|
|
|
|
Workshops |
|
SESSION I - Monday 12 November 2001 University of the
Sciences, Philadelphia, PA, USA PARTICIPANTS WHO WOULD BENEFIT: This session is intended for individuals who are interested in learning more about creating and using interactive tools and data collection instruments on the Internet. Pharmacoeconomic and outcome
tools on the Internet can allow data to be collected and results disseminated
from different policy perspectives: 1) patient, 2) provider, 3) hospital, 4)
managed care organization, 5) government, and 6) society. Interactive
internet models can facilitate the assessment of health care interventions or
services from different quantitative perspectives such as 1) cost of illness,
2) cost effectiveness, and 3) quality of life monitoring. Programming
examples will cover strengths and weaknesses of html, xml, mysql, javascript,
vbscript, cgi, perl, php, asp, wml, and java. Pros and cons of web authoring
software will be discussed. Internet and wireless examples will be presented
that can facilitate shared decision-making and incorporate patient
preferences into treatment pathways at the bedside. Web applications can
assist in maintaining outcome databases, facilitating disease management, and
generating continuous feedback for providers and patients. Internet pages can
now be accessed in patients’ homes, practitioners’ offices and anywhere with
the new handheld computers and web phones. Questions that should be asked in
evaluating Internet applications include: how much of the information is
built on clinical trials, has the web site been peer reviewed, has the
material been rigorously field-tested, are questionnaires valid and reliable,
is the information well referenced? WW2
NUMBER NEEDED TO
TREAT (NNT): IS IT A USEFUL BENCHMARK FOR THE EFFICIENCY OF THERAPIES? Caro Research Institute,
Concord, MA, USA PARTICIPANTS WHO WOULD BENEFIT: Researchers interested in learning more about the NNT concept which has been referred to as the “currency of Evidence Based Medicine”, and its potential role in economic evaluations. The NNT — the reciprocal of
the absolute risk reduction — was introduced in 1988 as an “easily understood
yardstick to describe the harm as well as the benefit of therapy and other
clinical maneuvers”. Although presented as a tool to facilitate clinical
decision-making, its use in public decision-making was already hinted at in
this first publication. The NNT has since been portrayed as a first
approximation to more complex measures such as cost-utility ratios, which are
considered the ultimate goal of Evidence Based Medicine, and NNT-based league
tables have been presented. In this workshop we will review the origin and
properties of the NNT and evaluate whether it indeed meets the criteria of a
useful decision-making tool: comparative, easy to understand and calculate,
standardized. Although the first three criteria appear to be met, we will
demonstrate how the modifications proposed over the years to address the
shortcomings of the original NNT in terms of standardization (i.e., time
horizon, reference risk, comparator and outcome considered) have greatly
increased its complexity while major issues of standardization remain.
Throughout this critical evaluation phase, direct input from the workshop
participants will be solicited. We will conclude by presenting our approach
to calculate an “adjusted” NNT for public health use. WW3
BEYOND
COST-EFFECTIVENESS: THE STRATEGIC VALUE OF PRODUCTIVITY-RELATED OUTCOMES IN
CLINICAL TRIALS, BURDEN OF ILLNESS STUDIES, AND POST-MARKETING RESEARCH 1The
MEDSTAT Group, Inc, Ann Arbor, MI, USA; 2The MEDSTAT Group, Inc,
Washington, DC, USA; 3The MEDSTAT Group, Inc, Cambridge, MA, USA
PARTICIPANTS WHO WOULD BENEFIT: Sponsors of clinical trials, burden of illness studies, and post-marketing research; Employers who offer medical care benefits and other benefit programs; Outcomes researchers; Drug formulary developers; Regulators. Until recently, productivity
at work has been overlooked as a measure of value in clinical trials, burden
of illness studies, and post-marketing research. However, recent advances in
data development capabilities and market pressures to differentiate the
plethora of drugs either under development or on the market have motivated a
concern for a broader set of relevant outcome measures. In addition, concerns
over the ability to document the total impact of drug therapy have emerged as
important issues, as pharmaceutical developers, health care providers,
employers, formulary developers, and policy makers strive to fully understand
their own or society’s return on investment in drug therapy. Recent evidence
suggests that, among leading employers in the U.S., productivity-related
metrics account for more than half (53%) of the cost of employer benefit
programs. In Europe, rigorously developed and supportable claims about the
impact of drugs under development or already marketed can influence the drug
approval process. In the U.S. and elsewhere, such evidence may influence
initial and subsequent formulary decisions, and the appropriate use of
alternative drug therapies. This workshop will illustrate how to 1) identify
productivity-related metrics, 2) collect and process data on these metrics,
and 3) use these data in sophisticated research studies designed for clinical
trials, to document the full burden of illness, or to legitimately support
post-marketing claims of the effectiveness of drug therapy. It will be shown
that better decisions can be made by incorporating a more complete set of
relevant outcome measures. WW4
THE ROLE OF
EXTRAPOLATION IN ECONOMIC EVALUATION: APPLICATION OF TECHNIQUES TO MEET
