Best practice in modeling and forecasting
Todd Johnson, director of forecasting, Kantar Health, and Lee Blansett, senior vice president, oncology market access, Kantar Health, talk to eyeforpharma about European and US strategies to control spending and manage budgets; David Robinson, Kantar Healths senior director of epidemiology services, discusses the utility of epidemiological analyses of complex diseases.
E4P: How do you think market modeling and forecasting initiatives vary in markets like the US and Europe?
Todd Johnson and Lee Blansett: From a forecasting perspective, we think there is more in common than different, but important country-specific differences in reimbursement and coverage certainly have to be taken into consideration. In Europe, most countries have either a dominant single-payer system, or as in Germany, multiple payers that offer a very consistent benefit. Each European country needs to be forecasted separately as time between EMEAs approval for marketing and a countrys decision to reimburse a product can range from essentially zero months to multiple years, or not ever.
Some countries also have significant differences in adoption from one region to another, a situation that can further alter a forecasts adoption assumptions. Finally, across-the-board price cuts have been legislated in some European countries in the recent past. These introduce an extra measure of uncertainty around pricing assumptions. A forecast must account for these reimbursement hurdles in some form. One approach is to apply sensitivity analysis around share and price variables. We recommend scenario development or the use of Monte Carlo simulation to determine the likely range of sales revenue rather than a single number for most European forecasts.
In the US, of course, we do not at this time have a single-payer system of reimbursement or even a majority-payer situation. We also do not have cost-effectiveness evaluation bodies like NICE. We do, however, have to consider the impact of third-party payers that may reimburse at different levels. We sometimes build models that segment populations by reimbursement coverage or more commonly include Formulary Tier as an attribute in primary market research to gauge the impact on penetration of new products. We develop blended peak shares based on the outcome of the research and the mix of payers. With healthcare reform, this may change somewhat in the future but it may be a while before we figure this one out.
Finally, another factor when forecasting US and European markets together is accounting for currency fluctuations when rolling up the models. Usually, we will consider static exchange rates in the long term, but recent economic conditions have necessitated more scrutiny for short-term forecasts. (For more on the US and Europe, see Different forecasting methods in the US and Europe.)
What do you mean by sophisticated analytical methods and how can they be used?
David Robinson: Sophisticated analytical methods could be interpreted to mean complicated forecasting methodologies or simply an established and proven thought process that allows one to sift through minimal and often contradictory data and yet arrive at a defensible estimate of current and future disease populations. One can develop the most complicated analytical methodology in the world but if you cant find the data necessary to populate the model, its worthless. Its all about making the best-educated and informed estimate based on what data is available. It is ironic that oftentimes the most complex diseases are those that have the least data available, especially outside the United States. (For more on complex diseases, see Forecasting for complex diseases.)
So, the ultimate challenge is to create an analytical process that creates an accurate picture of current and future disease populations and is transferable across geographies and markets where data availability can be highly variable (or non-existent). Goal number 1 is self-explanatory; i.e., without accurate patient estimates any resulting forecast will be suspect. The issue of transferability is too often overlooked. Without this, it is impossible to make comparisons of opportunity across diverse markets.
Learn more from Kantar Health at the Pharma Forecasting Excellence Summit in Boston from October 5-6, 2010.