Top 7 Challenges Faced by Pharma AI
AI can be leveraged to speed up the processes of collecting and accessing information to drastically shorten drug development time and keep prices of new treatments down, which is becoming critical as the cost of discovering and developing drugs skyrockets, says Michael Galarnyk, AI evangelist at cnvrg.io.
AI can speed up the processes of collecting and accessing information to drastically shorten drug development time and keep the price of new treatments in check. This is becoming more critical as the cost of discovering and developing drugs is skyrocketing.
How expensive is developing a new drug?A study published in 2020 concluded that the median research and development cost for a new treatment was$985 million. A substantial part of this cost was the high trial failure rate. Roughly seven out of eight compounds that enter the clinical testing pipeline are never developed.
This is where AI comes in. AI has the potential to find previously unexplored patterns, not immediately, which can lead to a new understanding of diseases, and the drugs designed to treat them.
For example, Astra Zeneca uses machine learning models to discover more quickly which genes can cause resistance to cancer treatments and Samsung has built an app to detect early Covid-19 infection.
Despite the high number of successes, there are unique challenges that need to be overcome for pharmaceutical (pharma) companies to reap the benefits of AI. Here are a few examples:
Smaller Datasets – Most AI algorithms need big datasets to learn from. Due to a large number of diseases and illnesses, and the relatively small number of incidences of each, creating large data for each type of medical condition is very challenging. A rule of thumb is that for machine learning models in the pharma industry to work efficiently, they need a minimum of 2 to 3 years of historical data. The high number of mergers and acquisitions can make this goal challenging, especially when the original source of the data can available. Because pharma data sets are usually smaller with fewer patients and fewer observations per patient which makes achieving meaningful insights is more challenging. Complex Data – At the same time, there can be fewer data sets there can be far more features for each dataset. Patient data may include information relating to their past and current health or illness, treatment history, lifestyle choices, and genetic data. It can also include biometric data, which is any measurable physical characteristic that can be measured by a sensor, or wearable device. As a result, patient data can include alphanumeric data, x-ray, pathological, and clinical test reports images in multiple formats like JPEG/JPG and digital imaging and communication for medicine (DICOM) format. AI systems for drug development need the ability to support varied and complex data.See More: How is Artificial Intelligence Impacting Healthcare? Complicated Data Labelling – Data labeling is more complex and requires highly specialized input. Consider all the different types of expertise required to identify skeletal, internal-organ, nervous system, and vascular abnormalities from X-Ray images. It isn’t just finding the necessary expertise that’s challenging, labeling can be tedious and time-consuming.
