No job name

Volume 10 • Number 5 • 2007
Principles of Good Practice for Budget Impact Analysis:
Report of the ISPOR Task Force on Good Research Practices—
Budget Impact Analysis
Josephine A. Mauskopf, PhD,1 Sean D. Sullivan, PhD,2 Lieven Annemans, PhD, MSc,3 Jaime Caro, MD,4C. Daniel Mullins, PhD,5 Mark Nuijten, PhD, MBA, MD,6 Ewa Orlewska, MD, PhD,7 John Watkins, RPh, MPH,8Paul Trueman, MA, BA9 1RTI Health Solutions, Research Triangle Park, NC, USA; 2University of Washington, Seattle, WA, USA; 3IMS Health, Brussels, Belgium;4Caro Research, Concord, MA, USA; 5University of Maryland, Baltimore, MD, USA; 6Imta, Erasmus University, Rotterdam, The Netherlands;7Centre for Pharmacoeconomics, Warsaw, Poland; 8Premera Blue Cross, Bothell, WA, USA; 9York Health Economics Consortium,York, UK A B S T R AC T
Objectives: There is growing recognition that a comprehen-
use and costs for the treatments and symptoms as would sive economic assessment of a new health-care intervention at apply to the population of interest. The Task Force the time of launch requires both a cost-effectiveness analysis recommends that budget impact analyses be generated as a (CEA) and a budget impact analysis (BIA). National regula- series of scenario analyses in the same manner that sensitivity tory agencies such as the National Institute for Health and analyses would be provided for CEAs. In particular, the input Clinical Excellence in England and Wales and the Pharma- values for the calculation and the specific cost outcomes ceutical Benefits Advisory Committee in Australia, as well as presented (a scenario) should be specific to a particular managed care organizations in the United States, now require decision-maker's population and information needs. Sensitiv- that companies submit estimates of both the cost- ity analysis should also be in the form of alternative scenarios effectiveness and the likely impact of the new health-care chosen from the perspective of the decision-maker. The interventions on national, regional, or local health plan primary data sources for estimating the budget impact should budgets. Although standard methods for performing and pre- be published clinical trial estimates and comparator studies senting the results of CEAs are well accepted, the same for efficacy and safety of current and new technologies as progress has not been made for BIAs. The objective of this well as, where possible, the decision-maker's own population report is to present guidance on methodologies for those for the other parameter estimates. Suggested default data undertaking such analyses or for those reviewing the results sources also are recommended. These include the use of of such analyses.
published data, well-recognized local or national statistical Methods: The Task Force was appointed with the advice and
information and in special circumstances, expert opinion.
consent of the Board of Directors of ISPOR. Members were Finally, the Task Force recommends that the analyst use the experienced developers or users of budget impact models, simplest design that will generate credible and transparent worked in academia, industry, and as advisors to govern- estimates. If a health condition model is needed for the BIA, ments, and came from several countries in North America, it should reflect health outcomes and their related costs in the Oceana, Asia, and Europe. The Task Force met to develop total affected population for each year after the new inter- core assumptions and an outline before preparing a draft vention is introduced into clinical practice. The model should report. They solicited comments on the outline and two be consistent with that used for the CEA with regard to drafts from a core group of external reviewers and more clinical and economic assumptions.
broadly from the membership of ISPOR at two ISPOR meet- Conclusions: The BIA is important, along with the CEA, as
ings and via the ISPOR web site.
part of a comprehensive economic evaluation of a new health Results: The Task Force recommends that the budget impact
technology. We propose a framework for creating budget of a new health technology should consider the perspective of impact models, guidance about the acquisition and use of the specific health-care decision-maker. As such, the BIA data to make budget projections and a common reporting should be performed using data that reflect, for a specific format that will promote standardization and transparency.
health condition, the size and characteristics of the popula- Adherence to these proposed good research practice prin- tion, the current and new treatment mix, the efficacy and ciples would not necessarily supersede jurisdiction-specific safety of the new and current treatments, and the resource budget impact guidelines, but may support and enhance local Address correspondence to: Sean Sullivan, University of Washington, Pharmaceutical Outcomes Research and Policy Program, Box357630, 1959 NE Pacific Street, Health Sciences Center, H-375, Seattle, WA 98195-7630, USA. E-mail: 2007, International Society for Pharmacoeconomics and Outcomes Research (ISPOR) Budget Impact Analysis Task Force Report recommendations or serve as a starting point for payers Keywords: budget impact analysis, economic evaluation,
wishing to promulgate methodology guidelines.
Whereas, CEA evaluates the costs and outcomes ofalternative technologies over a specified time horizon Definition and Intended Use to estimate their economic efficiency, BIA addresses the Budget impact analysis (BIA) is an essential part of a financial stream of consequences related to the uptake comprehensive economic assessment of a health-care and diffusion of technologies to assess their affordabil- technology and is increasingly required, along with ity. Admittedly, both CEA and BIA share many of the cost-effectiveness analysis (CEA), before formulary same data elements and methodological requirements, approval or reimbursement. The purpose of a BIA is to but there are important differences in how these data estimate the financial consequences of adoption and and methods are incorporated into the models because diffusion of a new health-care intervention within a of their different intended use. There may be circum- specific health-care setting or system context given stances where the CEA indicates an efficient technol- inevitable resource constraints. In particular, a BIA ogy while the BIA results indicate that it may not be predicts how a change in the mix of drugs and other affordable. In such instances, there is, unfortunately, therapies used to treat a particular health condition will no current scientific guidance on how to resolve this impact the trajectory of spending on that condition (see Fig. 1). It can be used for budget planning, forecastingand for computing the impact of health technology History of BIA changes on premiums in health insurance schemes.
