MISSION STATEMENT
The mission of the Center for Pharmaceutical Outcomes Research is to contribute to the improvement of patient health outcomes, primarily those related to potential or actual use of pharmaceuticals, through methodology development, evaluative research, translation of research findings to practice, and education.
Current decisions in health care management are strongly affected by cost containment concerns. In order to make rational decisions concerning the allocation of scarce health care resources, however, one must develop and utilize a knowledge base that details variations in outcomes of health care delivery. The relationship between costs and outcomes provides the key to arriving at efficient solutions to the allocation of limited health care dollars. One must not only accurately measure such outcomes, but also attach relative values to them so that decision makers may fully understand the costs and consequences of alternative treatment or allocation decisions. Thus, outcomes research must ultimately be based on the value of the outcomes in order to be of maximal usefulness.
In the areas of costs, outcomes and value of the outcomes, significant gaps exist between our current measurement abilities and those needed to supply better information for aiding decision makers. Moreover, much current evaluative research fails to utilize the full range of methods already available. This implies a need for education and research in both theoretical and methodological directions as well as in more obvious practical applications such as specific treatment assessments.
The Center for Pharmaceutical Outcomes Research (CePOR) based in the UNC School of Pharmacy is focused on developing methodologies and disseminating knowledge about treatment- related health and economic outcomes. Epidemiology and economics are the two primary disciplines supporting this inquiry. Epidemiologic tools enable the measurement of health outcomes; economic tools assign values and estimate costs of achieving selected outcomes.
The Center for Pharmaceutical Outcomes Research (CePOR) will focus on those economic, health status and humanistic outcomes related to the use of pharmaceuticals and pharmaceutical services. However, as the field of evaluation is a comparative one, research will not exclude evaluations of non-pharmaceutical interventions where appropriate.
Specific objectives for CePOR include:
(1) To develop and evaluate new tools, instruments and methods in patient outcomes research.
(2) To evaluate the effect of treatment strategies on patient outcomes.
(3) To assess the relationship of pharmaceutical care and pharmacotherapy on patient outcomes.
(4) To further graduate education through involvement of professional, graduate, and post-graduate students in research projects to develop a cadre of scientists trained in pharmaceutical outcomes research.
(5) To disseminate information of a theoretical and practical nature to the health care community.
CePOR's staff supplement traditional pharmaceutical science expertise with advanced degrees in economics, epidemiology and business administration. CePOR's formal and informal contacts with individuals and departments in UNC's Medical and Public Health Schools, other disease- specific multidisciplinary centers (e.g. Program on Aging, AIDS Clinical Trials Group), Duke University Medical School and Center for Health Policy, Research and Education provide additional talent and expertise on an as-needed basis. The in-house expertise and these institutional links make CePOR uniquely poised to provide solutions to the most pressing questions in today's competitive health care environment.
CENTER FOR PHARMACEUTICAL OUTCOMES RESEARCH
Concentration of expertise in areas of pharmaceutical outcomes assessment (both health and economic) at UNC combined with the great need for research at both practical and theoretical levels has led to the establishment of a research and educational center for pharmaceutical outcomes research. CePOR will focus on health and economic outcomes relating to pharmaceutical interventions; however, as the field of evaluation is a comparative one, research will not exclude evaluations of non-pharmaceutical interventions where appropriate. Specific objectives include:
(1) To develop and evaluate new tools, instruments and methods in patient outcomes research.
(2) To evaluate the effect of treatment strategies on patient outcomes.
(3) To assess the relationship of pharmaceutical care and pharmacotherapy on patient outcomes.
(4) To further graduate education through involvement of professional, graduate, and post-graduate students in research projects to develop a cadre of scientists trained in pharmaceutical outcomes research.
(5) To disseminate information of a theoretical and practical nature to the health care community.
CePOR has accumulated and continues to foster development of talent (by additional faculty recruitment and faculty development grants) in disciplines necessary to the task set out above. Expertise in statistics, epidemiology, and economics is conspicuous among CePOR faculty and staff. Responding to the need for talent in meta-analysis or for adjusting for bias in study designs, CePOR staff are trained in areas of statistics, decision analysis, Markov and queuing models and general simulation modeling. On a study-by-study basis, medical expertise (physician and pharmacist) will be necessary to inform analysts as to the natural history of diseases, pharmacokinetics, dynamics, etc. CePOR has considerable existing expertise in these areas and has institutional links to other departments within UNC as well as to external organizations to facilitate additional interactions in these areas (See Appendix I).
Extensive experience in conducting single-center and multiple-center studies for major pharmaceutical sponsors, contract research organizations, foundations and the federal government has earned CePOR's faculty a reputation for accurate, timely and relevant research. CePOR staff have concentrated on Phase IV and Phase V studies, however they have also supported drug development activities by providing cost-effectiveness analyses in Phase IIb clinical trials. Research collaboration in Phase I and Phase II studies also takes place with the General Clinical Research Center, School of Medicine at UNC-Chapel Hill and at the Program in Drug Development at the School of Pharmacy . In addition to measuring traditional health outcomes, CePOR has experience in evaluative research and methods development in patient attitudinal studies. CePOR's faculty are knowledgeable about measurement techniques that discern patient satisfaction and quality of care as defined by patient-based criteria.
Graduate education is an important component of CePOR's activities. The increasing complexities of the drug development process, and its interaction with managed health care, pharmaceutical marketing, and the increasing statutory requirements of the federal government, demand experts in pharmaceutical outcomes research. CePOR provides research training for graduate students interested in careers in industry, government, managed care, contract research organizations, and academia as pharmacoeconomists, pharmacoepidemiologists and quality-of-life researchers. Such a cadre of researchers is needed to ensure continued positive improvements in pharmaceutical outcomes throughout the world. Training at the Master of Science and Ph.D. levels for research on pharmaceutical outcomes is provided by affiliated departments. In addition, visiting scholars and fellows complement CePOR's educational activities. Degree programs are established through the Division of Pharmacy Administration in the School of Pharmacy in collaboration with the Department of Health Policy and Administration in the UNC School of Public Health.
