example. Audience participation will be sought on the appropriateness of bundled payment and key success factors.

DESCRIPTION:

Bundled payments have been used in the inpatient environment for many years. The focus of bundled payment is now shifting to the ambulatory care setting, which creates new challenges as its application broadens to include chronic conditions.  To successfully constrain costs and enable high quality care, it is critical to a) define what is included and excluded from a particular bundle in a manner that allows providers to appropriately manage patients without assuming excess and uncontrollable risk b) support patient variability while reducing unnecessary treatment variation, c) support the evolution of clinical practice to avoid disincentivizing clinicians from using new modalities, and d) establish an adequate quality measure to prevent a shift from over- to under-use of key resources. For each case example, panelists will a) indicate whether bundled payment can both constrain costs and enable high quality of care, b) explore whether each of the aforementioned factors is relevant to the case example and c) identify outstanding research questions that must answered to successfully implement a payment bundle. Audience will be polled on each case study and will have the opportunity to ask questions at the end.

Use of Real World Data

3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon J (Level 5)

W9: METHODOLOGICAL CHOICES FOR ANALYZING CLUSTER-CORRELATED DATA IN LARGE PATIENT DATABASES

Discussion Leaders:

Steve Sherman, MPH, Director, HEOR, Creativ-Ceutical, Chicago, IL, USA

Katia Thokagevistk, PhD, Manager, HEOR, Creativ-Ceutical, Paris, France

Firas Dabbous, PhD, Research Analyst, Epidemiology and Biostatistics, University of Illinois at Chicago, Chicago, IL, USA

Samuel Aballéa, PhD, Vice President, HEOR, Creativ-Ceutical, Paris, France

PURPOSE:

Cluster-correlated data are encountered in real-world databases, where patients and their encounters can be grouped into clusters; e.g., by institution and/or treating physician. Ignoring the dependence between patients from the same cluster may lead to biased results and underestimated variance. Selecting the right approach for modeling these data can prove challenging. The purpose of this workshop is to summarize and provide guidance for selecting the most appropriate methods for accounting for intra-cluster correlation (at each grouped level) in statistical modeling.

DESCRIPTION:

Depending on the structure of the data and the research question, several analytical approaches are available to account for multi-level data. The workshop will first describe and provide illustrations for the general approaches available for handling clustering in regression models. This summary will include details on 1) Mixed effects regression modeling, such as generalized linear mixed models, hierarchical linear models, random coefficient models, and linear random intercept models; 2) Fixed effects (FE) models, including linear and non-linear approaches; and 3) The method of generalized estimating equations (GEE). For each approach, model specifications and assumptions will be outlined. Empirical examples will illustrate the potential strengths and limitations of each method and help stimulate discussion with workshop participants. Furthermore, practical guidance for selecting the best methodology (e.g., the Hausman test of endogeneity) will be summarized and discussed. Propensity score matching will also briefly be introduced in the context of cluster-correlated data.


3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon K-L (Level 5)

W10: RARE DISEASES IN THE ERA OF BIG DATA: SELECTION BIAS IN SMALL SAMPLES

Discussion Leaders:

Nicole M. Engel-Nitz, Ph.D., Senior Researcher, HEOR, Optum, Eden Prairie, MN, USA

Zhimei Liu, PhD, Director, Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA

Cori Blauer-Peterson, MPH, Senior Research Analyst, Health Economics and Outcomes Research, Optum, Eden Prairie, MN, USA

Jonathan C. Johnson, MS, Associate Director, HEOR, Optum, Eden Prairie, MN, USA

PURPOSE:

Participants in this workshop will explore methodological issues related to performing outcomes research on rare diseases in large observational databases, with an emphasis on adjusting for selection bias when the study population is small.

DESCRIPTION:

The increasing availability of large administrative claims and electronic health records databases offers significant opportunities to assess real-world outcomes for patients with rare diseases. For rare diseases, defined in the US as a prevalence of <64 per 100,000 people, large databases can offer study sample sizes unseen outside of prospective registries. As with observational studies of more common conditions, comparisons of specific treatment or other subgroups in rare diseases may be subject to methodological challenges such as unmeasured confounding and selection bias. Using the example of the rare disease tuberous sclerosis complex (TSC), this workshop will review techniques to address issues of selection bias in the face of challenges specific to rare diseases; in addition to small sample sizes, these challenges include patient heterogeneity and the size of the network of clinicians who are disease experts. Tuberous sclerosis complex, estimated to affect 25,000-40,000 people in the US, is characterized by benign tumor growth which results in systemic manifestations including neurocognitive deficits, seizures, renal disease, and pulmonary disease. Analyses were performed to compare patients treated by clinics expert in TSC versus patients treated in other settings. Given small sample sizes, the reduction in study population with propensity score matching is undesirable; alternative approaches, including instrumental variables techniques and estimation of the true treatment effects in the face of unobserved confounders will be discussed as innovative applications to a rare disease population. Participants will engage in a critical comparison of each technique and its impact on study results through the use of an applied example. Discussion and questions will be solicited throughout the presentation.

Clinical Outcomes Research

3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon C, D, E (Level 5)

W11: CAN WE MAKE COMPARATIVE EFFECTIVENESS USEFUL FOR CLINICIANS AND PATIENTS OR IS IT JUST FOR HEALTH TECHNOLOGY ASSESSMENTS?

Discussion Leaders:

Edward J Mills, PhD, MSc, Visiting Associate Professor, Stanford University, Stanford, CA, USA

Christopher O'Regan, MSc, Head of Health Technology Assessment & Outcomes, Merck Sharp & Dohme Limited, Hertfordshire, UK

Sonal Singh, MD, MPH, Professor, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA

PURPOSE:

Comparative effectiveness and safety of therapeutics are often mandated by Health Technology Assessment (HTA) agencies to assess different interventions. Methods used to assess comparative effectiveness, such as indirect comparisons and network meta-analyses, are often statistically complex and many clinicians and patients are unable to interpret results. While HTA agencies request these methods, these approaches have not caught on widely in the clinical environment and it is possible this is due to complexity. This session will discuss examples where comparative effectiveness methods may change important health related decisions but have been ignored by clinicians and patients due to complexity. It will offer suggestions on making these approaches more transparent and conceptually easier.