Mauskopf et al. published an analytic framework for Users of BIA include those who manage and plan budget impact modeling in 1998 [1]. Others have for health-care budgets such as administrators of struggled with the need to include budget impact as national or regional health-care programs, administra- part of comprehensive economic evaluation [2–6].
tors of private insurance plans, administrators of Since the 1990s, several regions in the world including health-care delivery organizations, and employers who Australia, North America (Canada, United States), pay for employee health benefits. Each has a need for clearly presented information on the financial impact Hungary, Italy, Poland) and the Middle East (Israel), of alternative health-care interventions, yet each has have included a request for BIA alongside the CEA, different and specific evidentiary requirements for when submitting evidence to support national or local data, methods, and reporting.
formulary approval or reimbursement. Other coun- Budget impact analysis should be viewed as comple- tries have typically performed their own BI analysis mentary to CEA, not as a variant or replacement.
(The Netherlands) rather than requesting the BIA from Figure 1 Budget impact schematic. Adapted
from Brosa et al. [39].
Mauskopf et al. the manufacturer, although voluntary submission is complete report was then prepared by the cochairs, permitted. Country-specific guidelines for constructing and circulated to the Task Force members for review. A BIAs are also available [7–16]. These guidelines are face-to-face meeting of the Task Force was held to variable in terms of defining what constitutes a BIA discuss the draft and make revisions. This draft report and most of them provide only limited details on the was then sent to a group of primary reviewers chosen important factors in a BIA. An exception are the Polish to represent a broad range of perspectives. The review- guidelines [15], which provide precise recommenda- ers are identified in the Acknowledgments section of tions on perspective, time horizon, reliability of data the report. Following this review, a new draft was sources, reporting of results, rates of adoption of new prepared by the Task Force members and made acces- therapies, probability of redeploying resources, inclu- sible for broader review by all ISPOR members. This sion of off-label use, and sensitivity analysis.
final report reflects the input from all of these sources Despite the increased demand for BIA, a recent literature review indicates that the number of studiespublished in peer-reviewed journals is limited [17].
Purposes of the Document Some of these publications present cost studies that The purposes of this document are: 1) to develop a focus on the annual, 2- to 3-year or lifetime costs for a coherent set of guidelines for those developing or specific cohort of people or a representative individual reviewing budget impact analyses; and 2) to develop a being started on competing treatments [18–22]. A format for presenting the results of budget impact more limited number of published studies attempt to analyses that is useful for decision-makers.
estimate explicitly the financial and health-care service The intended audience is research analysts who impact of a new technology for a well-defined national perform budget impact analyses for health-care or health plan population [23–36]. There is ongoing decision-makers as well as health-care decision-makers debate as to whether BIAs should be publicly available who are responsible for local or national budgets.
for review and, if so, what parts should be published Others who may find this document useful include and/or made available for review upon request.
members of the press, patient advocacy groups, health-care professionals, drug and other technology manu- Task Force Process facturers, and those developing guidelines for their The cochairs of the ISPOR Task Force on Good Research Practices––Budget Impact Analysis, Jose- The panel recognizes that the methods for perform- phine A. Mauskopf and Sean D. Sullivan, were ing and reporting budget impact analyses continue to appointed in 2005 by the ISPOR Board of Directors.
develop. This report highlights areas of consensus as The members of the Task Force were invited by the well as areas where continued methodological devel- cochairs to participate, with advice and consent from opment is needed. The guidance is divided into three the ISPOR Board of Directors. Individuals were chosen main sections: 1) analytic framework; 2) inputs and who were experienced as developers or users of bud- data sources; and 3) reporting format.
getary impact models, who were recognized as scien-tific leaders in the field, who worked in academia, Recommendations for Analytic Framework industry, and as advisors to governments, and who For BIA, a description of the health condition, its came from several countries. This document reflects treatment and outcomes, is the essential component of the authors' own experiences developing budget the analytic framework. The purpose of a BIA is not to impact models and select publications, but is not produce exact estimates of the budget consequences of intended as a comprehensive review of the literature.
an intervention, but to provide a valid computing A reference group of ISPOR members from whom framework (a "model") that allows users to under- comments would be sought also was identified. The stand the relation between the characteristics of their Task Force held its first meeting at the ISPOR 10th setting and the possible budget consequences of a new Annual International Meeting in Washington DC in health technology (or a change in usage of current 2005 and held open Forums at the ISPOR 8th Annual health technologies). The BIA is a means of synthesiz- European Congress in Florence in 2005 and at the ing the available knowledge at a particular point in ISPOR 11th Annual International Meeting in Philadel- time for a particular decision-maker to provide a range phia in 2006.
of predictions specific to that decision-maker's infor- The Task Force reviewed other ISPOR guidance mation needs based on realistic estimates of the input documents that were developed to inform good scien- parameter values. Thus, the outcomes of the BIA tific conduct [37,38] and National Guidelines for BIAs should reflect scenarios that consist of a set of specific [7–16]. The Task Force held teleconferences and used assumptions and data inputs of interest to the decision- electronic mail to exchange outlines and ideas during maker rather than a scientifically chosen "base" or the subsequent months. Sections of the report were "reference" case based on assumptions and inputs prepared by Task Force members and a draft of the intended to be generally applicable.