II. Elements of Pharmaceutical Outcomes Research
Recent developments in medical program evaluations address the two primary concerns of cost and quality of care. Political concerns about the rate of increase in cost of medical care led to an intense interest in evaluating procedure- or treatment-specific costs in order to curtail them. However, there is a great danger that uninformed attempts to cut costs may, while achieving their goals, significantly sacrifice quality of care. Thus, it is essential that quality of care ought not be relegated to second place in deference to pure cost containment issues, but rather be given at least equal status.
The theory and practice of measuring quality of care has evolved over time. The recent evaluative history of quality has centered on the assessment of patient outcomes, appealing to the last of the sequence of structure, process and outcome as originally proposed by Donabedian (1966). It has been recognized that evaluating only structure or process in health care delivery falls short of the ultimate measure of interest--patient outcomes. While structure and process are not irrelevant for evaluating health care systems and technologies, neither are they sufficient. Measuring patient outcomes is paramount.
The concern for quality of outcome can be viewed from two perspectives. The first is concerned solely with measuring the quality of an individual patient outcome. Traditional quality-of-life studies fit into this framework - developing better notions of medical outcome than the more traditional (and cruder) notions of mortality, for example. The second approach includes the first in a technological sense, but subsumes it under a different spirit -- one that is population-based rather than individual-based. The basic structure of outcomes research involves defining the appropriate study population, determining treatment allocations and treatment outcomes, and ultimately assessing health outcomes, as shown in Figure 1.
The development of this second approach reflects a broad notion of economics - the balancing of cost and outcome concerns to determine efficient allocations of health care resources for populations. This notion of systemic quality recognizes that any action consumes resources that are then unavailable for other purposes. If those resources are not used in their most productive ways (to produce the most "health") then there is an inefficiency in the system.* The balancing of costs and consequences is crucial to the economic concept. It implies that sole concern for either cost limitation or quality maximization is inappropriate. The way to maximize the health of a given population is to balance quality and cost considerations for any individual. Savings at the individual level may be used to provide health to other individuals in a needy population. Thus, this second notion implies a recognition that achieving the highest quality of outcome for individuals may only come at the cost of the sacrifice of quality of care for other individuals. The notion of quality in this second formulation is a systemic one concerning the quality of outcome of the entire delivery system; however, as its building blocks, it uses refined notions of patient outcome, preferences for various outcomes and the costs of achieving them.
We are faced with the challenge of creating a "better" environment for health care delivery. Parts of the system demand an individual, patient-oriented focus while other parts demand or impose a population-based perspective. Both require better measures of patient outcomes and implications (health and economic) of achieving such outcomes. In order to better inform health care decision- makers about both cost and quality implications of alternative medical interventions, researchers must develop more refined methods of measuring costs, outcomes and the value of such outcomes and implement them to assess specific interventions. Furthermore, there is a great need for translation of information about both methodological advances and specific intervention evaluations to researchers and practitioners.
III. Important Issues in Cost and Outcome Measurement
Many of those with limited experience in the field of technology assessment in general or outcomes research in particular oversimplify the research issues by ignoring quality issues or by ignoring important elements relating to costs. Traditional concerns with drug acquisition cost alone, for example, ignore the potentially important issues of quality of outcome and potential cost savings implications to budgets other than those relating specifically to drugs (e.g. monitoring costs, laboratory tests, hospital lengths of stay).
The proper notions of cost are very broad and not obvious in many cases. They need not always be monetary, nor need they be in areas within the strict purview of particular budgetary authorities. Table 1 indicates a subset of potentially important costs that might need to be considered for a fully developed costing of the implications of using particular treatments. Some of these costs would not be easy to estimate. Standard tools from fields of epidemiologic and economic investigation can be used to provide estimates of frequencies of various intermediate outcomes and their costs.
Table 2 indicates that treatment effects are not always obvious either. More importantly the value of any given effect is not obvious (and may, in fact, differ substantially across individuals). Various methods exist to measure valuations of outcomes ( Table 3 ). A variety of different analytical approaches to outcomes research combines the costs and value or consequences ( Table 4 ). At present there are still conflicts of both a theoretical and practical nature that must be addressed in order to advance this field of outcomes research.
Of perhaps greater importance to accurate cost and value of outcome measures are accurate outcome measures themselves. There are sometimes significant differences between results obtained from clinical trial studies and the results to be expected in the "real world" of clinical practice. For lack of better terms, this distinction has been captured by the terms "efficacy" and "effectiveness" ( Table 5 ). Efficacy refers to theoretical results based, for example, on controlled clinical trials (Phase III or earlier), whereas effectiveness refers to the results to be expected in actual medical practice (possibly approximated by Phase IV or V studies). Effectiveness is defined as a drug's performance under naturalistic conditions in which the drug is prescribed, dispensed and taken by the patient in an environment typical of everyday medical practice. Factors such as concurrent diseases, noncompliance and less stringent medical surveillance may imply differences between the drug's efficacy and its effectiveness.
Much of health economic outcomes assessment in pharmaceuticals has relied on the randomized controlled trial for its outcome measurements. Table 6 traces a simplified version of the evolution of thinking about economic outcomes assessment in pharmaceuticals. The emergence of the Randomized Controlled Trial (RCT) as the perceived gold standard in economic assessment of pharmaceuticals has derived largely from (1) its superiority over casual "marketing anecdotes" of superior performance, (2) its familiarity within the pharmaceutical industry, and (3) its availability at the time when such outcomes information is needed. However, numerous problems may be identified with using RCTs to make inferences to the real world of clinical practice ( Table 7 ).
While it has long been recognized that RCTs are biased in numerous ways, such biases have been difficult to address in noneconomic contexts. Eddy (1990) has alleged the traditional approach to assessing bias in any type of study has been to either reject the study by virtue of having determined that it is too biased to be worthy of attention, or to accept it because of "acceptable" levels of bias which ignores the fact that it is nonetheless biased ( Figure 2).