DESCRIPTION:

Comparative effectiveness is often mandated by HTA bodies. Comparative effectiveness methods are, however, often statistically complex and many clinical readers are unfamiliar with it. This has resulted in a relatively low uptake of this approach by clinicians and patient groups. In an effort to improve the usefulness of comparative effectiveness methods, this session will discuss efforts to make these approaches interpretable to clinicians and non-methodologist groups. It will discuss feedback from clinical groups about the challenges of using this type of information and implementation in clinical practice. The session will engage the audience to discuss their challenges with using or applying comparative effectiveness approaches and seek input and suggestions to improve the utility of these approaches.

Economic Outcomes Research

3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon A, B, F (Level 5)

W12: ISSUES TO CONSIDER WHEN ESTIMATING HEALTH CARE COSTS WITH GENERALIZED LINEAR MODELS (GLMS): TO GAMMA/LOG OR NOT TO GAMMA/LOG? THAT IS THE NEW QUESTION

Discussion Leaders:

Jalpa A. Doshi, PhD, Associate Professor of Medicine & Director, Economic Evaluations Unit, Center for Evidence-based Practice and Director, Value-based Insurance Design Initiatives, Center for Health Incentives and Behavioral Economics, General Internal Medicine, Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, PA, USA

Henry Glick, PhD, Professor of Medicine, Division of Internal Medicine, University of Pennsylvania, Philadelphia, PA, USA

Andrew Briggs, DPhil, MSc, William R. Lindsay Professor of Health Economics, Health Economics & Health Technology Assessment, Institute of Health & Wellbeing, University of Glasgow, Glasgow, UK

PURPOSE:

The objectives of this workshop are to familiarize participants with issues that arise in the naïve use of generalized linear models and the preferred analytic approaches in the economic analysis of patient-level health care cost data.

DESCRIPTION:

The past decade has seen increasing adoption of methodologically superior techniques for the analysis of patient-level cost data.  The problematic use of log OLS models for the analysis of skewed health care cost data has declined and has increasingly been replaced by use of generalized linear models (GLMs).  However, use of GLMs requires selection of an appropriate family and link function based on the distribution of the data at hand. Review of published cost analyses instead indicates widespread routine adoption of gamma families and log links without any attempt to assess their appropriateness.  Hence, while “To log or not to log” OLS? is no longer the question, the new question is whether “To gamma/log or not to gamma/log” GLM? This workshop will provide an overview of GLMs including the availability of a host of family and link function options to model the cost distribution and the diagnostics tests available to inform selection of family and link functions.  Next, using examples from real-world data we will demonstrate the use of these diagnostic tests for the selection of appropriate family and link functions for the GLM models. Finally, we will outline the advantages and limitations of GLMs and more advanced analytic approaches in various contexts of cost estimation. The workshop will be highly interactive and practical in orientation and will provide examples to illustrate the methods to implement techniques as well as "do's" and “don't's”.

Patient-Reported Outcomes & Patient Preference Research

3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon G (Level 5)

W13: PATIENT- AND OBSERVER-REPORTED OUTCOMES (PROS & OBSROS) MEASUREMENT IN RARE DISEASE CLINICAL TRIALS — EMERGING GOOD PRACTICES

Discussion Leaders:

Katy Benjamin, PhD, MS, Director, Patient Reported Outcomes, ICON, Bethesda, MD, USA

Eleanor Perfetto, PhD, MS, Associate Professor, Pharmaceutical Health Services Research, School of Pharmacy, University of Maryland, Baltimore, MD, USA

Laurie Burke, MPH, RPh, Founder, LORA Group, Royal Oak, MD, USA

Donald L. Patrick, PhD, MSPH, Professor, Seattle Quality of Life Group and Biobehavioral Cancer Training Program, University of Washington, Seattle, WA, USA

PURPOSE:

To provide recommendations on measuring PROs & ObsROs in rare diseases. 

DESCRIPTION:

  Measuring how patients feel and function is increasingly considered to be an integral component of the development, review and regulation of new treatments.  Valid accurate information about patients’ treatment and condition-related experiences is collected in rare disease clinical trials. Nonetheless, there are significant challenges to developing, modifying and selecting PRO and ObsRO measures for rare disease treatment evaluation. Thus, standard methods and strategies of PRO and ObsRO development, validation and implementation, including those recommended by regulators, need to be interpreted in the context of the unique problems associated with RD populations. However, there is no one correct solution for addressing the myriad and diverse challenges that will arise when implementing PROs or ObsROs in a RD clinical trial and developing evidence to support their use within various RD contexts of use. The task force's recommendations focus on methods for PRO instrument selection, adaptation, and validation; optimal strategies for using observer reports; study recruitment; and modes of instrument administration. Conformance to regulatory guidance for the evaluation and proof of treatment benefit using PRO endpoints is taken into account in terms of how PRO measures and methods can be developed or adapted recognizing the unique challenges of RD.  The authors draw on the clinical outcomes assessment (COA) framework and use specific examples to discuss the potential obstacles involved in implementing or developing PRO and ObsRO endpoints for RD studies.  Panel members will present possible solutions to address common challenges that arise when working with RD populations, emphasizing pragmatic approaches to these challenges.  The audience will be invited to provide examples of problems they face for interactive discussion about the advantages and disadvantages of various options for developing and implementing  patient-centered measures that are valid and appropriate for context of use.


3:45 PM - 4:45 PM
Room: Grand Ballroom, Salon H (Level 5)

W14: CHARTING A PATH FOR THE ELICITATION OF PREFERENCES: PERSPECTIVES ON THE EFFECTIVE USE OF PREFERENCE-ELICITATION METHODS

Discussion Leaders:

Josephine A. Mauskopf, PhD, MHA, MA, Vice President, Health Economics, RTI Health Solutions, Research Triangle Park, NC, USA

Maarten J. IJzerman, PhD, Professor & Chair, Department of Health Technology & Services Research, University of Twente and MIRA Institute for Biomedical Technology & Technical Medicine, Enschede, The Netherlands

Juan Marcos Gonzalez, PhD, Senior Research Economist, Health Preference Assessment, RTI Health Solutions, Research Triangle Park, NC, USA

PURPOSE:

Introduce participants to alternative methods for preference elicitation and how these methods differ in terms of their complexity, the type of questions they can answer and the contexts in which they can be used most effectively.  Attendees will participate in a hands-on exercise to distinguish suitable approaches among a number of methods based on the features of research questions. The workshop will help people who are interested in conducting a preference study but lack the tools to identify an appropriate preference-elicitation method(s) that will satisfy their information needs. 