Budget Impact Analysis Task Force Report This section presents the Task Force recommenda- Perspective. Budget impact analyses are primarily tions for the key elements of the analytic framework intended to inform health-care decision-makers, espe- for BIA. It addresses the overall design, the perspective, cially those who are responsible for national, regional, the scenarios to be compared, the population, time or local health-care budgets. Therefore, the recom- horizon, costing, sensitivity analysis, discounting, and mended perspective is that of the budget holder. Thus, unlike a CEA, where the recommended perspective isthat of society, which includes all cost implications ofan intervention, a BIA needs to be flexible enough to generate estimates that include various combinations Proper design of the analytic framework is a crucial of health care, social service and other costs, depend- step in BIA. The design must take into account the ing on the audience.
current understanding of the nature of the health con- The drawing of budget boundaries is a highly local dition and the evidence regarding the current and new exercise. In particular, some budgets may have a very technologies. There are several dimensions that must narrow focus. For example, in one location the phar- be considered: acuteness of the health condition, macy budget holder will only be concerned with the whether it is self-limiting, and the type of intervention expenses for drugs but in another, this may be sub- (preventive, curative, palliative, one-time, ongoing, sumed within a total hospital budget. Thus, the per- periodic). These dimensions will affect the degree to spective of a given budget holder may cover very which time-dependence is important in the design, different elements according to location. Whereas it is how the size of the population is estimated, the unit of mandatory for the analyst to address the needs of the analysis (episode vs. patient, for example), how the selected budget holders, it is also desirable for the intervention uptake is addressed, and the choice of analytic framework to be able to encompass broader (or even narrower) budgetary envelopes. In this way, These guidelines cannot address the details of the analysis will not only be able to show the decision- design of the analytic framework, but rather highlight maker what they need to see, but also can extend the key aspects to consider. It is important that what- beyond that to provide a more comprehensive view of ever choices are made, they be clear, justified, and with the fuller economic implications of the intervention.
a view to the simplest design that will meet the needs ofthe analysis.
Scenarios to be compared. Budget impact analyses Whether or not a health condition model is needed generally compare scenarios defined by a set of inter- depends on the type of health condition and interven- ventions rather than specific individual technologies.
tions at issue. For a chronic health condition, where The reference scenario should be the current mix of time dependency tends to be a major concern, a health interventions for the chosen population and sub- condition model is likely to be needed. The model groups. The current mix may include no intervention should be constructed so that it is consistent both with as well as interventions that might or might not be a coherent theory of the natural history of the health replaced by the new intervention. It may also include condition and with available evidence regarding causal off-label use. Introduction of a new technology sets in linkages between variables. Techniques currently used, motion various marketplace dynamics, including such as Markov models, might be appropriate, but product substitution and possibly market expansion.
newer techniques such as discrete event simulation, These need to be modeled explicitly with realistic and agent-based simulation, and differential equations justifiable assumptions before the comparisons among models may be considered if they are likely to be scenarios can be made. Thus, the analysis should con- accepted by the decision-maker. It is important that sider how the current mix of interventions is likely to researchers be alert to advances in modeling methods change when the new intervention is made available.
as well as to methodology requirements of payers For example, the new intervention might be added to rather than commit them to a given technique exclu- all existing interventions or it might replace all of the sively. For acute, self-limiting health conditions where current interventions or only those in a particular drug the episode is the unit of analysis, simpler techniques class. These constitute the new scenarios.
using deterministic calculations may be used.
The BIA should be transparent regarding the All of these methods are supported by a variety of assumptions made about the current mix of interven- software which is continually evolving. The software tions and the changes expected as the new intervention chosen and the resulting model should be accessible is added to the mix. The budget impact model should to the users in the sense that it should allow them be designed to allow alternative assumptions regarding to review all the model calculation formulae and to the scenarios to be compared.
change the assumptions and other inputs interactively;indeed, even the design of the model may result from Population. The population to be included in a BIA collaboration with the intended users.
should be all patients who might be given the new Mauskopf et al. intervention in the time horizon of interest. Specifying the impact that might be expected when a steady state who is included in this population is not straightfor- would be achieved if no further treatment changes are ward. It depends, of course, on the approved indica- assumed. This will vary with the condition and with tion, but it also reflects local intended restrictions on the impact of the new intervention, but will generally use (and reimbursement), possible beyond-restriction be longer than the current budget period because of use, induced demand (i.e., the proportion of previously costs and benefits that accrue over time. Although time untreated patients who now seek treatment because of horizons that go beyond a few years are subject to improved outcomes, greater convenience, or fewer side considerable assumptions, they may in exceptional effects), and the extent to which practitioners adopt cases be required to cover the main implications of the the technology or change patterns of use of existing health condition (e.g., some vaccinations). In any case, ones. The budget impact model must be designed to results should be available disaggregated over time in allow for examination of the effect of alternative periods appropriate to the budget holder (e.g., quar- assumptions about the nature and size of the treated terly, annual, etc.). Hence, to be most useful, the population as well changes in its nature and size over output must be the period by period level of expenses time. The Task Force did not recommend inclusion of and savings rather than a single "net present value." off-label use of the new technology in these scenariossince generally accepted methods for doing this are not yet available.
The steps in costing are identifying the resource use Typically, these populations are open in the sense that may change, estimating the amount of change, that individuals enter or leave the population depend- and valuation of these changes. In a BIA, identification ing on whether they currently meet the analyst's crite- must be done according to the perspective and interest ria for inclusion (e.g., by developing the indication, of the budget holder (see above). Moreover, the meeting the intended restrictions, no longer having resource use considered should be that which is rel- symptoms, etc.). This is in contrast with CEA where evant to the health condition and intervention of inter- populations are closed (i.e., a cohort of patients is est over the chosen time horizon. The Task Force defined at the start and all remain members through- members did not reach agreement on whether or not out the analysis). For example, if one of the criteria future costs should be included for other health con- defining the population is a moderate severity of ditions that might be incurred when the new interven- illness, then patients with mild disease are not part of tion results in additional survival. On this point, the the population but may enter when the disease Task Force proposes that the analyst should use her/his progresses; similarly, patients who are initially in the best judgment, given payer requirements and perspec- population with moderate disease may exit as the tives, when including or excluding future unrelated illness advances to a severe stage.