A more rational (though complex) approach would be to measure bias and adjust results for that bias. In order to accomplish such adjustment for bias, one must appeal to complex methods of meta-analysis, a technique fast becoming a "buzz-word" in technology assessment without adequate recognition of the sophisticated analytical and statistical tools necessary to adequately perform the required tasks. Meta-analysis, or a related technique, cross-design synthesis (GAO, 1992), can be used to adjust results for bias or to incorporate various pieces of evidence from trials or observational studies to obtain better estimates of outcomes ( Table 8 ).
In summary, outcomes research can be defined as those research activities that measure the efficacy of treatments, identify factors which may render drug therapy less than optimal, predict effectiveness of treatments, evaluate interventions designed to optimize drug therapy and analyze the economic implications of treatment alternatives. Outcomes research should include both intended and unintended effects on health outcomes, stemming from the underlying biological parameters, as shown in Figure 3. Efficacy determinations are often made in the premarketing phase and effectiveness determinations in the postmarketing phase in the drug development. Experimental methods, mainly the randomized clinical trial, provide the efficacy data in the premarketing phase of the drug; effectiveness data is generally obtained employing observational methods in the postmarketing phase of the drug or by synthesizing results from different types of studies. A sample of study types is provided in Table 9 .
IV. Modeling Approaches in Outcomes Research
The latest methodological developments in outcomes assessment imply a continuing emphasis on advanced modeling techniques from the field of operations research. Models are both relatively inexpensive and very adaptable tools for predicting implications of changing the use of medical technologies. They supply an additional alternative to more traditional methods of outcomes assessment. Models can be described as "thought experiments" where all relevant issues are laid out for everyone to see and judge stage-by-stage. Models are quite useful at any stage of evaluation. The essence of a good model is that it strips away enough analytical complexity from the real world to make analysis of complex issues practicable, but not so much as to make results overly simplistic. Thus, the model is an analytical simplification of the real world. One omits details in order to be better able to answer the question at hand.
Early in the drug development process, significant uncertainty about various treatment parameters may exist (safety profiles, cure rates, etc.). At such a stage, some useful outcomes information may still be generated through the judicious use of modeling. In particular, models can divulge to decision makers the identity of the most important decision making parameters -- those upon which results hinge and, therefore, for which precise estimates are important. For example, comparative ADR rates may not yet be known, but if the impact of such events is high in health or economic terms, precise estimates will be essential to outcomes research. If, on the other hand, such events are not costly, precise estimates may not be of much value. One may speculate how a new drug which alters the probability of a particular ADR development will alter the cost- effectiveness of a drug over that of some other probability change. Prior to knowing what the true change is, one may develop a picture of the outcome implications of various different probabilities. In effect, modeling, by identifying key parameters for which precise estimates are needed, provides a cost-effective mechanism by which to improve any subsequent analysis. It shows analysts where the most analytical bang may be purchased for a given buck. Once the probability is estimated in patients, the model can indicate how any remaining uncertainty about the probability (or any other parameter) influences the conclusions of the model. Models may sometimes be used to discover the lack of economic viability of compounds early in development prior to expensive clinical trials. Models may also be used to help arrive at a good estimate of treatment effects, ADR rates, etc. Standard epidemiologic tools are used to estimate parameters on a study-by-study basis, and meta-analysis can provide a vehicle for combining estimates from several studies to obtain more precise estimates on which models can be built.
Models in outcomes research that incorporate recent developments in epidemiologic and economic methods aid in policy analysis and formulary decision making. As policy makers scrutinize health care budgets to allocate limited public health funds, results from outcomes research can be applied to various populations for which state and federal governments provide coverage. These decisions should be based on the perceived value of medicines at the population level. Similar analysis is relevant to any entity (e.g. HMOs) involved in medical and pharmaceutical reimbursement. At the societal level, legislators can utilize information from outcomes research models to set policy that affects research funding, either directly or through tax incentive programs. Thus, modeling is important not only to clinical analysis and reimbursement issues, but to a wide variety of policy implications.
Executive Committee:
The Executive Committee provides oversight and guidance for day-to-day activities.