EMERGING REQUIREMENTS 1University
of York, York, UK; 2City University London, London, UK; 3Innovus
Research (UK) Ltd, Amersham, UK PARTICIPANTS WHO WOULD BENEFIT: Researchers concerned with the generation of economic evidence to support the acceptance of new technologies. Economic evaluations ideally
require data on final outcomes measured over a reasonable period. This
sentiment is reflected in many of the official guidelines for undertaking
studies. However, at the time of product launch, data are available only for
intermediate outcomes, (e.g. percentage change in cholesterol), or final
outcomes are measured over only a short time period (e.g. months rather than
years). How can those concerned with generating evidence to support early
acceptance of new technologies reconcile this apparent conflict between
expectations and reality? The workshop will explore the calculation of life
years gained through analysis of non-parametric actuarial survival curves,
using two worked examples. One example will use data from a published
epidemiological cohort study of smokers to show how life expectancy varies
according to the age at which smokers successfully quit. The second example,
drawn from clinical trial data, estimates average life expectancy using the
"area under the curve" approach. Methods for extrapolation beyond the end of
the trial period will also be discussed. The workshop will conclude with a
discussion of the guidance about extrapolation offered in various official
guidelines for economic evaluation. WW5
RESPONSIVENESS OF
PATIENT REPORTED OUTCOMES IN CLINICAL TRIALS: ISSUES, ASSESSMENT STRATEGY,
AND INTERPRETATION Mapi Values, Lyon, France OBJECTIVES: The objective of the workshop is to present and discuss methods to assess and interpret the responsiveness of PRO measures and its links with the clinical validity. The discussion will be put on the specific context of the analysis of a PRO measure in a Clinical Trial for a claim. PARTICIPANTS WHO WOULD BENEFIT: Methodologists, Clinicians and Researchers designing and/or analysing PRO measures in Clinical Trials. Recent meetings with Health Authorities have shown the value of Patients Reported Outcomes (PRO) as end points in clinical trials (CTs). Nevertheless, application of the proper analysis strategy is essential to reach the requested level of evidence. Understanding the relationship between clinical outcomes and PRO in the treatment of a particular disease is central in this strategy. We will first review and present the concept of clinical validity and responsiveness and their relationships. Key references articles will support discussion. We will then consider concrete issues when moving from theory to practice, with examples from recent work performed on specific and generic instruments in different therapeutic areas (Gastroenterology, Central Nervous System, Cardio-Vascular, Cancer). Successes and failures will be analysed and discussed. The focus will be put on possible reasons for poor responses or biases such as true placebo effect, care effect, Hawthorne effect, non-linear response. The validity of clinical measures to determine responsiveness will be discussed as well. Based on our experience we will propose a set of analyses to be conducted on the CT data before unblinding. The potential interest of different statistical techniques such as structural modeling or multiple factor analysis will be discussed. In conclusion, a conceptual model describing the impact of the disease and its treatment on the PRO is a prerequisite for the interpretation of the changes in scores. This will increase the credibility and validity of the results when requesting for a claim.
WW6
HOW TO DEAL WITH
MULTIPLE OUTCOME MEASURES IN CLINICAL STUDIES? 1University
Medical Centre Nijmegen, Nijmegen, Netherlands; 2NV Organon, Oss,
Netherlands PARTICIPANTS WHO WOULD BENEFIT: All researchers who are involved in the analysis, interpretation and presentation of outcomes of clinical studies as well as reviewers of clinical evidence. Often there is not one exclusive outcome measure in a clinical trial that is adequate to assess the global effect of a medical intervention. We are frequently confronted with several distinct outcome measures (survival, clinical indicators, health status, adverse events), each providing specific and equally important information. As an example can be given the combination of two distinct outcome measures, quantity and quality of life, into a single measure based on the quality-adjusted life year (QALY). For the computation of QALYs, individual preference scores expressing health status are used. An alternative approach is the Q-TWIST method, distinguishing a limited number of distinct clinical health states. Each of these health states is assigned a preference score, and the proportion of patients in each health state at a specific time period is estimated, based on empirical survival curves. Moreover, depending on the study design and the disease of interest, we have to deal with dropout, missing values, repeated measurements, confounding factors and mortality. This workshop will provide an overview of the current analytic techniques to approach multiple outcome measures in clinical trials and their effectiveness. Based on the findings, we will demonstrate proposed analytic strategies to combine health status, survival, dropout and repeated measurements. In this workshop we will illustrate these analytical strategies by data from clinical studies in rheumatoid arthritis, psoriasis and heroin addiction. An interactive part of this workshop will consist of a comparison between two different analytic strategies on cost-effectiveness analysis.