In general, the resource use profile should reflect the Subgroups. The analytic framework should allow for actual usage and the way the budget holder values subgroups of the population to be considered so that these resources. Thus, the valuation of these resources budget impact information can be made specific to refers to the expenditures expected to accrue (in the these segments. Such aspects as disease severity or short-run variable costs only and in the long-run both stage, comorbidities, age, sex, and other characteristics fixed and variable costs) rather than the opportunity that might affect access to the new intervention, or its costs per se. It is the transaction prices that are rel- impact on the budget, might be taken into account.
evant, including any rebates or other modifiers that This may also inform decisions regarding use of the may apply. For example, in some countries, readmis- new technology as a "first line" intervention or reserv- sions within a certain period will not generate another ing for use in patients failing other alternatives. The payment and in other jurisdictions, the physician's fee choice of subgroups must be founded on available depends on the number of times the patient is seen clinical and other evidence from epidemiological within a period.
studies, local knowledge, and so on.
In some cases, the intervention alters resource use and, thus, the capacity of the system, but this may have Time horizon. Budget impact analyses should be pre- no direct monetary consequence for the budget holder sented for the time horizons of most relevance to the because the system will not adjust financially within budget holder. They should accord with the budgeting the time horizon (e.g., personnel may not be rede- process of the health system of interest, which is ployed or let go). It may still be desirable to describe usually annual. The framework should allow, however, this impact on health services because it has implica- for calculating shorter and longer time horizons to tions for planning health system organization.
provide more complete information of the budgetary The impact on productivity and other items outside consequences. A particularly useful extension of the the health-care system costs should not routinely be time horizon for a chronic health condition is to reflect included in a BIA as these are not generally relevant to Budget Impact Analysis Task Force Report the budget holder. One exception may be when budget Size and characteristics of affected population; impact analyses are intended to inform the decision- Current intervention mix without the new making of private health insurers or employers. Such organizations may have a vested interest in maintain- Costs of current intervention mix; ing a healthy and productive workforce and, thus, they New intervention mix with the new intervention; may be able to offset productivity gains against Cost of the new intervention mix; and increased health-care costs. Another exception may be Use and cost of other health condition- and health-care systems relying on tax payments where lost treatment-related health-care services.
production due to morbidity could have important These six elements can be combined to calculate the implications for the payment of health.
budget impact of changing the treatment mix. The Task Force recommends possible data sources forderiving the inputs for each of these elements. Apart There is considerable uncertainty in a BIA. Therefore, from efficacy and safety which are assumed to be gen- a single "best estimate" is not a sufficient outcome.
eralizable aspects of the interventions, the inputs are Instead, the analyst should compute a range of results local. In many jurisdictions, the required data may not that reflect the plausible range of circumstances the exist or may be difficult to obtain. Nevertheless, analy- budget holder will face. Indeed, it might be argued that ses should be as evidence-based as possible, with the analytic framework itself is the most important expert opinion only used where alternative sources of product of a BIA rather than any particular set of data are not readily available. If expert opinion is used, results. It is useful to consider both a most optimistic care should be taken to frame the questions and choose and most pessimistic scenario. Having said this, the the experts in ways that generate reliable and general- ranges to be presented must be based on realistic sce- izable information. For example, the experts should be narios regarding the inputs and assumptions—a task asked for responses to questions that they know the that should be done collaboratively with the decision- answer to (e.g., how often do you schedule follow-up makers because they are best placed to make many of visits for a certain type of patient). No matter what the the key assumptions and to supply data for the ranges data source, the BIA should include measures of the of input parameter values.
range of possible input parameter values.
Various forms of sensitivity analysis (univariate, multivariate, probabilistic, etc.) may be carried out.
Their usefulness depends on the amount and quality of Size and Characteristics of the Population available data and the needs of the decision-maker. For The estimated sizes of the population and of the rel- example, there is little point to an extensive probabi- evant subgroups over time are critical for a determina- listic sensitivity analysis when little is known about the tion of the budget impact. The ideal way to obtain this degree of variability and the extent of correlation estimate would be from the epidemiological data in the among parameters.
decision-maker's own population before and after theintroduction of the new technology. As these data are not usually readily available even for the current tech- As the BIA presents financial streams over time, it is nologies, various alternative methods can be used to not necessary to discount the costs. The computational provide default estimates for a budget impact model.
framework should be constructed so that the decision- One approach is to employ epidemiological data maker can readily discount these results according to from nationally representative populations, adapted to local practice back to a decision time point if they wish the age, sex, and racial mix of the decision-maker's overall population. This generally involves the appli-cation of successively more restrictive inclusion criteria to the decision-maker's overall population. This Like all models, those used for BIA must be valid process requires rates such as the prevalence of the enough to provide useful information to the decision- condition, the proportion of patients with a particular maker. The steps to be followed in validation are con- severity or usage pattern, and other relevant features ceptually identical to those already identified in the for the health condition and technologies being exam- ISPOR Modeling Studies Task Force Report and are ined. In addition, change in prevalence over the time therefore not repeated here [37].