DESCRIPTION:

Current focus on patient-centred care and interest in broader evaluations of the value of health interventions has increased practitioners’ interest in preference studies in recent times. A variety of preference-elicitation methods are available to health outcomes researchers to quantify the views of stakeholders in health care systems, but differences between these methods are not always apparent and conventions on their appropriate use are not definitive. Given the various contexts in which preference information can be used, it is vital for researchers to improve their understanding of the methods at their disposal, and the potential trade-offs faced when they commit to any particular method. We will describe some widely used preference-elicitation methods, including multi-criteria decision analysis, discrete-choice experiments and health-state utilities. Additionally, we will present examples of applications of each method and summarize the contexts under which each method has been used most successfully. Better understanding of the methods and their applications can help health outcomes researchers choose the most appropriate tool for their specific needs. Attendees will participate in a structured discussion about the appropriateness of various methods given specific research questions. They will be asked to state their own views on the type of the preference-elicitation method they would use in several examples of real-world applications. 

5:00 PM - 6:00 PM
WORKSHOPS - SESSION III
Health Policy Development Using Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon A, B, F (Level 5)

W15: THE ISPOR MCDA TASK FORCE: HOW BEST TO USE IT IN HEALTH CARE DECISION MAKING

Discussion Leaders:

Maarten J. IJzerman, PhD, Professor & Chair, Department of Health Technology & Services Research, University of Twente and MIRA Institute for Biomedical Technology & Technical Medicine, Enschede, The Netherlands

Nancy Devlin, PhD, Director of Rearch, Office of Health Economics, London, UK

Praveen Thokala, PhD, Research Fellow, School of Health and Related Research, University of Sheffield, Sheffield, UK

Kevin Marsh, PhD, Senior Director, Modelling and Simulation, Evidera, London, UK

PURPOSE:

The ISPOR Multi-Criteria Decision Analysis (MCDA) Emerging Good Practices Task Force will present recommendations for using MCDA to support and inform health care decisions.

DESCRIPTION:

There is growing recognition that MCDA methods have a role in improving health care decisions. Recent reviews have demonstrated an increase in the number of MCDAs published in health care, and regulators and payers are employing MCDA to support their decision making. This increased use of MCDA has drawn attention to several challenges. First, which of the many MCDA approaches – direct rating, swing weights, AHP, DCE – is the most appropriate for different health decisions? Second, how should these approaches be implemented? Third, can MCDA meet the specific needs of decision makers, such Health Technology Assessment (HTA) agencies’ consideration of opportunity cost, or the time-limits on clinician-patient consultations? Given the inadequate treatment of these issues in the existing literature, the ISPOR MCDA Taskforce was established to generate best practice recommendations. This workshop is designed to share the preliminary recommendations of the expert group assembled by the Task Force, including: 1) A typology designed to help distinguish MCDA techniques. 2) A summary of how the decision problem, the nature of decision makers preferences, the data demands of different techniques, the cognitive demands on participants, combine to determine the appropriate technique. 3) A best practice recommendations checklist for implementing MCDA, including how to: define the decision problem; generate criteria lists, elicit score and weights; measure performance, calculate aggregate scores; deal with uncertainty; and report findings. Recommendations for how MCDA can be employed for HTA, benefit-risk assessment, and clinician-patient decision.  The audience will learn key lessons to support the application of MCDA across the range of health care decisions, and will be invited to inform the development of Task Force recommendations by commenting on the preliminary output and sharing their experiences using MCDA.

Use of Real World Data

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon H (Level 5)

W16: MAXIMIZING THE UTILITY OF REAL WORLD EVIDENCE: INTEGRATION OF STRUCTURED EMR DATA, UNSTRUCTURED EMR DATA, AND BILLING DATA FOR ECONOMICS AND OUTCOMES RESEARCH IN ONCOLOGY

Discussion Leaders:

Mark S Walker, PhD, Vice President of Scientific Affairs, Vector Oncology, Memphis, TN, USA

Kathy L. Schulman, MS, Managing Director, Outcomes Research Solutions, Inc., Bolton, MA, USA

Arliene Ravelo, MPH, Associate Director, US Medical Affairs, Genentech, Inc., South San Francisco, TN, USA

Kim Saverno, PhD, RPh, Director of Pharmacoeconomics, Vector Oncology, Memphis, TN, USA

PURPOSE:

Access to detailed clinical data is driving new, innovative methodologies to enhance the value of EMR data in bringing clarity to important health outcome and pharmacoeconomic issues.  This session will describe synergies created by augmenting patient level structured EMR data with unstructured EMR data and practice-level billing data to conduct health economics and outcomes research in oncology.  Methodological considerations for use of each data source in oncology research will be featured.  Advantages and limitations relative to use of administrative data sources will also be discussed.

DESCRIPTION:

Approaches to optimizing use of real-world EMR data vary according to the aims of the research.  Dr. Walker will address the availability of clinical data in administrative vs. EMR sources, the completeness of structured EMR data, and types of clinical data generally available only by accessing unstructured EMR data.  Uses of unstructured data for complete case review, selective augmentation of structured EMR data, and algorithm development will be discussed.  Ms. Schulman will discuss selection criteria for data sources and optimizing ROI for outcomes research.  Ms. Ravelo will describe the use of clinical EMR data as a strategy to evaluate real-world practice patterns and treatment effectiveness.  Finally, Dr. Saverno will discuss the utility of EMRs and billing data as a basis for health resource and cost analyses, including limitations of these data sources, and the advantages that access to the unstructured EMR data bring to such analyses.


5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon K-L (Level 5)

W17: INFORMATICS AND INTEROPERABILITY: SPEAKING THE SAME LANGUAGE

Discussion Leaders:

Scott D Nelson, PharmD, MS, Post­doctoral Fellow, Medical Informatics, Department of Veterans Affairs, Salt Lake City, UT, USA

Olivier Bodenreider, MD, PhD, Branch Chief, Cognitive Science Branch, U.S. National Library of Medicine, Bethesda, MD, USA

Richard Boyce, PhD, Assistant Professor, Biomedical Informatics, University of Pittsburgh, Pittsburgh, PA, USA

Daniel C. Malone, PhD, RPh, FAMCP, Professor of Pharmacy, College of Pharmacy and Associate Professor, Mel & Enid Zuckerman College, University of Arizona, Tucson, AZ, USA bio

PURPOSE:

This workshop will provide a basic introduction to health information exchange, medication standard terminologies, and how to apply these concepts in pharmacy research.