William H. Campbell, Ph.D. (Chair) Dean, UNC-CH School of Pharmacy Abraham G. Hartzema, Ph.D., M.S.P.H., Pharm.D. Professor of Pharmacy Administration UNC-CH School of Pharmacy Kerry E. Kilpatrick, Ph.D. Chair, Department of Health Policy and Administration UNC-CH School of Public Health Carl M. Shy, M.D., Dr.P.H. Chair, Department of Epidemiology UNC-CH School of Public Health Advisory Board: The Advisory Board advises on long-term goals. William H. Campbell, Ph.D. (Chair) Dean, UNC-CH School of Pharmacy Allen E. Cato, M.D., Ph.D. President & CEO Cato Research Ltd. Sandra B. Green, Dr.P.H. Senior Director of Health Economics Research Blue Cross & Blue Shield of North Carolina Clifford R. Guy, M.D. Vice President and Medical Director Winston-Salem Health Care Plan Abraham G. Hartzema, Ph.D., M.S.P.H., Pharm.D. Professor of Pharmacy Administration UNC-CH School of Pharmacy Kerry E. Kilpatrick, Ph.D. Chair, Department of Health Policy and Administration UNC-CH School of Public Health Carl M. Shy, M.D., Dr.P.H. Chair, Department of Epidemiology UNC-CH School of Public Health Hugh H. Tilson, M.D., M.P.H., Dr.P.H. Vice-President, Epidemiology Surveillance and Pharmacoeconomics Burroughs Wellcome Co. Raymond J. Townsend, Pharm.D. Vice-President, Applied Healthcare Research Glaxo Inc. Director: Abraham G. Hartzema, Ph.D., M.S.P.H., Pharm.D. Assistant Director: Carolyn Morris Assistant Research Professor: Anna Almarsdottir Staff CePOR's core staff represents faculty members with an extensive track record in outcomes research. The core staff includes faculty with expertise in experimental design, patient behavior, quality of life studies, survey methods, epidemiology, economics, pharmaceutical policy, and pharmaceutical insurance. The faculty have extensive international experience and are familiar with pharmaceutical regulatory and evaluation policies in European and Pacific Basin countries. Andrea K. Biddle, Ph.D. Assistant Professor of Health Policy and Administration UNC-CH School of Public Health Policy analysis, research methodology, childhood vaccinations William H. Campbell, Ph.D. Dean, UNC-CH School of Pharmacy Data bases, data structure E. Christine Hanson, Ph.D. Assistant Professor of Pharmacy Administration UNC-CH School of Pharmacy Patient outcomes, quality-of-life outcome measurements Abraham G. Hartzema, Ph.D., M.S.P.H., Pharm.D. Professor of Pharmacy Administration, UNC-CH School of Pharmacy Clinical Professor of Health Policy and Administration, UNC-CH School of Public Health Pharmacoepidemiology William P. Johnston, Ph.D. Assistant Professor of Pharmacy Administration UNC-CH School of Pharmacy Mental health outcomes Kerry E. Kilpatrick, Ph.D. Chair, Department of Health Policy and Administration UNC-CH School of Public Health Economic modeling, decision analysis G. Joseph Norwood, Ph.D. Chair, Division of Pharmacy Administration UNC-CH School of Pharmacy Pharmaceutical benefit reimbursement Kit N. Simpson, Dr.P.H. Clinical Assistant Professor of Health Policy and Administration UNC-CH School of Public Health Pharmacoeconomics, cost effectiveness and utility modeling, cost- effectiveness in clinical trials Brian E. Rittenhouse, Ph.D. Assistant Professor of Pharmacy Administration UNC-CH School of Pharmacy Economics, utility assessment See Appendix II for a selection of faculty projects. CePOR draws upon the expertise of other faculty on the UNC-Chapel Hill campus on an as- needed basis. These faculty bring relevant expertise, such as clinical knowledge and methodological skills, depending on the nature of the project and the therapeutic area investigated. Collaborators may include physicians, dentists, nurses, pharmacists, or other health experts as well as biostatisticians and experts in psychometrics and data management. CePOR's core faculty serve as principal investigators or as consultants to various research projects.
A principle component of CePOR's mission is translation of research findings. To fulfill this objective, research papers and publications are presented to audiences which include pharmacists, physicians, legislators, insurers, industry experts, and other academicians.
The following is a sample of the expertise of faculty, as evidenced in recently completed studies or studies currently underway. A more comprehensive list is presented in Appendix II.
As Related to Drug Classes: Antihypertensives Anti-migraine products Anti-Parkinson agents Psycho actives Thrombolytics Vaccines As Related to Disease Categories: Anxiety Disorders CNS and related syndromes Drug resistant TB Hemophilia Mood Disorders Myocardial infarction Schizophrenia Sepsis Tardive dyskinesia Urinary incontinence As Related to Pharmacy Delivery Systems: Administrative environment Computer technology/modeling Drug utilization review Formulary Patient consultation Reimbursement As Related to Research Methodologies: Dynamic cohort studies Patient attitudinal studies Patient follow-up studies Policy evaluations (OBRA 90, Medicaid) Quality of life assessment/cognitive, physical and psycho-social functioning Randomized experimental trials Standard gamble/trade-off analysis
Some outcomes research can be accomplished wholly or partially with existing databases. A critical need in outcomes research is to accurately and completely represent medical care utilization for a person or for a defined population. This requires a linked dataset with attributes including: (1) patient or user characteristics, (2) medical (including pharmaceutical) care use, (3) provider information, (4) system characteristics, and (5) chronological accuracy. Each research project may require a unique database, ranging from the very simple to the very complex. In order to facilitate outcomes research CePOR will serve as an access point to various data resources that will support outcomes research.
Potentially valuable databases for outcomes research include those from multiple hospital information systems (e.g., University Hospital Consortium), managed care systems (Kaiser Health Plan), federal reporting systems (FDA Adverse Drug Reporting System), state government systems (Medicaid Management Information Systems), international systems (World Health Organization), proprietary systems from claims processors (Pharmaceutical Data Systems), and many others. Through a series of cooperative agreements with external organizations, plus purchase or maintenance of on-site data resources, CePOR will serve as a clearinghouse between potential research projects and available databases. CePOR will provide extensive documentation, and in some cases on-line access, for databases either actively supporting, or potentially useful for, patient outcomes research.
CePOR Database Catalog includes documentation and/or on-line access for the following databases: National Ambulatory Medical Care Survey (NAMCS) National Medical Care Expenditure Survey (NMCES) North Carolina Mental Health System Medicare Current Beneficiary Survey Data National Claims History File Federal Drug Administration Red Watch System Owen Health Care, Inc. University Hospital Consortium, Inc. Kaiser Health Plan (Southern California Region) Kaiser Health Plan (Northern California Region) Group Health Cooperative of Puget Sound United Health Care
VIII. Multimedia Modeling Laboratory
A multimedia computer laboratory is proposed as an integral part of CePOR. The Laboratory will be equipped with three multimedia RISC-based workstations running advanced statistical, epidemiological, econometric and mathematical simulation modeling packages, as well as decision rule-writing software. Networking capabilities with UNC mainframes allow CePOR to run larger data sets.
A primary purpose and goal of the laboratory is to handle large data sets, such as the National Medical Expenditures Survey and the National Ambulatory Care Survey, and to examine and evaluate the variability in disease outcomes as a function of pharmacotherapy and pharmaceutical care. Applying neutral network tools to these and other data sets will allow scholars to examine the underlying structure of the data, testing hypotheses about pharmaceutical care interventions, optimal levels of these interventions, and health and economic outcomes. Another primary goal of the simulation laboratory is to enable development and construction of computer simulation models of disease processes as influenced by treatment strategies.