horizon of the model because of new incident casesand people leaving the population through death orother changes in disease progression must be applied Recommendations for Inputs and
over time to ensure that the size of the population Data Sources
continues to reflect the prevalence with the current and There are six key elements requiring inputs for the new technologies. This approach is relevant when modeling framework of a BIA: people are the unit of analysis. For some conditions, Mauskopf et al. however, it is an episode of illness that is the unit of is the decision-maker's own database. If these data are analysis (e.g., a migraine attack), and then it is the not available, then published information on current frequency of episodes in the population that must be treatment patterns, such as the results of primary or estimated with the current and new technologies.
secondary data studies or medical text books or review Another approach is to obtain directly from provid- articles, can be used. In addition to these data sources, ers their estimates of the number of people in their market research data or expert opinion on current and setting who would be part of the relevant population evolving treatment patterns may be used.
based on their current and anticipated new treatmentpatterns and aggregating this up to the budget holder's Cost of Current Intervention Mix The cost of the current technology mix involves mul- Regardless of the method used, it is important for tiplying the decision-maker's valuation of the technol- BIA to estimate not only the starting size of the popu- ogy by the number of people who receive each one in lation (or number of episodes) but also the way these each population subgroup. These costs should include are likely to evolve over time with and without the new the acquisition of the product, administration or technology. Hence, for the typically used open popu- implantation or other procedure costs as well as any lation, estimates of the inflows and outflows must be monitoring over the relevant time horizon. Costs of managing any side effects should also be included in Given the difficulties in obtaining data to provide the cost of current technology mix as a separate line accurate estimates of the population size, analysts should consider multiple sources including national The BIA should address the impact of compliance statistics, published and unpublished epidemiological and persistence with therapy on the cost of treatments.
data in the relevant, or similar, settings; registries; This must take into account whether the payer bears naturalistic studies carried out for other purposes; the cost anyway (e.g., even if poorly compliant, the claims data; and even expert opinion. The calculations patient still picks up the prescription). The assump- used to derive the population estimate should be pre- tions regarding compliance rates and persistence with sented in disaggregated format so that a decision- treatment should be based on the best available evi- maker could adjust the calculations to reflect their dence, which may come from database studies or spe- cific date collection or expert opinion. The relativecompliance and persistence on therapy should be Current Technology Mix reported at various time intervals. If patients do not fill For each population subgroup, it is necessary to iden- all the recommended prescriptions, then the cost of tify the interventions used currently and estimate the treatment should be reduced. In addition, the cost to proportion of patients using them, or proportion of the decision-maker should take into account drug dis- episodes in which they are used. Technologies may counts and patient deductibles and copays.
include no active treatment as well as drugs, devices,surgical or other modes of treatment. Some people New Technology Mix may receive more than one type of treatment which The new technology mix depends on the rate of uptake should be recorded separately in the current technol- of a new technology as well as the extent to which a ogy mix table. Table 1 gives an example of what these new technology replaces current technologies or is input parameters might look like. Although labeled added to them. The rate of uptake is likely to change "current," this technology mix may also evolve over over time as physicians and patients become familiar time even in the absence of the new technology and with a new technology. There are several ways to esti- this must be taken into account in budget impact mate the new technology mix. One way is to use the producer's estimates of market share over the first few Once again, the best data source for the current years after launch if these data are made available. An technology mix for the different population subgroups assumption must then be made as to whether the newintervention will be given in addition to current tech-nologies or whether it will substitute for some or all of Current technology mix the current technologies. For example, a new technol- ogy might reduce the use of a subset of the currentlyused technologies equi-proportionately (e.g., all drugs Drug A (combination of drugs B and C) Drugs B and C in separate doses in a particular class) or it might be added to all of the current technologies. The assumptions should be transparent and the model structured so that the Drugs C and D in separate doses budget impact of alternative assumptions about the new technology mix can be calculated. Another way to estimate the new technology mix is to incorporate Budget Impact Analysis Task Force Report Figure 2 Adherence and effectiveness. Notes:The relationship between effectiveness and adherence may be estimated based on observed data or expert
opinion or pharmacokinetic and pharmacodynamic data. The relationship in this figure is based on expert opinion:
Effectiveness Relative to Trial Data = Adherence rate (AR) if AR ⱕ 30%.
Effectiveness Relative to Trial Data = 1 - exp [-5 ¥ (AR - 0.2287)] if AR > 30%.
directly in the analytic framework usage rules that Effectiveness Analysis alongside Clinical Trials Task account explicitly for the new treatment pathways Force Report be used but simplified where possible and available, thus explicitly modeling how people switch adapted so that the estimates of the health outcomes to the new drug. For example, they may only switch are generated from a population perspective and pre- when they have failed on current therapy. Other ways sented for each year that is included in the BIA [37,38].
of estimating the new technology mix involve extrapo- For an acute or episodic illness, this adaptation is lating previous experience on product diffusion with straightforward. For a chronic or progressive illness, the same technology in other settings or with similar this adaptation may require an extension of the cost- interventions in the budget holder's setting.
effectiveness health condition model to account for theopen population and time-dependencies required for a Cost of New Technology Mix Costing of the new technology mix follows the same The BIA must be transparent about the assumptions process as for the current mix except that for technolo- made about the impact of noncompliance or reduced gies not yet on the market, the price may have to be compliance on effectiveness and about safety issues assumed if it is not yet set. In this case, we recommend associated with underutilization or overutilization of that the assumed technology cost be transparent and treatment and must allow them to be changed. If there justified. In addition, any uncertainty in the price are no published data on the relationship between should be readily able to be incorporated into alterna- compliance and health outcomes, then either pharma- tive scenarios for the sensitivity analyses.