DESCRIPTION:

What is health information exchange (HIE) and how can we bridge the 'data islands' to improve the delivery and evaluation of health services? Health care data currently exist locked up in many different islands making data integration, exchange, and aggregation difficult. Bridges between these 'data islands' are needed, but how do we get everyone to speak or understand the same language? HIE and interoperability require standard data exchange formats and terminologies to promote understanding across systems. This workshop will provide an introduction to health information exchange, standardized data exchange formats and pharmacy terminologies, the pharmacy "Rosetta stone" RxNorm, and application of these topics, such as using RxNorm to identify medications of interest for health outcomes studies, especially when the products are off-patent and made by numerous manufacturers. We will also discuss the Observational Health Data Sciences and Informatics (OHDSI, ohdsi.org) collaborative research network and indicatives. OHDSI is bringing together standardized data to develop a global knowledge base of all available electronic information for all drugs and health outcomes of interest pertinent to drug safety. Participants will be encouraged to share their experiences and views, and will have opportunity to provide feedback and question the panel.

Clinical Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon I (Level 5)

W18: SURVIVING THE SUBMISSION: BEST PRACTICES IN OVERALL SURVIVAL EXTRAPOLATION IN ONCOLOGY

Discussion Leaders:

Sanatan Shreay, PhD, Associate Director, Gilead Sciences, Foster City, CA, USA

Andrew Briggs, DPhil, MSc, William R. Lindsay Professor of Health Economics, Health Economics & Health Technology Assessment, Institute of Health & Wellbeing, University of Glasgow, Glasgow, UK

Rachel Beckerman, PhD, Principal, Value Demonstration, CBPartners, New York, NY, USA

Petros Pechlivanoglu, PhD, Health Economist, Toronto Health Economics and Technology Assessment Collaborative, Toronto, ON, Canada

PURPOSE:

The purpose of this workshop is to understand the limitations manufacturers are faced with when needing to extrapolate overall survival over a lifetime horizon in oncology indications, review the different survival extrapolation approaches that can be taken, and assess HTA bodies’ considerations when evaluating submissions based on such data.

DESCRIPTION:

Within the oncology space, payers typically rely on the ‘gold standard’ clinical endpoint of median overall survival (OS) to assess a new agent’s clinical value. However, if a new product is extremely effective, median OS may not be reached over the course of study follow-up.  Even if median OS has been reached, up to 50% of events may have yet to occur, requiring some form of survival extrapolation when building comparative effectiveness or cost-utility models over a lifetime horizon. This necessity for extrapolation can increase uncertainty in decision-making from the payer perspective. Different methodologies exist – both parametric and non-parametric – to extrapolate OS data, but each is associated with its own limitations. This workshop will first provide an overview of why such data limitations are inherent in many oncology trials. Next, we will assess the utility of potential survival analysis approaches in oncology, particularly in challenging situations, such as when median OS has not been reached or when the data have been captured from a single arm trial.   Finally, an analysis of recent case studies will elucidate whether payers in key HTA markets view the submission of immature OS data as a significant issue, as well as how different survival extrapolation approaches have been considered. Participants will come away from the workshop with a sense of best practices in modelling overall survival to meet payer specifications.

Economic Outcomes Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon C, D, E (Level 5)

W19: MODELING IN ONCOLOGY: THE TAMING OF THE SHREWS?

Discussion Leaders:

Noemi Muszbek, MSc, Senior Research Scientist, Evidera, London, UK

Sorrel Wolowacz, PhD, Head, European Health Economics, RTI Health Solutions, Manchester, UK

Agnes Benedict, MSc, Executive Director, Evidera, Budapest, Hungary

PURPOSE:

In oncology escalating costs, changing treatment paradigm with newer treatment modalities (e.g. immune therapy and antibody-drug conjugates), extensive off-label use, and evolving biomarker research highlight the limitations of current gold standard approaches in modeling cost-effectiveness (e.g. use of progression-free survival in model structure, cohort models, applicability of standard data requirement and indirect comparisons). This workshop aims to discuss these challenges, the current research and potential solutions. It is also planned to be the initial step in assessing and discussing the scope of a future special interest group for oncology modeling and to indicate priority topics of discussion. 

DESCRIPTION:

The workshop will begin with a brief questionnaire (using an app) to be completed by the audience to elicit oncology modeling issues encountered by the audience. The discussion leaders will introduce the challenges faced in evaluating oncology drugs. The implications of these challenges for the current ‘gold standard’ modeling methods will then be reviewed.  Increasing understanding of patient heterogeneity, individualized treatments and novel trial design for adaptive licensing need to be translated into different modeling approaches and model structures, and suggest the consideration of lifecycle and disease models. The specific issues of data requirements (e.g. oncology specific utilities, lack of data in palliative care, issues with the surrogacy and measurement of PFS, methods for extrapolation, implications of correlation between overall survival and PFS) will be listed and selected ones discussed in detail with examples. Audience answers to the questionnaire will be analyzed during the presentations; the audience will then be invited to participate in an extended open discussion of oncology specific issues in cost-effectiveness modeling, focusing on the issues highlighted by the survey, with problems encountered that would require further discussion and research.


5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon G (Level 5)

W20: INCORPORATING SOCIAL VALUES INTO COST-EFFECTIVENESS ANALYSIS

Discussion Leaders:

Christopher McCabe, MSc, PhD, Professor, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada

Mike Paulden, MA, MSc, Senior Research Associate, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada

James F. O'Mahony, PhD, Post-Doctoral Researcher, Department of Health Policy & Management, Trinity College Dublin, Dublin, Ireland

PURPOSE:

The purpose of this workshop is to provide attendees with a clear understanding of how social values ought to be incorporated within cost-effectiveness analysis. The workshop will be interactive, with attendees asked to consider what social values should be incorporated within cost-effectiveness analysis and to discuss the potential consequences of various approaches to doing so. The workshop will consider how the conventional model of the cost-effectiveness threshold may be updated to account for conflicting social values held by different decision makers.

DESCRIPTION:

The first component of the workshop will provide an overview of cost-effectiveness analysis and introduce attendees to the conventional model of the cost-effectiveness threshold. Arguments for and against broadening the consideration of cost-effectiveness analysis from health outcomes to a more complex measure of social value will be presented. Attendees will be asked to participate in a discussion about what social values ought to be incorporated into cost-effectiveness analysis. The second workshop component will explore the potential consequences of integrating a more complex measure of social value within cost-effectiveness analysis, using examples drawn from the earlier discussion. Attendees will be asked to critique a recent proposal by the UK’s National Institute for Health and Care Excellence (NICE) that attempted to integrate social values arguments into cost-effectiveness analysis by adjusting the cost-effectiveness threshold. In the final workshop component, attendees will be introduced to a more complex model of the cost-effectiveness threshold that addresses a number of limitations with the conventional model. This updated model includes explicit consideration of social values and a recognition that different decision makers may hold conflicting social value judgments. The workshop will conclude with a discussion of the implications of these and other considerations for the use of cost-effectiveness analysis in real-world practice.

