Understanding demand for any product is important. For orphan drugs, accurate economic models and forecasting is essential.
Orphan drugs are often expensive to produce. Pharmaceutical companies don’t want to create stockpiles that may expire before they reach patients. Yet under-production poses the risk that rare disease patients will not receive essential medications.
By choosing the right data and using effective models, pharmaceutical companies can develop better economic forecasts. Digital tools like artificial intelligence and machine learning can aid this process.
Finding the Right Data
Pharmaceutical forecasts have long relied on academic studies and market research data, much of which is collected by pharmaceutical marketing teams. While this information provides some insight, its scope and accuracy are limited.
“Users trying to query assumptions faced a lack of reliability and transparency, especially during times of rapid change,” writes David Wolter, vice president, consulting at IQVIA.
Real-world data offers a way around these limitations. According to Wolter, the benefits of real-world digitized data include:
- Up-to-date information, even in times of change or uncertainty.
- Reliable real-world information gathered by physicians from patients.
- Deeper insights gleaned from layering real-world datasets for a comprehensive view of patient journeys.
- More accurate forecasting via the application of artificial intelligence and machine learning.
The digital revolution offers to overcome the challenges of limited, outdated information. In doing so, however, it poses a new challenge: Overwhelming forecasters in an avalanche of information. Striking a balance that incorporates the right type and source of information has become a new focus for pharma economic modeling.
For example, in a study published in PLOS One, researchers Hazhir Rahmandad, Ran Xu and Navid Ghaffarzadegan examined previously created predictive models of the COVID-19 pandemic. The researchers found that many of these models omitted two essential data categories: real-world information on data transmission and human behavior in response to pandemic stressors. The models also failed to account for randomness in important ways.
By adding this information, Rahmandad, et al. were able to improve the accuracy of long-term predictions regarding the pandemic’s spread. Similarly, choosing the right data is a must for pharmaceutical marketing and analysis teams who wish to accurately predict demand for rare disease treatments.
Putting Data to Work
Gathering, standardizing and selecting data for economic forecasting is an essential first step. Choosing or developing effective models is vital as well.
In a 2022 article published in Frontiers in Medicine, researcher Aurelija Burinskiene notes that different forecasting models for drug sales have different applications. For instance, Burinskiene writes that while the naive method provides effective three- to five-year forecasts, it is less effective for single-year forecasts than Holt’s linear method. Choosing the right method can help ensure better results.
Digital tools, including artificial intelligence and machine learning, remain at the forefront of economic model creation and forecasting. In a 2022 paper in Neural Computing and Applications, Abdul Aziz Abdul Rahman and fellow researchers explore the use of shallow and deep neural networks in creating demand forecast models to assess future demand for eight distinct groups of pharmaceutical products. Their goal: to create a model that offers better data on which to base sales and marketing strategies for key medications.
Artificial intelligence and machine learning offer promise. Yet their use “requires careful quality and applicability assessment,” write Anne A. H. de Hond and fellow researchers in an article in npj Digital Medicine. Taking the time to assess models thoughtfully will pay off in better results and more efficient decision making.
When the right models are applied to accurate and timely real-world data, economic forecasting can provide better insights for orphan drug makers. The results may be better control of drug production and more efficient movement of these medications from their manufacturers into the hands of patients who need them.
Images by: andreypopov/©123RF.com, antonioguillem/©123RF.com