cokinetic or pharmacodynamic data or expert opinionare possible alternative data sources. Figure 2 presents Use and Cost of Other Condition-Related a hypothetical example of the relationship between adherence and effectiveness that was generated using Although the health outcomes associated with differ- expert opinion.
ent technologies are not generally estimated explicitlyas part of a BIA, we recommend that they be estimated Recommendations for Reporting Format
and added to the BIA through changes in the cost oftreating the health condition of interest. Thus, alterna- This section presents a recommended reporting format tive technology mixes are likely to result in changes in for BIAs. The format presented below should be under- the symptoms, duration, or disease progression rates stood as the preferred ISPOR structure for the reporting associated with the health condition and, thus, in of any study regarding BIA. In view of the decision- changes in the use of all other condition-related health- maker-specific scenario basis that we have recom- care services. These changes will have an impact on the mended to be adopted for BIA, this format gives only health plan budget.
general directions for reporting.
In order to compute these changes in health out- comes and the associated changes in costs over the time horizon of the BIA, we recommend that estima- The introduction of the report of a BIA study should tion techniques similar to those described in the ISPOR contain all the necessary relevant epidemiological, Modeling Studies Task Force Report and the Cost- clinical, and economic information.
Mauskopf et al. Epidemiology and treatment. The introduction of a plished and the target audience (i.e., for which BIA study should present relevant aspects of the preva- decision-making body the study is intended). Ideally, lence and incidence of the particular disease as well as the model should be flexible enough to model the information on age, sex, and risk factors.
perspective of the budget holder and those of otherstakeholders with whom the budget holder must inter- Clinical impact. The clinical information should act. This requires disaggregation into the various cost consist of a brief description of the pathology, includ- components and categories of interest to these parties.
ing underlying pathophysiological mechanisms, and of In all cases, the perspective should be clearly stated and the prognosis, disease progression, and existing treat- transparent to the budget holder.
ment options, all of which are relevant to the design ofthe BIA study.
Model description. This section should contain a com-plete description of the structure of the BIA model, Economic impact. The economic impact information including a figure of the model. The description should should include any previous related studies on the allow the reader to identify outcomes for all treated condition of interest and associated therapies, for patients during the study period, including patients example, previous BIA studies in the condition of inter- with treatment failure.
est for another technology, cost-of-care studies, andcost-effectiveness studies.
Input data. The parameter values assumed for all theclinical data items and all the cost data items for all the scenarios modeled should be presented in the report.
This section should contain a detailed description of The level of detail should be such that the reader could the characteristics of the new technology compared duplicate all the calculations in the model.
with the current technologies: indication, onset ofaction, efficacy, side effects, serious adverse events, Data sources. The sources of model inputs should be intermediate outcomes, and adherence. A summary of described in detail. The strengths, weaknesses, and the clinical trials is given, including information on the possible sources of bias, that may be inherent in the design, study population, follow-up period, and clini- data sources used in the analysis, should be described.
cal outcomes.
Selection criteria for studies and databases should bediscussed and an indication is given of the direction and magnitude of potential bias in the data sources The objective of the BIA should be clearly stated. This which were used.
will be tied to the perspective(s).
Data collection. The methods and processes for Study Design and Methods primary data collection (e.g., for a Delphi panel) and The report should specify the design of the BIA, which data abstraction (e.g., for a database) should be will usually involve a modeling study. The following described and explained. The data collection forms characteristics of the model should be described.
which were used in the study should be included in theappendix of the report (e.g., the questionnaire for Patient population. This paragraph should clearly the Delphi panel, or the abstraction protocol for the specify the study population. The report should iden- tify and justify differences between the clinical trialpopulations and the BIA population.
Analyses. A description of the methods used toperform budget total and incremental analyses should Technology mix. The chosen technology mix with and be provided. The choice of all of the scenarios without the new technology should be discussed and presented in the results should be documented and justified. The choice of the technology mix is primarily based on the local treatment patterns and clinicalguidelines and this choice should be justified.
ResultsBoth total and incremental budget impact should be Time horizon. The time horizon(s) for the study presented for each year of the time horizon. Both should be presented and its choice justified. The choice annual resource use and annual costs should be pre- for the study period should be appropriate to the sented. The estimates of resource use should be listed budget holder.
in a table (if possible classified by technology applica-tion, technology side effects, and condition related) Perspective and target audience. This paragraph which shows the change in use for each year of the should clearly identify the perspective(s) from which time horizon. Another table should show the aggre- the study is performed, the costing that is accom- gated and disaggregated (e.g., pharmacy, physician Budget Impact Analysis Task Force Report visit, outpatient tests, inpatient care, and home care) study audit reports and the names and addresses of costs over time after applying costing information to participating experts and investigators.
the resource use. In general, budget impact estimatesshould be presented as a range of values, based on Budget Impact Computer Model alternative possible scenarios rather than a single point Because budget impact models need to be flexible enough to provide budget impact estimates for differ- Annual health outcomes for each year of the time ent health-care decision-makers, it is critical that the horizon do not need to be reported, but may be pre- software used to perform the model calculations is sented if these results are of interest to the decision- designed with both default input parameter values makers. For example, the health outcomes might be of based on credible national or local values and with the interest to the decision-makers when a large budget capability for the user to enter values that represent impact is accompanied by large health benefits.
their own particular situation. The model should be The results of the scenarios (sets of assumptions and programed so that the user can restore the original inputs and outcomes) analyzed should be described.
default parameters easily.