CePOR is strategically located on the Campus of the University of North Carolina at Chapel Hill. This campus is one of the nation's leaders in the Health Sciences, with top-ranked Schools of Dentistry, Medicine, Nursing, Pharmacy and Public Health. The campus has a large number of research programs including Centers for Cancer, Cystic Fibrosis, Hemophilia, Arthritis and many more, providing an interdisciplinary approach to research in these clinical fields. The School of Public Health with departments of epidemiology, biostatistics and health policy and analysis is located next to Beard Hall. CePOR is able to draw on the expertise in all these fields.
A major hospital is located on the campus; a large VA facility and the Duke hospitals are located within a short twenty minute driving distance. Collaborative arrangements with the North Carolina Association of Nursing Homes and the North Carolina Association of Long-Term Care Facilities provide access to institutionalized patients, while relationships with the Area Health Education Centers and the Family Medicines Clinics throughout the state provide access to ambulatory patients. Through the University's Computer Sciences Department and the Data Management Division of the Center for Health Services Research, secondary data analysis on third-party data, such as Medicaid, can be handled.
CePOR's location is in Beard Hall on the UNC-Chapel Hill campus which also is the home of the School of Pharmacy . The UNC campus is one of the three academic anchors of the Research Triangle Park (RTP), the other anchors being Duke University and the North Carolina State University . Several major pharmaceutical companies are located in the RTP, including Burroughs Wellcome and Glaxo, Inc. Furthermore, the RTP and its surroundings are home to several major contract research organizations such as Quintiles, Cato Research Ltd., ClinTrials Inc., PPD, the Research Triangle Institute , and several others. Based on the specific needs of a project, contractual, or sub contractual arrangements with any of these organizations can be made.
X. CePOR's Educational Activities
Newsletter "Newsletter" Editor: C. Daniel Mullins, Ph.D.
The newsletter is a quarterly publication that will chronicle current developments in outcomes research. The newsletter will contain an editorial, a lead story, several relevant methodological development reviews, a watch on the international regulatory scene and a brief synopsis of the relevant literature. It is targeted towards the pharmaceutical industry, contract research organizations, governmental agencies and academia. Sponsorship for the newsletter is solicited.
Short Course in Pharmacoeconomics
A short summer course is under development. This 3 day summer course will be provided in conjunction with the very successful clinical epidemiology course which is provided yearly at the University of North Carolina campus. International experts in pharmacoeconomics will serve as faculty for the course which will provide introductory training to representatives from government, contract research organizations, health care providers and payers, and industry.
Symposia and Seminars
Free-standing symposia and seminars will be organized under the auspices of CePOR. These may range from training programs for the pharmaceutical industry or other agencies to symposia dedicated to a specialized topic area. Such symposia may also be organized as part of conferences organized by other associations or organizations. Individual sponsorship will be encouraged for a variety of forums on selected topics.
CePOR sponsors periodic seminars on issues relating to outcomes research. Recent seminars include: Dr. Amiram Gafni McMaster University Department of Clinical Epidemiology and Biostatistics Presentation: "Controversies in Utility Assessment: Healthy Years Equivalents as a Step Forward" David R. Holtgrave, Ph.D. Centers for Disease Control Presentation: "Markov Processes" Joseph Lipscomb, Jr., Ph.D. Duke University Department of Community and Family Medicine and Institute of Policy Sciences and Public Affairs Presentation: "Methodological Challenges in Applying Cost-Benefit and Cost-Effectiveness Analysis to Pharmaceutical Decision Making: An Overview" David B. Matchar, M.D. Duke University Center for Health Policy Research and Education Presentation: "Patient Outcomes Research Teams: Stroke" Dr. Bernie O'Brien McMaster University Department of Clinical Epidemiology and Biostatistics Presentation: "In Search of Power and Significance: Issues in the Design and Analysis of Stochastic Economic Appraisals in Health Care" Dev S. Pathak, D.B.A. Ohio State University School of Pharmacy Presentation: "Cost Utility Analysis: Problems and Prospects" Eleanor Perfetto, Ph.D. Agency for Health Care Policy and Research Center for Medical Effectiveness Research Presentation: "Research Activities at AHCPR" Dennis A. Revicki, Ph.D. Medical Technology Assessment and Policy Research Center Battelle Human Affairs Research Centers Presentations: "Overview of Medical Technology Assessment and Policy Research at Battelle" and "Cost Utility Analysis of Maintenance Treatment for Depression" Syam Sarma, Ph.D. PPD, Inc. Outcomes Research Division Presentation: "Developing Outcomes Research Analyses from Clinical Trials Data"
Working Paper Series
CePOR sponsors a working paper series that is peer-reviewed. It presents the work of CePOR- affiliated staff in the pre-journal submission phase of such work and disseminates it under the auspices of CePOR to the educational and professional community. The purpose of the Working Paper Series is to facilitate dissemination and discussion of CePOR-related research on a pre- publication basis. Views expressed in these papers remain those of the author(s) and do not necessarily reflect the opinions of CePOR staff or sponsors.
Fellowships
CePOR provides several fellowships to sponsor visiting scholars and graduate students pursuing research in areas relating to its mission. The review and selection process for fellows is competitive, based on academic qualifications and research potential. Both general research fellowships and specific area fellowships are awarded, depending on funding source stipulations.
References Donabedian A. Evaluating the quality of medical care. The Milbank Memorial Fund Quarterly 1966 (July); 44(3), Part 2. Eddy D. Should we change the rules for evaluating medical technologies. In: A. Gelijns (ed). Modern Methods of Clinical Investigation. Washington DC: Institute of Medicine, National Academy Press, 1990. GAO. Cross Design Synthesis: A New Strategy for Medical Effectiveness Research. Washington DC: U.S. General Accounting Office, March, 1992.