WW7
WHICH DATA CAN WE
GET? – THE EFFECT OF ELECTRONIC PATIENT RECORDS AND SECURITY REQUIREMENTS ON
DATA AVAILABILITY deCODE genetics Inc, Reykjavik, Iceland OBJECTIVES: The purpose of this workshop is to further the understanding of data requirements, security, and privacy issues in health care databases. PARTICIPANTS WHO WOULD BENEFIT: All pharmacoeconomics and outcomes researchers who are interested in new developments in secondary health care data availability and wish to learn specifically about data requirements, security and privacy issues in health care databases. Health care researchers have increased access to clinical data as electronic patient records (EPRs) are implemented. The possibilities for linkage of diverse databases enables the study of complex research questions utilizing new methods. With the advent of stricter privacy rules globally researchers are increasingly faced with data security and patient privacy issues. In the first part of the workshop leaders will provide ideas of data requirements for EPRs. Sophisticated classification and coding schemas as an integral part of an EPR are a prerequisite for the successful development of clinical databases as they enable large scale collection of standardized clinical data suitable for observational research. The participants will be asked to formulate ideas of data requirements of pharmacoeconomic and outcomes studies and how EPRs need to be developed to meet these needs. In the second part, the methods employed in building a clinical health care database in Iceland are shown. Issues of database security and patient privacy will be discussed as examples of new developments in the field of health care databases.
SESSION II - Monday 12 November 2001 WW8
THE AMCP FORMULARY
SUBMISSION FORMAT: IMPLICATIONS FOR PHARMACOECONOMIC RESEARCH PROGRAMS 1RTI
Health Solutions, Research Triangle Park, NC, USA; 2PAREXEL
International, Baltimore, MD, USA PARTICIPANTS WHO WOULD BENEFIT: 1. Health economics personnel participating in clinical development teams 2. Clinical and marketing personnel participating in clinical development teams 3. European health economics academicians interested in keeping abreast of global developments in the use of health economics for decision making. The goal of this workshop is to review the recently promulgated guidelines for formulary submissions by the Academy of Managed Care Pharmacy in the US and to determine the implications for pharmacoeconomic research programs by companies planning to submit new products to US health plans for reimbursement. Although the US has a fragmented health care system, it is likely that the AMCP guidelines for formulary submission will be broadly adopted by US health care plans. The guidelines include requirements for pharmacoeconomic evaluations of new drugs. Because of the size of the US market, it is important for company personnel to understand these new guidelines so that they can ensure that their clinical development program will result in data that will satisfy these requirements. The review of the guidelines will indicate what type of economic evaluations are desired (e.g. cost-consequence, cost-effectiveness, cost-utility), what data sources are preferable (e.g. random controlled clinical trials, meta-analyses, extrapolation from multiple data sources), and how different population subgroups should be treated. Examples of changes in typical pharmacoeconomic research programs will be developed in an interactive session with the workshop participants using specific disease examples such as schizophrenia and community acquired pneumonia.
BUILDING A
PHARMACOECONOMIC STRATEGY FOR A PRODUCT 1MEDTAP
International, London, UK; 2MEDTAP International, Jisp,
Netherlands PARTICIPANTS WHO WOULD BENEFIT: Those wishing to understand more about the process of defining and commissioning pharmacoeconomic activities by the pharmaceutical industry. Pharmacoeconomic and other health outcomes activities undertaken by pharmaceutical companies are usually devised to fit into a strategy for development of each product. The selection and programming of activities depends on the clinical development plan, the competing products and therapies on or near to market, the potential “value arguments” which may be supportable for the product and the target audiences for this information. At an early stage an environmental scan may give information on the burden of disease and potential market size, and early “feasibility” modeling may help to explore the economic implications in different projected levels of effectiveness. Activities undertaken alongside clinical trials will give information used in submissions for reimbursement approval and marketing of the product immediately after launch. Modeling may again be required to project health status and resource use where direct measurement has not been possible. Post-launch studies may be used to confirm the value of the product in a real world setting, or to identify patient groups where cost-effectiveness is particularly favourable. Following a brief introduction, participants in this workshop will work together in small groups to plan part of a pharmacoeconomic strategy for a (fictitious) product being developed to treat cerebrovascular disease. The focus will be on activities that take place during the later stages of drug development. Tasks to be undertaken during the workshop will include identification of ‘value arguments’ based on a briefing to be provided, choosing countries and audiences for economic messages, justifying data collection activities and the scheduling of these activities.
EXPERTS BOARD IN
HEALTH ECONOMICS AND OUTCOMES RESEARCH: INSTRUCTIONS FOR USE Association CRééS, Fontenay-sous-Bois, France OBJECTIVES: The purpose of this workshop is to draw an “instructions for use” of experts board in health economics and outcomes research. From the definition of an experts board to the final impact expected, this workshop is the place to share experiences on this topic between actors, sponsors and the present audience. PARTICIPANTS WHO WOULD BENEFIT: All health economics or outcomes researchers who wish to ensure to their projects a good credibility with regard to health authorities and/or scientific or medical community. Constitution of an experts board in HEOR has to share common principles between members of the board, sponsor of the board and future audience of the board. Whatever the territory where the board plans to work on (local, national or worldwide), the following items need to be clearly defined to meet objectives initially targeted. Some of them are proposed thereafter: Definition: experts board is a way to gather expertise in HEOR and in medical science. This is the place to orient the HEOR research, to get consensus on responses on key issues of the project and to gain credibility on the research. Composition: a balance needs to be found between HEOR experts, public health experts, clinicians and with regards to the origin of the experts (university, hospital, institutions, health authorities, CRO, sponsors, patients representative). Type of Board: Advisory Board, Executive Board, Experts panel or a mix of each? Scope of work: complete or partial development of HEOR program (HEOR protocol, analysis plan development, cost of illness studies, economic model construction, epidemiological data analysis, survival data analysis, customisation, report development, publication plan). Knowledge of disease program on management, cost, environment of the disease. Development and/or interpretation of QoL scale, utility scale, patient/prescriber satisfaction questionnaire. Communication on these various topics by participation to publication, symposium, press conference. Organisation: level of implication of each members (adviser, doer) and decision making process. Moreover, way of functioning, assets and issues raised by experts board will be presented and discussed during this workshop: The Association CRééS (Cercle de Réflexion évaluation économique de la Santé: an association of french speaking pharmaco-economists) leading the debate.
