Are USA payers moving beyond pharmaceutical budget impact models and embracing cost–effectiveness analysis?

Assessing current trends for health economic modeling in the USA

The Academy of Managed Care Pharmacy (AMCP) issued revised guidelines (version 2.1) for formulary submissions to managed care for the appraisal of new medical interventions to be reimbursed in April 2005 [1]. The AMCP guidelines call for, among other things, the inclusion of a cost–effectiveness model along with a pharmaceutical budget impact model (BIM) to be included within the dossier. The BIM is designed to provide an estimate of the financial impact of a new technology, usually only on the pharmacy budget, as BIMs only include drug costs, network and other discounts, rebates, copayments and other benefit design parameters [1]. The manufacturer is left with the decision to include either a cost–effectiveness model, a BIM, both or neither within the AMCP dossier. The usefulness of models included within an AMCP dossier to payers is unclear and is related to a host of factors, including credibility of the manufacturer or the modeling group/company, simplicity, understanding of modeling techniques and the capabilities of personnel within these payer groups to understand and comprehend cost–effectiveness or budget impact models. Generally speaking, modelers have moved to the simplest approach and away from complex models and techniques for fear of having the end customer discard the model and results completely. To this end, the early trends in AMCP dossiers were towards the simple Excel spreadsheet models that focused only on pharmacy budget impact, thereby keeping the analysis simple and straightforward. Unfortunately, it gives an incomplete picture of the true economic value of a new medical intervention. By contrast, our international counterparts in countries, such as Australia, Canada and the UK, all have formalized guidelines that call for submission of a cost–effectiveness model with a budgetary impact model being optional or not required at all [2–4].

The USA has lagged behind other countries in formalizing cost–effectiveness analysis and modeling, while at the same time having those results used for formulary decision making. While other countries have also experienced significant inflation in the healthcare sector, the response to this has been more complete, with guidelines for the economic appraisal of new medications. Australia was the first country to formally adopt economic modeling methodology in 1990, with the most current revision of these guidelines published in the autumn of 2002 [3]. The Australian guidelines outline the requirements for a budget impact analysis, which include drug/intervention costs but also requires the manufacturer to estimate savings from side effects for existing treatments, savings to the health budget for fewer medical procedures, along with savings to the government health budget for fewer long-term outcomes (e.g., using antihypertensives to prevent stroke) [3]. The Australian guidelines, therefore, require a broader definition of budgetary impact by including morbidity offsets, whereas the AMCP guidelines in the USA do not make this distinction and rely more heavily on pharmacy budget impact. More specifically, the AMCP guidelines separate the budget impact model completely from the cost–effectiveness model and treat them as separate entities. The Canadian guidelines treat budgetary impact analysis completely different by labeling it a ‘financial impact analysis’ and stating that budgetary impact analysis can be undertaken if requested by the primary decision maker in Canada, for which the economic evaluation is intended, but is not part of the formal economic evaluation process [2]. The National Institute for Health and Clinical Excellence guidelines request that a manufacturer project the budgetary impact of a new intervention to the National Health Service in the UK with little additional direction [4].

Budget impact analysis in the USA is also changing. The time is fast approaching where managed care plans will require more than a simple budget impact model on an Excel spreadsheet and manufacturers will be required to provide more sophisticated modeling and estimation techniques. Moreover, with the beginning of Medicare Part D prescription drug coverage scheduled to begin on January 1, 2006, the Center for Medicare and Medicaid Services is likely to make a determination in the near future about economic evaluation guidelines for products to be reimbursed through this very large and brand new drug benefit program in the USA.

Diabetes is a good case study, contrasting the need for payers to have easily understandable models with an appropriate level of sophistication to accurately project morbidity and mortality into the future. The costly diabetes complications often occur many years, or decades, into the future and require sophisticated modeling techniques (e.g., Markov state transition models) in order for payers to estimate the value of one medical intervention compared with another. A simple budget impact analysis on an Excel spreadsheet that takes into account only pharmacy costs is going to undervalue interventions in the diabetes treatment armamentarium that prevent costly morbidity years or even decades into the future. While the AMCP guidelines cover the inclusion of both a budget impact and cost–effectiveness model, many healthcare plans often concentrate on the budget impact of a new intervention first, rather than weighing the true economic value of a new intervention to the healthplan. With this said, we must acknowledge that the average American remains within one healthplan in the USA for approximately 3 years, thereby making it more difficult for the plan to recoup any costs for newer diabetes medical interventions when the payoff in lower morbidity could be years into the future and it is unlikely that the person will still be within their original healthplan. It is encouraging that USA payers seem to be heading down the road of a broader definition of economic value rather than simple budgetary impact, but it must also be recognized and respected that USA healthplans are required to weigh both the immediate and long-term economic benefits of new medical interventions in order to remain viable business entities. A consensus panel organized by the American Diabetes Association has taken this a step further by issuing guidelines for computer modeling of diabetes and its complications [5]. The guidelines were designed to provide a road map for modelers in the diabetes area, such that modeling expectations are made clear at the beginning and payers could perhaps have more confidence in the results. Disease modelers would be well served to promote transparency in all cost–effectiveness and budget impact models while, at the same time, encouraging extensive validation and rigorous testing of assumptions. Perhaps then, payers will begin to have more faith in the estimates of economic value arising from these models both now and into the future and begin to use them more extensively in formulary inclusion decisions.