Patient-Reported Outcomes & Patient Preference Research

5:00 PM - 6:00 PM
Room: Grand Ballroom, Salon J (Level 5)

W21: STATISTICAL METHODS USED FOR THE ASSESSMENT OF NON-REDUNDANCY AMONG CLINICAL TRIAL ENDPOINTS

Discussion Leaders:

Elizabeth D. Bacci, PhD, Senior Research Associate, Outcomes Research, Evidera, Seattle, WA, USA

Randall H. Bender, PhD, Senior Psychometric Statistician, Outcomes Research, Evidera, Bethesda, MD, USA

Joseph C Cappelleri, PhD, MPH, MS, Senior Director, Pfizer Inc., Groton, CT, USA

Kathleen W. Wyrwich, PhD, Executive Director, Center of Excellence, Outcomes Research, Evidera, Bethesda, MD, USA

PURPOSE:

This workshop will explore the issue of redundancy and non-redundancy among clinical trial endpoints. Endpoint redundancy occurs when two or more clinical outcome assessments measure the same concept. A closely related issue is the question of redundancy of treatment effects. That is, are treatment effects on conceptually different endpoints distinct and non-overlapping? An introduction to this topic, including regulatory perspectives, will be presented, followed by a discussion of several statistical techniques that can be used assess for non-redundancy among clinical trial endpoints.

DESCRIPTION:

Study endpoints can display redundancy when information obtained from one endpoint duplicates or provides overlapping or modest unique benefits beyond what is captured with other endpoints. However, it may be important to determine if a potentially new endpoint is measuring something similar as assessed by a currently used and accepted endpoint or if a treatment is associated with or affects multiple outcomes, independent of the effect on a primary outcome of interest. This session demonstrates specific techniques for the assessment of redundancy or non-redundancy among clinical endpoints. In the first session, Dr. Bacci will provide an in-depth introduction to this topic and regulatory perspectives on the importance of the assessment of non-redundancy. In the second session, Dr. Cappelleri will discuss the problem of empirical redundancy and illustrate methodological advances from the field of organizational research that are also applicable and useful to clinical trial endpoints. In the third session, Dr. Bender will demonstrate the use of a structural equation modeling approach to estimating simultaneous equations for the assessment of non-redundancy of treatment effects using real-world data, including design considerations that can help bolster the statistical argument.  Dr. Wyrwich will moderate the session and conclude the formal segment of the workshop with a summary incorporating themes of all three presentations before commencing additional discussion with the audience.

Wednesday, May 20, 2015
1:45 PM - 2:45 PM
WORKSHOPS - SESSION IV
Health Policy Development Using Outcomes Research

1:45 PM - 2:45 PM
Room: Grand Ballroom, Salon H (Level 5)

W22: HOW TO DESIGN AN ANALYTIC STRATEGY FOR EVIDENCE GENERATION FOR DECISION MAKERS

Discussion Leaders:

Sean D. Sullivan, PhD, MSc, RPh, Professor & Dean, Pharmaceutical Outcomes Research and Policy Program, University of Washington, Seattle, WA, USA

Omar Dabbous, MD, MPH, Head of Value Evidence Analytics, GlaxoSmithKline, King of Prussia, PA, USA

Lou Garrison, PhD, Professor, Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy, University of Washington, Seattle, WA, USA bio

Rafael Alfonso-Cristancho, MD, PhD, MSc, Director, Value Evidence Analytics, Research and Development, GlaxoSmithKline (GSK), King of Prussia, PA, USA

PURPOSE:

This workshop will demonstrate how to anticipate and address controversial study design and analysis issues to promote transparency and credibility with decision makers. We will highlight the differences in the evidence requirements for health care payers and describe the key types of designs and study features that must be considered to meet the health care payer evidence needs. Participants will be able to discuss the criteria for successful engagements of key stakeholder for the development of clinical studies for payers.

DESCRIPTION:

While randomized controlled clinical trials have long been viewed as the ‘gold standard’ of evidence, this approach alone may not provide sufficient information from which health care payers can make decisions about resource allocation. Pragmatic Clinical trials, retrospective and observational studies in all patient populations, comparing different therapeutic approaches within routine clinical care, may give data that more closely reflects the real-life. Early engagement with the stakeholders, so as to anticipate and address controversial issues, as well as communication with stakeholders throughout the process, to explain negative decisions and offer the chance to dispute them, can lead to success. Finally, analytical strategy, methodological rigor and transparency, through the involvement of well-respected clinical and nonclinical researchers is required. In the first session, Sean Sullivan, will describe the current challenges for the development of payer evidence, focusing on their needs for decision making and the gaps from the evidence provided to them. On the second session, Omar Dabbous, will present the basis to develop an analytic framework to improve payer evidence generation and the methods/approaches to evaluate evidence credibility. Finally, Lou Garrison, on the third session, will focus on how to achieve success and overcome barriers of implementation showing specific examples. Rafael Alfonso will moderate the session and provide a discussion of all three presenters, with audience participation at the end.

Use of Real World Data

1:45 PM - 2:45 PM
Room: Grand Ballroom, Salon J (Level 5)

W23: INTEGRATED LONGITUDINAL DATA: HOW DYNAMIC DATA COLLECTION CAN BIAS ESTIMATORS AND POSSIBLE SOLUTIONS

Discussion Leaders:

Henry J. Henk, PhD, Vice President & Principal Consultant, Optum, Eden Prarie, MN, USA

William H Olson, PhD, Leader, Research and Analysis Strategy, Janssen Scientific Affairs, LLC, Titusville, NJ, USA

PURPOSE:

Administrative claims databases provide a comprehensive and longitudinal record of heath care encounters and drug dispensing but lack clinical detail. Merging clinical data from electronic medical records with claims data to create “integrated” data can overcome this shortcoming. Integrated data can be used to more precisely identify the relevant study population and may allow one to better control for confounding when administrative claims data are used to conduct observational comparative effectiveness research. A missing data problem arises when longitudinal clinical data is collected dynamically. The most common approach in dealing with this issue is to ignore it and treat the non-missing data as if it were randomly selected from all patients regardless of their history. However, analyzing such data in this way can lead to seriously biased estimates of parameters of interest due to selection bias (i.e., bias that results from selection on common effects) unless the missing data is missing completely at random. The purpose of this workshop is to present methodologies that can mitigate such bias.