THE USE OF
INTEGER/LINEAR PROGRAMMING MODELS IN DECISION MAKING 1RTI
Health Solutions, RTP, NC, USA; 2Eli Lilly, Indianapolis, IN, USA
PARTICIPANTS WHO WOULD BENEFIT: Health outcomes researchers, health care decision-makers, and those responsible for resource allocation and budgeting would benefit from this workshop. Many types of decision
analytic techniques are used in outcomes research to assist in making health
care resource allocations. ILP/LP models, like decision trees and Markov
processes, are used to make these allocations. Specifically, ILP/LP
complement decision trees and Markov processes in that their outputs can be
used to obtain an optimal resource allocation that considers all feasible
alternatives under all applicable constraints. Proof is in that these models
are used extensively in other industries to achieve objectives such as
minimizing costs and maximizing resource allocation effectiveness. This
workshop presents the use of integer and linear mathematical programs and
their application in health care decision making for health authorities. We
describe their properties, program formulation, and how ILP/LP results can be
applied to various health care decisions. Examples to evaluate and compare
health care and non-health care interventions are presented. The workshop
includes an interactive demonstration of solving these models and
interpreting results.
WW12
BEYOND THE QALY:
USING CONJOINT ANALYSIS TO QUANTIFY HEALTH-CARE PREFERENCES AND EQUIVALENT
MONETARY VALUES 1Research Triangle Institute, Research Triangle Park, NC, USA; 2Research Triangle Institute, Manchester, UK OBJECTIVES: Participants will learn how to use conjoint analysis to elicit and analyze patient and public preferences for health-care treatments and how these estimates relate to QALY measures. PARTICIPANTS WHO WOULD BENEFIT: Pharmacoeconomic researchers who are interested in alternative techniques for measuring the benefits of health care treatments and products. Conjoint analysis often is employed in marketing research, but increasingly also is being used in outcomes research. This workshop will explore the advantages of conjoint analysis (CA) approaches in evaluating health outcomes relative to the standard QALY approach. CA is a stated-preference technique that presents subjects with a series of scenarios or bundles of treatment characteristics. Experimental-design considerations guide scenario construction to ensure statistical independence among characteristics. Subjects then rank, rate or choose alternatives to indicate their preferences. Unlike other preference elicitation methods, subjects’ responses can be checked for stability, consistency, and other utility-theoretic requirements. When correctly analyzed, these data provide estimates of utility weights useful for cost-utility analysis and other purposes. If cost is included as a treatment characteristic, estimated utility weights also can be scaled into monetary measures of the value to patients. Value to patients is simply the change in a subject’s income that yields the same increase or decrease in satisfaction. Selecting appropriate ranges of treatment characteristics in European health-care systems will be discussed. Participants also will act as subjects in a simple CA exercise to understand how the method works. Participants will see examples of applications that illustrate principles of survey design, show how to cope with subjects’ cognitive limits, demonstrate statistical analysis, and compare utility and value estimates for pharmaceuticals, health outcomes, patient satisfaction, longevity, and health delivery mechanisms.
VISUAL ANALOG
SCALES: DO THEY HAVE A ROLE IN THE MEASUREMENT OF PREFERENCES FOR HEALTH
STATES? 1McMaster
University, Innovus Research Inc, Burlington, ON, Canada; 2Innovus
Research Inc, Burlington, ON, Canada PARTICIPANTS WHO WOULD BENEFIT: Those who must design, execute, commission or interpret studies that involve direct utility measurement. The VAS has long been used
as an instrument for the measurement of health state utilities. It is an
attractive approach because it is simple, quick to administer, and lends
itself to self-completion. Recent findings, however, have raised serious
questions about its validity in all settings. Should its use now be
abandoned? Should its use continue unabated? Or should it be used in more
limited ways? We will argue for the latter. In the workshop we will review
the two theoretical foundations for the VAS – one from decision
sciences/economics and one from psychology/psychophysics. We will review the
three major measurement problems with the VAS – anchor specificity, context
bias and end-aversion bias. Methods of bias correction will be presented.
