Forecasting Branded and Generic Pharmaceuticals

Research output: Contribution to journalArticlepeer-review

Electronic versions

  • Kostas Nikolopoulos
  • Samantha Buxton
    Swansea University
  • Marwan Khammash
    University of Sussex
  • Philip Stern
    University of Exeter
We forecast UK pharmaceutical time series before and after the time of patent expiry. This is a critical point in the lifecycle, as a generic form of the product is then introduced into the market, while the branded form is still available for prescription. Forecasting the numbers of units of branded and generic forms of pharmaceuticals dispensed is becoming increasingly important, due to their huge market value and the limited number of new ‘blockbuster’ branded drugs, as well as the imposed cost for national healthcare systems like the NHS. In this paper, eleven methods are used to forecast drug time series, including diffusion models (Bass model & RPDM), ARIMA, exponential smoothing (Simple and Holt), naïve and regression methods. ARIMA and Holt produce accurate short term (annual) forecasts for branded and generic drugs respectively, while for the more strategic horizons of 2–5 years ahead, naïve with drift provides the most accurate forecasts.
Original languageEnglish
Pages (from-to)344-357
JournalInternational Journal of Forecasting
Volume32
Issue number2
Early online date22 Jan 2016
DOIs
Publication statusPublished - 1 Apr 2016
View graph of relations