By Lyra Gao, Evardra Bell, and Raphael Cervantez (Analysts 2020–21)
The high failure rate of clinical trials is a significant bottleneck in drug development, which lies at the foundation of the pharmaceutical industry. Despite increased investment into drug discovery, fewer and fewer new drugs reach the market with each passing year due to this shortcoming. The re-emergence of artificial intelligence (AI) is a solution to this previously unaddressable problem and can dramatically increase drug development success, recovering up to $800M–$1.4B USD.
Currently, only 10% of drugs advance clinical trials and move onto FDA approval. Because clinical trials are the penultimate stage of the development process, a failed clinical trial undermines not only the investments into the trial itself, but also virtually all prior investments — a damage typically costing $800M–$1.4B USD.
The clinical trial market has been steadily growing and the number of clinical trials undertaken has increased. As a result, the market of clinical trial management systems (CTMS) has also increased. In 2018, the CTMS market was scheduled to grow at a 12.6% CAGR to reach a market size of $1.06 billion by 2022. This estimate has now doubled to $2.186 billion due to the re-emergence and realization of AI.
As CTMS start-ups are helping optimize clinical trials using AI, two new companies that caught our attention are Trials.ai and Intelligencia.ai:
- Trials.ai, a San Diego-based start-up, founded by Kim Walpole and Brad Puitt, works with researchers to give optimized recommendations given clinical trial specifications. It uses data mining of trial-related documents to generate its recommendations.
- Intelligencia.ai, a New York-based startup founded by Dimitrios Skaltsas and Vangelis Vergetis, estimates potential risk factors to inform researchers. Like Trials.ai, Intelligencia.ai uses massive data mining. However, instead of developing recommendations, Intelligencia.ai assesses the potential success of a trial and the areas that need improvement before implementation.