Power curves to convert VAS scores to utilities will be discussed. Two
appropriate roles for VAS will be suggested – VAS as an introductory task
before standard gamble or time trade-off, and VAS adjusted by a
study-specific power curve with or without debiasing. An example of a VAS in
the latter role will be presented -- a two-stage, self-completion VAS in a
urology application. Participants will be asked to contribute to the workshop
by sharing their own experiences and ideas, and by participating in the
discussion.
WW14
MULTICULTURAL AND
MULTILINGUAL QUALITY OF LIFE INSTRUMENT DEVELOPMENT: A COMPARISON OF
SEQUENTIAL, PARALLEL AND SIMULTANEOUS APPROACHES AND THEIR IMPACT ON
TRANSLATION AND ADAPTATION 1Evanston
Northwestern Healthcare, Evanston, IL, USA; 2MEDTAP International,
Bethesda, MD, USA PARTICIPANTS WHO WOULD BENEFIT: Pharmacoeconomic researchers who are interested in multicultural and multilingual instrument development and translation issues. Multinational clinical trials have become a necessity in today's global economy to expedite testing and approval of new pharmaceutical treatments. As quality of life (QOL) instruments are used in more clinical trials and research worldwide, the question of how such instruments were initially developed becomes increasingly important. Three major questionnaire development approaches have evolved over the past decades. In the sequential approach, a scale is developed in one culture and later adapted to other cultures through translation. In the parallel approach, a scale is developed in multiple cultures with the goal of deriving a common set of items across countries. Like the parallel approach, the simultaneous approach begins with multi-national input in the earliest stages of development but also allows item content and question format to differ across language versions according to the cultural and normative experiences of each country. No single approach is suitable for all situations; therefore the decision of which development methodology to choose will depend on the purpose of the study and other practical considerations. Examples of questionnaires developed through each of these approaches will be presented, along with the impact on the translation process and on instrument psychometric and equivalence issues. The interactive portion of this workshop will focus on a discussion of the advantages and disadvantages of each approach and practical guidelines for conducting instrument development, translation, adaptation and measurement across cultures.
SESSION III - Tuesday 13 November 2001 WW15
IMPLEMENTING THE UK
NICE CONCEPT IN ANOTHER EUROPEAN COUNTRY: NICE TRY OR "SALADE NICOISE"? 1Ghent
University, HEDM, MEISE, Belgium; 2Ghent
University Hospital, GENT, Belgium; 3BIGE, Zaventem, Belgium
PARTICIPANTS WHO WOULD BENEFIT: Decision makers on price and reimbursement, policy makers, researchers conducting studies for price and reimbursement purposes. Health economic evaluations in decisions on price and reimbursement of pharmaceuticals will formally be applied in Belgium as from January 2002. Belgium has a policy of positive lists whereby companies must submit a file for decisions on reimbursement. The preparation of the new policy was inspired by the NICE initiative in the UK, whereby it was attempted to “learn from the experiences with NICE”. Several issues occurred: for instance, many of the expected evidence is predictive and often based on modeling techniques, while the true value of new drugs can only be assessed in the market. Hence, the decision maker only wants to adopt new drugs once there is more evidence, while the supplier only is able to provide real-life evidence once the drug is adopted. An additional problem is the size of the target population: although, according to priority setting theory, this size should not play a role in priority decisions, the budget impact of decisions is obviously affected. Risk sharing and re-examination after 2-3 years marketing of new drugs were proposed for novel and improved decision making. Next, it was discussed whether a threshold for defining cost-effective care must be applied, and whether QALYs were to be applied. On the procedural level, the issue of independent assessment of the delivered evidence was raised. The workshop overviews and discusses the arguments that were used in these and other debates. The participants are encouraged to suggest alternative scientific and pragmatic arguments.
ASSESSING
PHARMACOECONOMIC AND QUALITY OF LIFE OUTCOMES IN GLOBAL PATIENT REGISTRIES 1Quintiles
Late Phase, Levallois-Perret, France; 2Quintiles Late Phase, San
Francisco, CA, USA; 3Quintiles Late Phase, Boston, MA, USA
PARTICIPANTS WHO WOULD BENEFIT: All health outcomes and pharmacoeconomic researchers interested in learning about design, methodological, interpretation, and logistical issues associated with conducting global patient registries. Pharmacoeconomic and QOL data are routinely collected in prospective, observational, global outcomes registries. However, designing global registries to measure health and pharmacoeconomic outcomes cross culturally poses unique methodological and logistical challenges to researchers. Standardization of designs and data collection processes are made difficult due to differences in technologies, treatment norms, available resources, cultural expectations and characteristics of the health care system across countries and continents. Design, methodological and logistical challenges faced by researchers in implementing global registries will be reviewed in this workshop. Specifically, cross cultural design issues, methods for handling cultural variation in perception of disease and treatment-related side effects, statistical issues regarding pooling of data, and interpretation of QOL outcomes will be addressed. The measurement and collection of pharmacoeconomic outcomes from multiple countries also will be discussed. The guided interactive part of this workshop will focus on designing a hypothetical global registry that includes QOL and pharmacoeconomic outcomes. Guidelines for implementing global registries will be presented. Additionally, real world examples of multinational disease registries will be reviewed to illustrate potential solutions to global design, measurement, analytic, interpretation, and logistical challenges.