DESCRIPTION:

This workshop will demonstrate how, and under what circumstances, missing longitudinal data can lead to biased results if the analysis ignores the missing data mechanism and provide an introductory review of methodologies to assess and reduce the magnitude of such bias. Particular focus will be placed on alerting researchers to how common missing data processes can significantly bias results. The workshop will be interactive in nature that will encourage audience participation through a working example based on claims data from a large US payer linked to clinical data from electronic health records. By the end of the workshop, it expected that the participants will be fairly comfortable with the concept of selection bias, its relationship to missing data resulting from linking multiple datasets, and statistical methods that may offer a solution.

Clinical Outcomes Research

1:45 PM - 2:45 PM
Room: Grand Ballroom, Salon I (Level 5)

W24: A PRACTICAL APPROACH TO UNDERSTAND THE CONCEPTS AND METHODS USED TO ASSESS HETEROGENEITY AND INCONSISTENCY IN NETWORK META-ANALYSES

Discussion Leaders:

Varun Ektare, BPharm, MPH, Senior Analyst, Pharmerit International, Bethesda, MD, USA

Dipen Patel, PhD, Director, HEOR, Pharmerit International, Bethesda, MD, USA

Berhanu Alemayehu, PhD, Director, Outcomes Research, CORE, Primary Care Merck, Gaithersburg, MD, USA

Sonya J Snedecor, PhD, Executive Director, Health Economics, Pharmerit International, Bethesda, MD, USA

PURPOSE:

This workshop will present clear and simplified descriptions to facilitate an understanding of similarity, homogeneity and consistency including a) definitions, sources, and impact on network meta-analysis (NMA) results and b) detection and handling methods using sample exercises.

DESCRIPTION:

NMAs are commonly used to generate comparative effectiveness evidence among a landscape of treatments. Similarity, homogeneity, and consistency are the key assumptions of NMA methodology which should always be examined and reported. The assumption of similarity requires all trials included in a network to be comparable in terms of key factors that can be potential treatment effect modifiers. There are no formal tests for similarity, but it can be assessed using qualitative methods, which will be discussed in this workshop. Heterogeneity is present when outcomes from different trials vary more than expected from random variation. Heterogeneity can be a consequence of variations in patient characteristics, outcome definitions or measurements among different clinical trials. Standard qualitative and quantitative methods exist to detect heterogeneity. Methods used to manage heterogeneity when detected, such as random-effects modeling, sub-group analyses, and meta-regression will be discussed. Consistency within an NMA refers to reasonable agreement between the results of direct and indirect evidence within a network of clinical trials. Methods for testing inconsistency within a network such as the Bucher method, inconsistency model, and node-splitting method are often hard to understand and implement. These methods will be illustrated and discussed using examples. Finally, the implications of violating the NMA assumptions will be discussed as their presence may necessitate reconsideration of study inclusion criteria in order to ensure sufficient comparability among included trials. The workshop audience will be urged to participate in simple exercises to assess similarity, homogeneity and consistency in sample data/networks. Methods and examples will be illustrated utilizing Microsoft Excel and WinBUGS software.

Economic Outcomes Research

1:45 PM - 2:45 PM
Room: Grand Ballroom, Salon G (Level 5)

W25: MODELING TREATMENTS FOR RARE DISEASES: METHODOLOGIC CONSIDERATIONS, CHALLENGES & POTENTIAL SOLUTIONS

Discussion Leaders:

Michele Kohli, PhD, Director, Health Economics and Outcomes Research, Optum, Burlington, ON, Canada

Debbie L. Becker, MSc, Director, Health Economics and Outcomes Research, Optum, Burlington, ON, Canada

Milton C Weinstein, PhD, Henry J. Kaiser Professor of Health Policy and Management, Harvard T. H. Chan School of Public Health, Boston, MA, USA

Pablo Lapuerta, MD, Executive Vice President & Chief Medical Officer, Lexicon Pharmaceuticals, Princeton, NJ, USA

PURPOSE:

To discuss methodologic considerations, challenges, and solutions in designing cost-effectiveness and budget impact models for new treatments for rare diseases. The workshop will be of interest to health economics managers in the pharmaceutical industry and applied researchers in academic and consulting environments.

DESCRIPTION:

When assessing the cost-effectiveness or budget impact of a new treatment for a rare disease, it can be challenging to design and populate a model due to limited information on disease epidemiology, natural history, existing treatments, health care resource utilization, and health-state utility.  In addition, efficacy trials often have small sample sizes in important subgroups, which limits the ability to estimate efficacy of the new treatment. An ISPOR working group is in the process of cataloguing the challenges associated with the conduct of health technology assessment for rare diseases. In this workshop, discussion leaders will draw from their experiences in addressing some of these challenges in the context of budget impact and cost-effectiveness models of specific rare diseases. For example, in developing a model for a rare genetic disorder that causes intellectual disability, heterogeneity in disease progression and treatment outcomes made it difficult to define meaningful health states and to populate a Markov model with transition probabilities, costs, and utilities. In modeling a rare oncology disorder, challenges arose in estimating health care utilization for events that were defined endpoints in a clinical trial but not readily accessible from claims data. In developing a model for an inherited disorder that causes chronic pulmonary impairment, differences in outcomes by birth cohort and time- and age-varying risk factors presented challenges for modeling disease progression and survival. The workshop will also provide a forum for structured discussion on approaches for overcoming these challenges and workshop participants will be encouraged to share their thoughts and experiences.

Patient-Reported Outcomes & Patient Preference Research

1:45 PM - 2:45 PM
Room: Grand Ballroom, Salon K-L (Level 5)

W26: BACK TO THE FUTURE: LEARNING THE LESSONS IN VALUING EQ-5D-3L HEALTH STATES

Discussion Leaders:

Paul Kind, Professor, Academic Unit of Health Economics, University of Leeds, Leeds, UK

Roisin Adams, PhD, Deputy Head, National Centre for Pharmacoeconomics, Dublin, Ireland

Mônica Viegas, PhD, Professor, Economics Department, Center for Regional Development and Planning, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil

PURPOSE:

Over the past 20 years EQ-5D-3L has evolved as the most widely used generic measure of health-related quality of life with a notable role in economic evaluation. During that time a large number of national value sets have been generated, predominantly using methods based on a 1993 UK protocol. Despite its longevity this approach to valuing EQ-5D-3L health states has not been without its problems and changes to that standard protocol were proposed to deal with these. This workshop collates more recent experience and for the first time presents an overview of the current methods used for valuing EQ-5D-3L health states together with reports from national valuation studies that have benefited from the revised methodology.