Intra-University Liaisons Cecil Sheps Center for Health Services Research (Gordon DeFriese, Ph.D.) Department of Biostatistics (Gary Koch, Ph.D.) Department of Epidemiology (Carl Shy, M.D.; Harry Guess, M.D.) Department of Health Policy and Administration (Kerry Kilpatrick, Ph.D.) General Clinical Research Center (Eugene Orringer, M.D., Director) Lineberger Cancer Research Center (Joe Pagano, M.D., Director) School of Medicine (Steven Downs, M.D.) Other Liaisons Agency for Health Care Policy and Research (Eleanor Perfetto, Ph.D.) Group Health Cooperative of Puget Sound (Michael von Korff, Ph.D.) Kaiser Foundation (Merwyn Greenlick, Ph.D.; Rich Johnson, Ph.D.) Office of Technology Assessment (Jane Siske, Ph.D.) Research Triangle Institute (Donald C. McLeod, M.Sc.) Inter-University Liaisons Brunel University, U.K. (Martin Buxton, Ph.D.) Center for Health Policy, Research and Education; Duke University (Frank Sloan, Ph.D.; David Matchar, M.D.; Joe Lipscomb, Ph.D.) McMaster University (G.W. Torrance, Ph.D.; Bernie O'Brien, Ph.D.; Amiram Gafni, Ph.D.; David Feeney, Ph.D.) Ohio State University (Dev Pathak, Ph.D.) York University, U.K. (Michael Drummond, Ph.D.) Pharmaceutical Industry Wellcome LTD, U.K. Mr. Brian Lovatt Schering Plough International David Sugano, Ph.D. Ciba, AG Switzerland Erol Caglarcan, Ph.D. Contract Research Organizations (CROs) PACT A Besselaar Company 150 Radner-Chester Road, Suite D200 St. Davids, PA 19087 Phone: (215) 687-5454 James A. Bannon, Pharm.D. Senior Vice-President, Medical Operation Summer internships for graduate students Orphan Medical Bert Spilker, M.D, Ph.D. 13911 Ridgedale Drive Minnetonka, Minnesota 55305 Phone: ( 612) 541-1868 ClinTrials, Inc. Jim Higgins, Ph.D. Phone: (919) 460-9005
I. Past Research
The Relevance of Searching for Effects Under a Clinical Trial Lamp Post: A Key Issue. Pre-marketing randomized clinical trials of pharmaceuticals reveal important information, but it is not necessarily valid to use such trials in economic analyses. This paper juxtaposes results of economic studies based on (1) inclusion criteria in a clinical trial and (2) likely population that will actually receive the drug in a practice environment. The results from these two approaches are significantly different even having altered only the assumption concerning the population receiving the drug and not adjusting for all other significant artificial attributes of trials as compared to medical practice.
The Economics of Screening and Outcome Valuation. The economics of prevention programs and screening in particular depend crucially on the accuracy of screening tests (as measured by sensitivity and specificity). However, it is also crucial that the value of the screen must also depend on the effectiveness of any medical activity that is consequent to the test results. Furthermore, the effectiveness itself is somewhat less than maximally informative. More important is the valuation of any effectiveness of treatments. Thus, the accuracy of a test is only one aspect that must be considered in assessing the value of a screen. Its value hinges on the ability of subsequent interventions to generate valuable outcomes. Whether this expanded assessment of a screen is ultimately "worth it" then depends on the costs of achieving the outcomes.
Managed Care Formulary Decision Making. The purpose of this study was to determine what drug factors are important for P&T committee members to select drugs onto their formulary. Individuals rank ordered hypothetical drugs and conjoint analysis was used to evaluate the importance of specific factors.
Incontinence in Long-Term Care Facilities. A series of studies was completed to assess incontinence and causal factors in a large sample of resthome residents. Incontinence prevalence in this population was approximately 28 percent. Literature analysis revealed a large series of reports implicating anticholinergics and other drugs in incontinence, however no causal evidence is documented. Modeling behavioral, environmental, socio-demographic and drug exposure or incontinence pointed out that cognitive impairment and lack of nursing attention may be the most pertinent causes of incontinence. A research population of different types of incontinence was established for future drug study purposes.
Impact of a Clinical Pharmacist's Emphasis on Patient Communication on the Public's Attitude Toward Pharmacy. This project assessed the impact of instituting a clinical pharmacist in a family practice clinic on the public's attitude toward pharmacy over time.
Effect of Pharmacist's Knowledge on Quality of Consultations. This project assessed the extent to which knowledge and attitude of pharmacists influenced the quality and extent of pharmacist consultation with both patients and physicians.
Assessment of the Extent to Which Prescribing Patterns Differed Between Medicaid and Private Paying Patients. One focus was on the impact of payment status on the quantity of drugs dispensed per prescription. The other focus was on the differences in the prescribing patterns of primary care physicians.
Evaluation of Effects of Capitation Fee for Pharmacy Services. The project involved converting community pharmacies from a fee-for-service (FFS) form of reimbursement to a capitation basis of reimbursement under the Medicaid program. The studies focused on assessing the differences among experimental (capitation) pharmacies as compared to the control (FFS) pharmacies before and after the implementation of the new reimbursement system. The studies assessed differences in drug-drug interactions, quantity of drugs dispensed per prescription, use of over-the-counter drugs, use of generic substitutes and use of therapeutic substitutes among the capitation pharmacies as compared to the fee-for- service pharmacies, both before and after implementation. The assessment focused on changes in costs and quality resulting from the intervention.
Potential Inconsistencies Between Cost-Effectiveness and Cost-Utility Analyses: An Upstairs/Downstairs Socioeconomic Distinction. Cost-utility analyses (CUA) have traditionally been used when outcomes have an important quality dimension. In the absence of such a quality dimension, cost-effectiveness analysis (CEA) is often used. The quality adjustment in the denominator of a CUA could, however, be equally well done by a cost adjustment in the numerator via, for example, a willingness to pay assessment). This paper indicates that while these substitute methods are equally valid, they can lead to inconsistent rankings of treatment alternatives.