COST OF ILLNESS
STUDIES: DO WE STILL NEED THEM? 1MEDTAP
International, London, UK; 2MEDTAP International, Inc, London, UK;
3MEDTAP International, Jisp, Netherlands
PARTICIPANTS WHO WOULD BENEFIT: Pharmacoeconomic researchers who undertake or commission COI studies. COI studies broadly fall into two types: ‘prevalence-based’ - where the cost of managing a condition over a given time period is measured for all those in a given population with a condition, irrespective of when their treatment started; or ‘incidence-based’ – where the cost of newly treated patients is measured over a given time-period from treatment initiation. COIs frequently include estimates of foregone productivity for those suffering from a condition or undergoing a treatment, as well as direct costs of medical and social care. COIs have been shown to be of limited value in health care decision making. By definition they are non-comparative: they do not present information about the incremental costs of introducing a new therapy, the associated impact on budgets and changes in health outcomes. In this workshop participants will explore the reasons why these studies remain popular in the pharmacoeconomic and clinical literature. Reasons may include: input into drug development portfolio selection, prioritisation of indications for licensing, ascertainment of treatment patterns and cost elements for use in future model-base economic studies; meeting regulatory and local requirements for burden of disease/care information; creation of ‘noise’ in the market for the condition of interest. At the outset participants will be asked to complete a questionnaire asking about their experience and perceptions of the value of COI studies. Results from this questionnaire will be fed back and used as a starting point for a structured discussion. Participants will also be asked if they know of examples where COI analyses have made an important difference in health care decision-making.
FATIGUE ASSESSMENT
AS AN OUTCOME IN CANCER CLINICAL TRIALS On behalf of the EORTC Quality of Life Group, Brussels, Belgium OBJECTIVES: The purpose of this workshop is to provide an overview of fatigue as a concept of Quality of Life Assessment and explain how it can be integrated into cancer clinical trials. PARTICIPANTS WHO WOULD BENEFIT: All researchers with an interest in fatigue working in the field of cancer outcome assessment. Fatigue is undoubtedly one of the commonest symptoms experienced by cancer patients, presenting a challenging and complex phenomenon for measurement. The effect of fatigue on patients’ quality of life (QOL) is considerably influencing most domains. Repeatedly, studies show that fatigue correlates directly with overall QOL, greater fatigue often leading to poorer outcomes. Lately a considerable number of studies stressed the complex problem faced by cancer patients experiencing fatigue either during or after treatment. Furthermore, fatigue can continue for many years post treatment and for those patients with progressive disease, fatigue can lead to a loss of overall QOL that is extremely limiting. Over the past decade, researchers and clinicians have become increasingly aware of the negative influence of fatigue on QOL, devoting considerable time to developing fatigue assessment tools. However, whatever the reason for the increased interest, researchers are still asking basic questions "How is fatigue assessed?” “Is there not a gold standard to measure or compare fatigue?” “What are the key dimensions of fatigue?” Perhaps researchers should assess fatigue briefly and simply by focusing on a single question such as "Do you feel tired?" If researchers and clinicians do choose measures to examine fatigue, should different measures be used with different cancer patients at different times in the disease cycle? When information is collected from patients about fatigue, how should researchers interpret this? For example, what level of score on any given fatigue measure would suggest clinical intervention? What are the key properties in a fatigue measurement that researchers should look for? We will explain these issues and encourage audience involvement in debating fatigue assessment issues.
MEASURES AND
METHODS FOR ASSESSING PATIENT-REPORTED OUTCOMES RTI Health Solutions,
Research Triangle Park, NC, USA
PARTICIPANTS WHO WOULD BENEFIT: Anyone who wants to learn more about developing strategies for the measurement of patient-reported outcomes and the survey methods used to collect these data could benefit from this workshop. Health outcomes researchers charged with collecting patient-reported outcomes, either within or independent of clinical trials, are likely to benefit most. This workshop will begin with an overview of what patient-reported outcomes are and how they are measured. To illustrate the range of instruments available for assessing just one such outcome, we will briefly describe the three major categories of quality of life measures, including disease- or condition-specific questionnaires, generic health profiles and generic preference-based measures. Guidelines will then be offered to assist in developing a measurement strategy that is appropriate for the goals of the research. Issues for consideration include the type of instrument to be used, the psychometric properties of the various options, the cognitive complexity of the items and formatting (if self-administered), the availability of translations, flexibility regarding the mode of administration, and previous research involving the population and therapeutic area under study. The discussion will then shift to techniques for collecting survey data with special attention to the reduction of missing data that results from item and survey nonresponse. The benefits and potential drawbacks to collecting information from patients through mail, telephone, computer-assisted, and Web-based surveys will be described as well as the use of panel maintenance and tracking methods to reduce the number of patients lost to follow-up. Recommendations will then be offered to assist attendees in choosing the optimal mode of administration, depending on the context and ultimate purpose of the survey.