DESCRIPTION:

The study of values for EQ-5D-3L provides an important and continuing portal to a significant field of research. Its relatively compact descriptive system with 243 health states supports experiment and innovation that is more difficult with more complex systems. The accumulated publications related to the valuation of EQ-5D-3L means that for the time being it remains the de facto standard in its field. The primary objective of this workshop is to provide access to practical experience of EQ-5D-3L valuation. Issues that have emerged cover both conceptual and practical elements of the EQ-5D-3L system including aspects of the descriptive system used to classify health states, the implementation of Time Trade-Off methods, survey design and data processing issues. A summary of these issues will be presented alongside potential options for remedial action. To exemplify the impact of such modifications, relevant details from two contemporary valuation studies using a modified protocol design in contrasting national settings (Ireland and Brazil) will be described. Workshop attendees will be encouraged to share their experiences of working with EQ-5D-3L either in eliciting social preference weights or applying them in economic evaluation.

3:00 PM - 4:00 PM
WORKSHOPS - SESSION V
Health Policy Development Using Outcomes Research

3:00 PM - 4:00 PM
Room: Grand Ballroom, Salon I (Level 5)

W27: WHEN BETTER THINGS HAPPEN TO A GOOD MODEL: A DEVELOPMENT OF THE DIFFERENCE IN DIFFERENCES (DD) MODEL INTO A DIFFERENCE IN DIFFERENCES IN DIFFERENCES IN DIFFERENCES (DDDD) MODEL WHEN ANALYZING THE EFFECT OF AN INTERVENTION

Discussion Leaders:

Hemant Phatak, PhD, Group Director, Global Health Economics & Outcomes Research, Bristol-Myers Squibb Company, Princeton, NJ, USA

Ya-Chen Tina Shih, PhD, Professor of Health Economics, & Section Chief, Section of Cancer Economics and Policy, Department of Health Services Research, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Yanru Qiao, MS, Data Analyst, Department of Clinical Pharmacy, The University of Tennessee College of Pharmacy, Memphis, TN, USA

Junling Wang, PhD, Associate Professor, Department of Clinical Pharmacy, The University of Tennessee College of Pharmacy, Memphis, TN, USA

PURPOSE:

This workshop aims to: (1) showcase the application of a difference-in-differences-in-differences-in-differences model (DDDD), developed as an extension to the difference-in-differences (DD) model, to analyze the effects of an intervention; (2) demonstrate how to program the DDDD model in Stata.

DESCRIPTION:

Difference-in-differences (DD) model is a quantitative analysis technique commonly used to examine the effect of an intervention in observational studies. Mimicking an experimental design, DD calculates changes over time in study outcomes for the intervention group and the control group, and attributes the differences in these changes between the two groups before and after the intervention as the effects of the intervention. DD model has becoming increasing popular because of the growing interests among the research community in conducting comparative effectiveness research using secondary databases. However, an even more powerful DDDD model has not been used much, partly because its programming using existing statistical software is daunting. For this workshop, Shih will first provide an overview of the DD model as well as its extended difference-in-difference-in-differences (DDD) model as well as their applications in the current literature. Next, Wang will showcase the development and application of a DDDD model in a recent study published by the research team. The study objective was to determine whether the implementation of Medicare Part D correlates to changes in racial and ethnic disparities among individuals ineligible for Medicare medication therapy management (MTM) services and MTM-eligible beneficiaries. Using DDDD model, this study found that there were greater racial and ethnic disparities among MTM-ineligible individuals than among MTM-eligible individuals even after Part D implementation, suggesting that MTM eligibility criteria may perpetuate existing disparities. Third, Qiao will demonstrate how the programming process was achieved in Stata for the DDDD model. Sufficient time will be allocated throughout the workshop for audience to share their related experiences.


3:00 PM - 4:00 PM
Room: Grand Ballroom, Salon K-L (Level 5)

W28: HORIZON SCANNING – IDENTIFYING AND ESTIMATING FUTURE IMPACT OF EMERGING INNOVATIONS ON U.S. HEALTH CARE

Discussion Leaders:

Elise Berliner, PhD, Director, Health Technology Assessment, Center for Evidence and Practice Improvement, Agency for Healthcare Research and Quality, Rockville, MD, USA

Karen Schoelles, MD, SM, Project Director, AHRQ Healthcare Horizon Scanning System and Director, ECRI Institute Evidence-based Practice Center, Technology Assessment, ECRI Institute, Plymouth Meeting, PA, USA

Marcus Lynch, PhD, Senior Horizon Scanning Analyst, Technology Assessment, ECRI Institute, Plymouth Meeting, PA, USA

Jalpa A. Doshi, PhD, Associate Professor of Medicine & Director, Economic Evaluations Unit, Center for Evidence-based Practice and Director, Value-based Insurance Design Initiatives, Center for Health Incentives and Behavioral Economics, General Internal Medicine, Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, PA, USA

PURPOSE:

This workshop will introduce attendees to horizon scanning as a method for identifying and monitoring innovations that have high potential to disrupt clinical care, the health care system (i.e., infrastructure, care processes and settings, staffing), patient outcomes and costs. The AHRQ Healthcare Horizon Scanning System, in operation since December 2010, encompasses a broad range of interventions, including drugs, devices, diagnostics, procedures, behavioral health and care delivery innovations that claim to address a substantial unmet need for AHRQ's priority conditions. The audience will apply the described methods for assessment of selected topics.

DESCRIPTION:

The methods used in the AHRQ Healthcare Horizon Scanning System were developed based on a review of the literature, discussions with leaders of other national horizon scanning systems, an expert panel meeting, and refinements made as the team gained experience implementing the system. Dr. Berliner will discuss AHRQ’s interest in horizon scanning and use of horizon scanning outside the United States. Dr. Schoelles will describe the inclusion criteria and the evaluation of topics for potential impact on health outcomes, the delivery system and disparities. She will describe current use of the System outputs by payers and other groups. Dr. Lynch will present examples of drugs, devices and implants evaluated in a pilot project to add basic cost analyses to Horizon Scanning assessments. Dr. Doshi will discuss the challenges in conducting “rapid review” cost analyses of emerging technologies. Participants will engage in an exercise of assessing interventions for potential impact on unmet health needs and on the health care system. We will solicit feedback on the work presented and solicit ideas for future uses of the data in the System.