Economic Incentives and Disincentives for Efficient Prescribing. This discusses how the economic notion of efficiency challenges traditional medical notions of efficiency and suggests a way to remove some of the derivative ethical conflicts. It also examines the role of economic incentives in inducing physicians to act in the name of "efficiency."
Community Pharmacists as Immunization Advocates: Cost Effectiveness of a Cue to Influenza Vaccination. This research describes a pharmacist-based intervention - a cue to immunization for influenza and assesses its effectiveness and economic impact and cost.
Investigation of the Relationship Between Purchasing Generic Prescriptions, Generic Groceries and Generic OTC Products. The primary purpose of this study was to determine whether there was a relationship between the purchasing of generic prescription drug products (GPDP) and other generic products. In other words, the purpose was to ascertain whether individuals who frequently purchase generic products will be more likely to purchase GPDP, and also to determine if frequent purchasers of GPDPs perceive GPDPs differently than brand name prescription products.
II. Current Research
Drug Knowledge Acquisition by Children in Different Stages of Development. This study is conducted in collaboration with the EEC-COMAC program. Two groups of children, six years and ten years of age, are interviewed to assess factors that may explain children's knowledge about medications and children's perceived benefits of medications. Children's developmental theory states that they achieve autonomy in making decisions about illness behavior between the ages of six and ten. Knowledge concerning factors that are important in the evolvement of autonomy can assist in effective structuring of drug education and information for this age bracket. Such education and information may help to establish cohorts of informed rationale decision makers in drug therapy.
DTacP Vaccine Patient Follow-Up Study. This is a 30,000 patient follow-up study in a large HMO designed to develop the safety profile of the vaccine as mandated by the U.S. Food and Drug Administration. The objectives of the study are to assess the incidence rate of major adverse events in: 1) a population of children 15 to 18 months old in temporal relationship to their fourth vaccination; and 2) a population of children 4 to 6 years old in temporal relationship to their fifth vaccination.
Functioning and Medication Management in Older People. The study's objective is to improve cognitive, psycho-social and physical functioning in older patients with complex and multiple medications. A referral clinic is established for patients 65 years of age who have over ten concurrent medications prescribed by multiple physicians. In a randomized experimental trial, the effect of rationalizing drug therapy on the health status of the elderly patients is evaluated.
Prophylactic Recombinant Factor VIII Therapy on Bleeding Incidences and Related Patient Outcomes. Juvenile hemophilia patients are placed on prophylactic recombinant Factor VIII therapy in a randomized, controlled experimental trial. A prophylactic dosing schedule is developed to trough levels of circulating Factor VIII of 71%. Controls receive "on demand therapy" in response to bleeding incidences. Health and cost outcomes are evaluated.
Theoretical and Conceptual Development of Pharmaceutical Outcomes Research. Theoretical and conceptual research in pharmaceutical outcomes evaluation is focused on the relationships between biological parameters, clinical manifestations and health outcomes and the relevance of each of these levels of measurement for economic evaluations. Moreover, structure and dependence of each of these levels of outcomes is a fruitful area for conceptual development.
Off-Label Use of Drugs. Characteristics of the off-label uses of classes of drugs are evaluated, according to drug, drug class, physician specialty type, indication and patient follow-up. Methods will be developed to provide for cost-effective clinical assessment of documented off-label use of these drugs (historical controls).
Use of Vaginal Antifungals Among Medicaid Patients. This study assesses the use of vaginal antifungal products among Medicaid patients in North Carolina in comparison to the use of these products in two control states. North Carolina does not cover any over-the- counter drug products whereas some other states provide coverage for selected over-the- counter products. It is hypothesized that in North Carolina, there has been a shift away from usage of the lesser expensive vaginal antifungals which are now available over-the- counter towards the use of more expensive prescription products as a number of popular drugs in this category have shifted from prescription to nonprescription status. Inasmuch as the nonprescription product is less expensive, this policy may lead to an increase in drug costs for this therapeutic category. It is also hypothesized that the more open formulary approach used by the control states may be a more appropriate policy in terms of the total impact on drug and health costs.
Generic Versus Disease Specific Quality of Life Instruments. Researchers addressing the impact of a drug on the quality of life of the patient are frequently confronted with the question of whether it's more appropriate to use a generic or a disease specific quality of life instrument. This study hypothesizes that it's more appropriate to use generic instruments when there is only one disease present but it's more appropriate to use disease specific instruments when there are co-morbidities. The study will involve data analysis from a number of clinical studies over several major therapeutic categories where both disease- specific as well as generic instruments were used.
Obtaining Practice-Relevant Information from the Artificial Environment of Clinical Trials. Early clinical trials provide a very different environment than the real world of medical practice for assessing pharmaceutical health and economic implications. This work shows how protocol-induced effects in controlled trials may be estimated and eliminated in an economic assessment based on an early phase clinical trial so that results are more accurately reflective of real world environments.
Another Deficit Problem: Clinical Trials v. Clinical Practice. Clinical trials, while a perceived gold standard for much standard assessment of pharmaceuticals, are potentially quite distortive of results to be expected in a medical practice environment. This paper describes numerous problems with trials indicating how the randomized controlled trial paradigm falls short of what is required in technology assessment and suggests an alternative approach to predicting the economic value of pharmaceuticals.
Patient -Specific Information from the Randomized Response Interview. This paper examines the information provided by the randomized response interview, a technique typically used for estimating the sample proportion of respondents engaging in behavior of a type where it is expected that direct questioning would elicit false responses (e.g. sexual or criminal behavior). This paper indicates that contrary to prior claims, this technique does reveal information about individual behavior (albeit probabilistically). The estimated sample proportion of those engaging in the behavior under examination is not the probability of the behavior for any individual. Rather, responses to the randomized response interview yield added information specific, at least, to the elicited responses (yes or no) and, in some cases, to other characteristics of respondents.