ITEM RESPONSE
THEORY AND ITS APPLICATIONS TO HEALTH OUTCOMES MEASUREMENT Evanston Northwestern
Healthcare and Northwestern University, Evanston, IL, USA
PARTICIPANTS WHO WOULD BENEFIT: All quality of life and outcomes researchers who are interested in IRT models and applications. Although item response theory (IRT) models have been developed and widely used in educational and psychological testing for several decades, their use in healthcare settings has just mushroomed. Theoretical sound IRT models coupled with available software make it possible for outcomes researchers to develop and refine outcomes assessment instruments for use in clinical trials and research. This workshop will provide an overview of IRT models and how they can be appropriately applied to health outcomes assessment. Specifically, this workshop will discuss the following topics: 1) dichotomous versus polytomous models; 2) unidimensional versus multidimensional models; 3) scale construction; 4) exploratory vs. confirmatory item-level factor analysis; 5) instrument equating; 6) differential item functioning; 7) item banking and computerized adaptive testing; and 8) software availability. Examples using empirical data and annotated computer output will be provided and discussed. Guidelines to the selection of models and software will also be provided.
SESSION IV- Tuesday 13 November 2001 WW21
MAXIMIZING THE
SCIENTIFIC AND STRATEGIC VALUE OF PATIENT REGISTRIES Ovation Research Group,
Highland Park, IL, USA PARTICIPANTS WHO WOULD BENEFIT: Health economists seeking to better understand patient registries and their advantages relative to other outcomes research initiatives. The objectives underlying industry-sponsored patient registries — typically large, observational, prospective research programs involving data collected from physicians and, often, directly from patients — vary from safety surveillance to market penetration. Accordingly, these programs are designed and implemented by staff with a broad range of responsibilities, from marketing and product management to clinical research. The most successful patient registries achieve strategic value for the sponsor while serving as a vehicle for scientific exploration into the clinical, economic, and humanistic impact of a drug, device, or disease state. As such, outcomes researchers should — but seldom do — play a central role in the development of registries. This workshop will begin with an interactive discussion focusing on the audience’s knowledge of and experience with patient registries. The workshop will then present brief case studies of several existing registry programs and their respective histories, achievements, and successes. A four-step approach for designing and executing patient registries will be presented, including techniques employing state-of-the-art technologies for data collection and communications. The workshop will conclude with suggestions for bringing coordination of patient registries under the authority of health economics and outcomes research departments.
HEALTHCARE DATABASE
AS A SERVICE CONTRACTING TOOL – BETWEEN INFORMATION DESERT AND INFORMATION
FLOOD: EXPERIENCE OF POLISH HEALTH CARE SYSTEM REFORM Silesian Regional Sickness
Fund, Katowice, Poland PARTICIPANTS WHO WOULD BENEFIT: all who use or plan to use healthcare database (in research or in “practice”), to learn strengths and weaknesses of the databases. The health care system reform that has been introduced in Poland is an attempt to base health service provision on market competition between service providers. The pivotal element of the new system is an institution of independent sickness fund that is responsible for purchasing health service. The Silesian Regional Sickness Fund (insures 5 million members) established computer network linking all the contracted hospitals. The information from the database is used to “watch” the medical services. Hospitals report information on patient age, gender, place of living, performed procedures, disease (plus co-morbidity), length of hospitalisation, places of treatment, result of treatment, etc. Logical and medical coherence of the data is checked. But coherent information not necessarily means true one. There is danger that road leading from information desert to information ocean may end up in virtual reality. The data flows smoothly, the quality of the information is checked, its coherence controlled but results of the analyses are sometimes unexpected - enormous mortality rate differences in regions or cardiology hospitals (from 6% in one university hospital to 14% in the other), patients “transferred” between hospitals (artificial costs creation), strange stroke patient population structure, etc. We discover facts or we are too information greedy and create “own” reality?