These scenarios may consist of optimistic, pessimistic, The model should be programed as easy-to-use and most likely input values determined from the sen- spreadsheets. For example, all input parameters would sitivity analysis of the key variables from the perspec- be presented on one input worksheet and outputs dis- tive of the decision-maker. We recommend that the played in one or more worksheets in a logical manner results of all sensitivity analyses be presented as a that summarizes the findings for the user. Graphical Tornado diagram.
output is often useful in the model. Introductory work- Inclusion of Graphics sheets should be included to describe the structure,assumptions, and use of the model. All sources and Graphical snapshots of the model's structure and data assumptions associated with input parameters should can be useful in summarizing for the user, who may be displayed with the parameters themselves and full wish to copy them for inclusion in their own internal references should be included on a reference work- reporting. Use of the following tools is recommended: sheet. The model calculations should be accessible tothe user and clearly and comprehensively presented.
Figure of the model. A graphical representation of the In many cases, the budget holder will be interested model structure makes it easier for the budget holder in modeling from more than one perspective. In such to understand what is represented by the outputs.
cases, model developers are encouraged to design the Simple flow diagrams are recommended to be included user interface so that the user can toggle between the with the model description.
different perspectives easily.
The user should be able to change easily any of the Table of assumptions. Listing the major assumptions input parameters. Color coding the input cells is a in tabular form can improve the transparency of the useful way of doing this. Changing the inputs allows model, particularly to the relatively inexperienced user the user to test various input scenarios. It may be and should be included with the model description.
useful to provide sample scenarios.
Finally, we recommend that the model be pro- Tables of inputs and outputs. Similarly, collecting the gramed so that the user can readily perform sensitivity model inputs and their data sources and outputs in analyses of relevance to their population.
tables provides a useful snapshot for the user andshould be included with the text on input data anddata sources.
Schematic representation of sensitivity analysis. Ana- Budget impact analysis is important, along with CEA, lysts should be encouraged to use diagrams (such as as part of a comprehensive economic evaluation of a Tornado diagrams which show graphically the impact new health technology. Some published examples of on the budget impact of feasible ranges of each input budget impact analyses are described in the review by parameter) as a simple way of capturing the key Mauskopf et al. [17]. We propose here a framework drivers of the model and presenting them to the user for creating budget impact models, guidance about the and should be included along with the text on the acquisition and use of data to make budgetary projec- results of the scenario analyses.
tions and a common reporting format that willpromote standardization and transparency. Adherence Appendices and References to these proposed good research practice principles The enclosure of relevant appendices to reports is would not necessarily supersede jurisdiction-specific encouraged. The appendices may cover the intermedi- budget impact guidelines, but may support and ate results (e.g., of individual Delphi panel rounds), enhance local recommendations or serve as a starting Mauskopf et al. point for payers wishing to promulgate methodology Format Version_2_1 Final_Final.pdf August 2005).
The following individuals provided suggestions and com- 10 NHS. National Institute for Clinical Excellence.
ments on the first draft of the Task Force Report: Sang-Eun Guide to the Methods of Technology Appraisal (ref- Choi, PhD, MPH, Health Insurance Review Agency, Korea; erence N05515). 2004. Available from: Karen Lee, MA, Canadian Agency for Drugs and Technolo- gies in Health, Canada; Maurice McGregor, MD, McGill August 2005).
University, Canada; Penny Mohr, MA, Centers for Medicare 11 Annemans L, Crott R, Degraeve D, et al. Recom- and Medicaid Services, USA; Ulf Persson, PhD, The Institute mended structure for reporting economic evaluation for Health Economics, Sweden; Jose-Manuel Rodriguez on pharmaceuticals in Belgium. Pharm World Service Barrios PharmD, MPH, MSc, Medtronic Iberia, Spain; Rod Taylor, PhD, MSc, University of Birmingham, UK; David 12 Collège des Economistes de la Santé. French Thompson, PhD, i3 Innovus Research Inc., USA; Jill van den Guidelines for Economic Evaluation of Health Care Bos, MA, Milliman USA, USA; and Johan van Luijn, RPh, Technologies. September 2004. Available from: Health Care Insurance Board, The Netherlands. The authors http: // / source / France_ wish to thank the 23 ISPOR members from 11 countries who provided detailed comments on an earlier version of the report, Jerusha Harvey from the ISPOR office for her excel- 13 Szende A, Mogyorosy Z, Muszbek N, et al. Method- lent administrative support in all aspects of the Task Force ological guidelines for conducting economic evalua- process and Executive Director of ISPOR, Dr Marilyn Dix tion of healthcare interventions in Hungary: a Smith, PhD, for her institutional support.
Hungarian proposal for methodology standards. EurJ Health Econom 2002;3:196–206.
14 Capri S, Ceci A, Terranova L, et al. Guidelines for economic evaluations in Italy: recommendations from 1 Mauskopf J. Prevalence-based economic evaluation.
the Italian Group of Pharmacoeconomic Studies.
Value Health 1998;1:251–9.
Drug Inf J 2001;35:189–2001.
2 Baltussen R, Leidl R, Ament A. Real world designs in 15 Orlewska E, Mierzejewski P. Proposal of Polish guide- economic evaluation. Bridging the gap between clini- lines for conducting financial analysis and their com- cal research and policy-making. Pharmacoeconomics parison to existing guidance on budget impact in other countries. Value Health 2004;7:1–10.
3 Anis AH, Gagnon Y. Using economic evaluations to 16 Guidelines for the Submission of a Request to Include make formulary coverage decisions. So much for a Pharmaceutical Product in the National List of Health Services. Version 4. April 2002. Available 4 Trueman P, Drummond M, Hutton J. Developing guidance for budget impact analysis. Pharmacoeco- PE%20GUIDELINES_Israel.pdf (Accessed Septem- ber 10, 2006).