Implementing the Randomized Response Interview to Assess Compliance in a Pharmacy. This project is a pilot study of the Randomized Response Interview (RRI) in the area of compliance assessment. It specifically selects likely noncompliant patients and follows a direct question regarding compliance with a similar question for which the RRI is used. Estimated proportions of noncompliers will be compared between the two methods and respondent comfort with the RRI will be assessed. If the RRI shows potential superiority, it will be tested in a number of settings and indication areas.
Dispensing Errors at an Outpatient Pharmacy. This project is an observational study that looks at the errors committed during the normal dispensing activity of a pharmacist in an outpatient setting. It evaluates the causes of errors, the time requirements for checking, the distractions that occur and the implications for future policy changes. The results of this study will draw attention to areas where effective time management is necessary toward reducing distractions. It will also focus the attention of administrators on the need for better quality control within their facility.
Developing and Evaluating Cognitive Services for Non-Literate and Visually Impaired Patients in Community Pharmacy Sites. The primary objective of this study is to develop and evaluate a pictorial labeling system and counseling protocol that meet the special needs of non-literate and visually impaired patients. To evaluate the system, a randomized crossover design using 10 sites will be conducted. The pictorial system will be evaluated on the basis of patient comprehension, compliance and health status.
III. Future Projects
Cost/Benefit, Cost/Effectiveness and Cost/Utility Analysis of Pharmaceutical Interventions. Included here is the development and application of broadly relevant outcome measures such as quality of life measures. The field is moving away from general outcome measures toward disease-specific measures. The development of more precise models and more efficient methods for assigning dollar values for changes in disease status and treatment processes are also needed to establish cost profiles for various treatment protocols. This costing research requires a detailed understanding of the clinical course of the disease, methodological issues and the potential clinical manifestations of the treatment regimen. Comparison to alternative therapeutic choices, such as medication or surgery, is needed for full economic evaluation.
Determinants of Success in Dealing with Managed Care Authorities Regarding Reimbursement for Drug Products. Traditional contact points with providers and insurers of health care are constantly being altered by newer developments such as the emergence of the importance of evidence of cost-effectiveness. When more expensive drugs save resources in nonpharmacy areas, traditional contacts of industry sales are less relevant. Physicians have less control over prescription decisions, and pharmacy directors see their budgets increase with costs being saved in areas external to their purview. The new environment requires new contacts within health care organizations. Relevant research would focus on the effectiveness of alternative approaches to demonstrating cost- effectiveness of pharmaceutical interventions in managed care settings.
Assessing the Impact of Drug-Pricing Policies for Public Programs. What are the effects upon the cost of treatment and quality of life of the elderly if drug benefits are included under Medicaid? Do restrictive Medicaid pricing policies save money or reduce access to promising, cost-effective medicines? How are the treatments prescribed by physicians affected by the pricing and formulary policies of government programs? How do the 1990 federal budget amendments that tie Medicaid drug prices to private sector negotiated prices impact pharmaceutical manufacturers, health insurers, health care providers and patients?
Importance of Compliance in Health and Economic Outcomes Research. What are the benefits of increasing patient compliance. What effect does such action have on patient outcomes such as health status, costs, disease condition, satisfaction, etc.? What interventions are effective in enhancing compliance and thus improve patient outcomes? Measuring the effects of compliance and compliance interventions may be beneficial in both patient and economic outcomes.
Assessing the Utilization and Effectiveness of Pharmaceutical Interventions in Various Organizational Settings. The drug benefits paid by Health Maintenance Organizations are typically more generous than fee-for-service insurance plans. To what extent does this difference in coverage affect the use of pharmaceuticals among plan members? Does variance in usage of pharmaceuticals translate into different outcomes in terms of inpatient admissions, readmissions, days lost from work and total medical costs to the plan? Can HMOs benefit from a rigorous internal appraisal of the costs and effects of coverage policies for specific drugs? Also, what effect, if any, does mail order pharmacies have on patient outcomes?
Development of Computer Software for Computer Modeling of Economic and Health Outcome Analysis. For assisting in-house and external marketing efforts, economic and health outcome analysis can greatly benefit from use of computer simulation models. Such models permit clinicians or other decision makers to forecast quickly and easily the most probable economic and therapeutic results for particular patients or populations under varying clinical, economic and organizational conditions. Greater confidence in the analysis can be achieved as clinicians are able to fine-tune the analytic model to their own situations.
Measuring the Effect of Changes in the Payment Policies of Public Health Programs on the Use and Effectiveness of Medicines. The implementation of hospital payments based on Diagnosis-Related Groups (DRGs) dramatically changed the incentives of providers for Medicare patients and for Medicaid patients in those states using DRG-based Medicaid payment systems. Federal legislation reduced the prices Medicaid paid physicians for 245 "overvalued" procedures by 15 percent in April 1989, and by another 15 percent in January 1991. did this dramatic shift in pricing result in substitution of pharmaceutical interventions for surgery? If so, what effect does this shift have on total program costs and patient outcomes? To what extent have prescribing patterns in hospitals been affected by prospective payment incentives?
Contacts deleted from text, but do not want to lose their names: CATO Research Ltd. Allen Cato, M.D., Ph.D. David Bernstein, Ph.D. 4364 S. Austin Avenue Durham, NC 27713 Phone: (919)361-2286 Abt Associates Cambridge, Massachusetts * The University of North Carolina distinguishes three types of centers as defined in the Academic Program Development Procedures, January 1990. A research institute or center is defined as follows: A research institute or center is a formal organization which has research as its primary mission. A research institute or center may also provide instructional or public services programs, but such units do not have jurisdiction over academic curricula. They may offer courses in cooperation with academic units. * Effeciency is not the sole concern in decision making; equity considerations, to take but one example, may imply that economically efficient solutions are not politically acceptable. However, it is always valuable to indicate the economic implications of various options or decisions so that any tradeoffs between, for example, equity and efficiency become apparent to decision makers. (br201)