INCONGRUITY BETWEEN
THE USE OF HEALTH ECONOMIC ANALYIS AND BUDGETARY IMPACT ANALYSIS 1Medtap
International, London, England; 2Medtap International, Jisp,
Netherlands PARTICIPANTS WHO WOULD BENEFIT: Those responsible for pricing and reimbursement of new drugs at local and global offices of pharmaceutical industry, and decision makers using economic and financial data. Carrying out studies
according to a pharmacoeconomic guidelines should promote the provision of
reliable, reproducible and verifiable insight into the therapeutic value of a
drug, the costs involved in using the drug and the possible savings that can
be made on other drugs and/or therapies. For policy-making, apart from the
cost-effectiveness of a drug, the consequences for the macro-costs are also
important: for this purpose a financial analysis is used. A financial
analysis must provide insight into all the financial consequences of the
introduction of a new drug. If the financial analysis is going to play a
fundamental role in reimbursement decisions comparable with pharmacoeconomic
analysis, it is vital that the methods and procedures used in such
evaluations are as carefully scrutinised and refined. Finally the weight of
the relationship between the pharmacoeconomic and budgetary impact data in
the decision making process needs to be defined. The workshop will
concentrate on the UK and The Netherlands, but is also relevant for most of
the other West-European countries. We will start with a description of the
differences between health economic and budgetary impact analyses:
(perspective, time horizon, outcomes), which may lead to conflicting
situations. In this workshop following a brief introduction, participants
will discuss the differences between health economic and budgetary impact
analyses. Tasks to be undertaken will be to identify the potential conflict
between both types of data, and explore possible solutions in the light of
participants' experiences with submissions in the absence of detailed budget
impact analysis guidance. WW24
USING PROPENSITY
SCORES TO ADJUST FOR TREATMENT SELECTION BIAS: REVIEW OF METHODS PLUS AN
EXTENSION FOR USE IN STUDIES WITH THREE OR MORE TREATMENT GROUPS RTI Health Solutions,
Research Triangle Park, NC, USA
PARTICIPANTS WHO WOULD BENEFIT: Pharmacoeconomic researchers who are responsible for designing non-randomized studies of healthcare interventions and healthcare decisionmakers who are responsible for evaluating and interpreting the results of those studies. In non-randomized studies,
differences in patient characteristics can influence treatment selection
which, in turn, can lead to biased estimates of treatment effects. Propensity
scores are one technique to address treatment selection bias. A propensity
score is the probability that a patient will be assigned to a particular
treatment or intervention. Typically, a logistic regression analysis is used
to estimate a patient’s probability for receiving Intervention A (versus
Intervention B) using observed pre-treatment characteristics. Propensity
scores may be used in at least three ways to reduce treatment selection bias:
(1) matching, (2) stratification; and (3) adjustment in regression analysis.
Most published applications of propensity scores to reduce treatment
selection bias are applied in studies comparing two interventions. Through
the use of multinomial logistic regression analysis, we extend the
two-treatment paradigm to include the evaluation of three interventions on
health outcomes. The guided interactive part of the workshop will use a case
study involving patients with schizophrenia or schizoaffective disorder.
Participants will be invited to brainstorm factors that might influence
selection of an atypical antipsychotic for the treatment of schizophrenia.
Participants will also be asked to critique an actual propensity score model
developed in this disease area.
WW25 PSU and The On-Line Guide to Quality-of-Life Assessment (OLGA), State College, PA, USA OBJECTIVES: This workshop develops basic skills for evaluating alternative health status and quality of life instruments to be used in assessing health outcomes and health-related quality of life in clinical trials and pharmacoeconomic studies. Key to the selection of relevant measures is the adoption of a systematic approach for evaluating questionnaires, rating scales and classification systems for use, singly or in combination, in different types of studies. We will present and discuss an approach that includes essential elements in the instrument selection process, including issues in specifying purpose of study and target population, identifying relevant domains to satisfy these study parameters, and evaluating measurement properties to assure meaningful results. PARTICIPANTS WHO WOULD BENEFIT: Pharmacoeconomic and health outcomes researchers who want to increase their understanding of the issues involved in selecting meaningful and responsive quality of life instruments for use in clinical trials and other evaluative studies Health outcomes and quality-of-life assessment is becoming increasingly important in the evaluation of pharmaceutical products, in terms of labeling claims and product promotion as well as in terms of formulary decisions. Each of these applications requires an assessment strategy that provides information relevant for decision-making. How is a successful assessment strategy developed? What criteria, including reliability, validity and responsiveness, should be used when evaluating instruments for use in a clinical trial or pharmacoeconomic study? This workshop will address these questions and introduce an analytic framework that participants can apply in their daily experience.
WW26
METHODOLOGICAL
ISSUES OF DESIGNING QUALITY OF LIFE IN CANCER CLINICAL TRIALS On behalf of the EORTC
Quality of Life Group, Brussels, Belgium
PARTICIPANTS WHO WOULD BENEFIT: All outcomes researchers who have a limited knowledge of developing cancer clinical trial protocols. Quality of life (QL) is an increasingly important outcome in clinical trials. In the future we expect QL to be regarded as a standard endpoint, both primary and secondary in many trials and the key to helping prove the value of treatments. Regulatory authorities are now seeing the value of QL in many therapeutic areas. It is clear that it will become over the coming years an important aspect of approval for new drugs as well as in labeling and promotional claims. However, while this may be the case, the field of QL assessment is fraught with many challenges, in particular relative to the design and collection of good quality data. Therefore, the following presentations highlight the challenges that face researchers when designing clinical trials and collecting data. The first presentation discusses the definitions of QL and provides a basic overview of how to design a robust QL study in a cancer clinical trial, based on developments at the EORTC. Emphasis will be placed on explaining key issues such as selecting measures, timing of assessment, dealing with missing data and reporting the results that one must consider when designing such studies within large Phase III clinical trials. The second presentation will focus on putting these methodological issues into practice and reporting on ongoing study protocols undertaken with Hodgkin’s disease patients on active treatment. After training, we will expect the audience to be able to contribute to developing a Phase III clinical trial and have some ideas on how to design the QL component of a cancer clinical trial. |
|
Fourth Annual European Conference Main Page |