Use of Real World Data

3:00 PM - 4:00 PM
Room: Grand Ballroom, Salon G (Level 5)

W29: STRATEGIES FOR ASSESSING THE PATIENT-LEVEL ECONOMIC IMPACT OF CANCER DIAGNOSIS

Discussion Leaders:

Veena Shankaran, MD, MS, Assistant Professor, Division of Medical Oncology, University of Washington, Seattle, WA, USA

Amy Davidoff, PhD, MS, Senior Research Scientist, Cancer Prevention and Control, Yale School of Public Health, New Haven, CT, USA

Donatus Ekwueme, PhD, Senior Health Economist, Division of Cancer Prevention and Control, Centers for Disease Control and Prevention, Chamblee, GA, USA

PURPOSE:

The purpose of this workshop is twofold:  first, to review the strengths and limitations of available and emerging data sources for assessing cancer survivors' economic burden and second, to engage the audience in discussions around strategies at the patient, provider, or policy level to mitigate the economic burden of cancer treatment.

DESCRIPTION:

The impact of cancer diagnosis on patients’ and caregivers’ financial well-being (employment and productivity, out-of-pocket spending, assets and savings) is an emerging concern in an era of rapidly rising cancer treatment costs and growth in high cost-sharing health plans in the United States.  In this workshop, we will provide a conceptual framework for financial toxicity, and a detailed discussion of important sources of data to evaluate the economic burden of cancer treatment:  (i.) Medical Expenditures Panel Survey (MEPS) Experiences with Cancer Survivorship Supplement, (ii.) National Health Interview Survey, (iii.) Medicare Current Beneficiary Survey, (iii.) Medicare Part D claims data, (iv.) linkage of SEER to U.S. Bankruptcy Court records (Washington State), (v.) commercial credit reporting agencies (e.g. TransUnion, Equifax).  We will discuss previous and ongoing work using these data sources and linkages as well as the strengths and limitations of each.  We will end the session with an interactive discussion of potential strategies to mitigate the economic burden of cancer treatment.

Economic Outcomes Research

3:00 PM - 4:00 PM
Room: Grand Ballroom, Salon J (Level 5)

W30: WARRANTING BUDGET PREDICTABILITY THROUGH MANAGED ENTRY AGREEMENTS AND INSURANCE-BASED MECHANISMS

Discussion Leaders:

Olivier Ethgen, MSc, PhD, Adjunct Professor, Health Economics, University of Liege, Liege, Belgium

Augustin Terlinden, MSc, Health Economist, BLUE ANTIDOTE, Tervuren, Belgium

PURPOSE:

Stretched health care budget are increasingly inciting payers to contract on market access conditions with the manufacturers, the so-called managed entry agreements (MEAs). Many of the MEAs contracted so far are based on financial considerations and primarily intend to provide the payer with some degree of budget predictability. This workshop will describe the methodological foundations and the strategic questions encountered in the design of financial-based MEAs solution for innovative products. In this view, the workshop will explore insurance-based mechanisms with the intervention of a third-party insurer in the enacting and operationalization of MEAs.

DESCRIPTION:

As a short introduction, financial-based MEAs will be illustrated with few of the MEAs already in place. Then, we will show how the financial risks both parties (the payer and the manufacturer) are exposed to when they enter into MEAs can be modelled and quantified. The workshop will detail how, and under which conditions, these risks can be outsourced to an insurance company. Conditions for insuring these risks embodied in a MEA will be analyzed. Pricing such risks is conditional to owning sufficient historical data (which is not the case for genuine innovation) and having a good understanding of the risks dynamics. In addition, the insured parties (the payer and the manufacturer) need to be incentivized against all form of moral hazard. All those methodological steps will be exemplified with a hypothetical oncology product. Finally, directions for future researches will be proposed and opinion from the audience on the future applications of MEAs will be sought through an interactive Q&A session.

Patient-Reported Outcomes & Patient Preference Research

3:00 PM - 4:00 PM
Room: Grand Ballroom, Salon H (Level 5)

W31: DEVELOPMENTS AND COMMUNICATION SINCE THE ISSUANCE OF FDA’S GUIDANCE TO INDUSTRY ON PATIENT-REPORTED OUTCOMES: THEN AND NOW

Discussion Leaders:

Brooke Witherspoon, BA, Associate Director, Research & Operations, Endpoint Outcomes, Boston, MA, USA

Somali Misra Burgess, PhD, CEO & Research Director, Strategic Outcomes Services, Mission Viejo, CA, USA

Paivi Miskala, MSPH, PhD, Founder, ProCon Global, LLC, Rockville, MA, USA

PURPOSE:

Since the release of the US Food and Drug Administration’s (FDA) Guidance to Industry on Patient-Reported Outcomes (PRO) measures in December 2009, FDA and experts in the field have provided additional information to help developers navigate patient-focused outcome measurement in clinical trials and the qualification of clinical outcome assessments (COAs), including PROs. The purpose of this workshop is to provide health outcomes researchers, who are not necessarily PRO experts, a basic understanding of the FDA PRO guidance document with a focus on new information that has been disseminated by both regulators and experts within the PRO community since the guidance was issued.

DESCRIPTION:

The discussion leaders, including a former FDA Study Endpoints and Labeling Development (SEALD) reviewer and representatives from industry and consulting, will present a brief overview of the principles included in the FDA’s PRO Guidance, including the types of evidence that should be considered when evaluating if a PRO measure is well-defined and reliable in a specific context of use to support medical product labeling claims. They will also present additional published information from experts in the field that will help to fill some of the gaps in the FDA's PRO Guidance document. Subsequently, the focus will be on how some of this information can be applied to specific situations. FDA’s Drug Development Qualification Program and the related guidance document will also be discussed. By outlining the FDA’s PRO Guidance recommendations and identifying gaps that have since been filled or addressed, workshop attendees will be better equipped to navigate the current landscape of PRO development and validation as it relates to FDA regulatory submissions in supporting medical product labeling claims. Participants will be encouraged to share their experiences and engage in discussion regarding how to best address common PRO scenarios given developments since the issuance of FDA’s PRO Guidance.

Contact ISPOR @ info@ispor.org | View Legal Disclaimer
© 2018 International Society for Pharmacoeconomics and Outcomes Research.
All rights reserved under International and Pan-American Copyright Conventions.