Problem
Drug demand forecasting is sensitive to well-known external variables because they are susceptible to seasonality, such as pollinosis or influenza, as well as market competition; however, sometimes they are unknown, such as a pandemic. The drug demand forecasting systems used by pharma companies based on influenza epidemic curves and statistical analysis of previous year’s flu drug purchasing trends were not sufficient to predict ‘normal’ drug demand following a year in which the influenza virus had all but vanished due to masks, vaccinations, and distancing. As a result, the post pandemic forecasting of drug production and distribution had left pharmacies and consumers without the drugs needed for symptom management. In uncertain and unpredictable environments, demand forecasting requires accurate, reliable forecasting models able to analyse large amounts of qualitative and quantitative data.
Moxoff solution
Based on the multiplicity and heterogeneity of economic-political-social, health and technological variables required to make accurate predictions of drug demand, a broad Artificial Intelligence model based on Machine Learning algorithms, integrated with mathematical modelling systems, was used. The integration of AI, Machine Learning and mathematics has allowed both the extraction of information and the output of simulations from a large amount of qualitative (changes in consumer behaviour) and quantitative (numbers or vectors of numbers) data describing the variables outside the market and helping to influence the forecasts. The forecast scenarios obtained from the trained algorithms on the company’s workflows allow the pharmaceutical industry to make production planning choices that are more representative of new possible scenarios; they also provide a flexible and rapid forecasting model to modify planning as unpredictable events occur.
- Flexibility, speed and accuracy of demand forecasting
- Customised business workflow simulations that can adapt to highly variable scenarios
- Support for production, distribution and marketing planning