5 Koopmanshap MA, Ruttem FF. The drug budget silo 17 Mauskopf JA, Earnshaw S, Mullins CD. Budget mentality: the Dutch case. Value Health 2001;6(Suppl.
impact analysis: review of the state of the art. Future Drugs Ltd 2005;5:65–79.
6 Drummond M, Brown R, Fendrick AM, et al.
18 Simon GE, von Korff M, Heligenstein JH, et al. Initial ISPOR Task Force. Use of pharmacoeconomics antidepressant choice in primary care. Effectiveness information—report of the ISPOR Task Force on use and cost of fluoxetine vs triciclic antidepressants.
of pharmacoeconomic/health economic information in health-care decision making. Value Health 2003; 19 Mullins CD, Ohsfeldt RL. Modeling the annual costs of postmenopausal prevention therapy: raloxifene, 7 Commonwealth Department of Health and Ageing.
alendronate, or estrogen-progestin therapy. Manag Guidelines for the Pharmaceutical Industry on Prepa- Care Pharm 2000;9:150–9.
ration of Submissions to the Pharmaceutical Benefits 20 Ozminkowski RJ, Wang S, Marder WD, et al. Cost Advisory Committee. Canberra, September 2002.
implications for the use of inhaled anti-inflammatory medications in the treatment of asthma. Pharma- 21 Grima DT, Burge RT, Beckr DL, et al. Short-term (Accessed August 2005).
cost-effectiveness of bisphosphonate therapies for 8 Canadian Coordinating Office for Health Technology postmenopausal osteoporotic women at high risk of Assessment (CCOHTA). Guidelines for Economic fracture. P&T 2002;27:448–55.
Evaluation of Pharmaceuticals: Canada. 2nd edn.
22 Nurnberg HG, Duttagupta S. Economic analysis of November 1997. Available from: sildenafil citrate (Viagra) add-on to treat erectile dys- function associated with selective serotonin reup- take inhibitor use. Am J Therapeutics 2003;11:9–12.
9 Academy of Managed Care Pharmacy Format for For- 23 Mason JM, Moayyedi P, Young PJ, et al. Population- mulary Submissions. Version 2.1. April 2001. Avail- based and opportunistic screening and eradication of Budget Impact Analysis Task Force Report Helicobacter pylori. Int J Tech Assess 1999;15:649– 32 Dykstra DM, Beronja N, Menges J, et al. ESRD managed care demonstration: financial implications.
24 Mauskopf JA, Tolson JM, Simpson KN, et al. Impact Health Care Financ Rev 2003;24:59–75.
of zidovudine-based triple combination therapy on an 33 Bigal ME, Rapoport AM, Bordini CA, et al. Burden of AIDS Drug Assistance Program. J Acquir Immune migraine in Brazil: estimate of cost of migraine to the Defic Syndr 2000;23:302–13.
public health system and an analytical study of the 25 Doyle JJ, Casciano J, Arikian S, et al. A multinational cost-effectiveness of a stratified model of care. Head- depressive disorder (MDD): a comparison of cost- 34 Tonelli M, Winkelmayer WC, Jindal KK, et al. The effectiveness between venlafaxine, SSRIs, and TCAs.
cost-effectiveness of maintaining higher hemoglobin Value Health 2001;4:16–31.
targets with erythropoietin in hemodialysis patients.
26 Mauskopf JA. Meeting the NICE requirements: a Kidney Int 2003;64:295–304.
Markov Model approach. Value Health 2000;3:287– 35 Casciano J, Amaya K, Doyle J, et al. Economic analy- sis of oral and topical therapies for onychomycosis of 27 Zabinski RA, Burke TA, Johnson J, et al. An eco- the toenails and fingernails. Manag Care 2003;12:47– nomic model for determining the costs and conse- quences of using various treatment alternatives for the 36 Sorensen J, Andersen LS. The case of Tumor Necrosis management for arthritis in Canada. Pharmacoeco- Factor-a inhibitors in the treatment of rheumatoid nomics 2001;19(Suppl. 1):S49–59.
arthritis: a budget impact analysis. Pharmacoeconom- 28 Orlewska E. The cost-effectiveness of alternative thera- peutic strategies for management of chronic hepatitis B 37 Weinstein MC, O'Brien B, Hornberger J, et al. Prin- in Poland. Value Health 2002;5:404–20.
ciples of good practice for decision analytic modeling 29 Mauskopf J, Muroff M, Gibson PJ, Grainger D. Esti- in health care evaluation: report of the ISPOR Task mating the costs and benefits of new drug therapies: Force on good research practices—modeling studies.
Value Health 2003;6:9–17.
Schizophr Bull 2002;28:619–35.
38 Ramsey S, Willke R, Briggs A, et al. Good research 30 Nuijten MJC, Rutten F. Combining a budgetary practices for cost-effectiveness alongside clinical trials: impact analysis and a cost-effectiveness analysis using the ISPOR RCT-CEA Task Force report. Value Health decision analytic modeling techniques. Pharmacoeco- 39 Brosa M, Gisbert R, Rodríguez Barrios JM y Soto J.
31 Tavakoli M. Modelling therapeutic strategies in the Principios, métodos y aplicaciones del análisis treatment of osteoarthritis: an economic evaluation of del impacto presupuestario en sanidad. Pharmaco- meloxicam versus diclofenac and piroxicam. Pharma- economics Spanish Research Articles 2005;2:65–

Source: http://www.heads.gr/downloads/17.%20Budget%20Impact%20Analysis.pdf