Five Reasons to Embrace Data-Driven Drug Development

Five Reasons to Embrace Data-Driven Drug Development

Five Reasons to Embrace Data-Driven Drug Development

The growth of the pharmaceutical and biotechnology industry depends on successful clinical trials. The cost of developing a new drug that gets approval is estimated at around $2-3 billion. The longer the trial lasts and the more patients it has involved, the larger the loss. The researchers at Johns Hopkins Bloomberg School of Public Health observed that trials which failed at phase I or II wasted around $6 million, while those failed at phase III resulted in a loss of approximately $77 million. Not to mention, the time it takes for an approved drug to reach the market is 10-12 years.

 

Drug discovery and development timeline

Image source: www.yourgenome.org

 

The estimation of $2-3 billion includes a large amount wasted on failed drugs. It has been observed that 9 out of 10 clinical trials fail. However, the technique pharma relies on is a simple equation – more compounds in trial phase = higher the probability of success. This means, to get one approved drug, multiple molecules are put into clinical development, requiring large expenses spent on R&D. A data-driven approach is, therefore, the key to accelerating research and development for more successful clinical trials, both in biotech and pharma companies.

 

1) Big Data in Pharma is a treasure trove of unknown information

 

When pharma faced the Big Data explosion it was a perfect opportunity to unearth previously hidden data. However, pharma didn’t know how to manage the exploding volume of data. Data silos naturally formed and data monopoly was achieved by the big players. Today, we have plenty of data that we can leverage. With time we learned to tame the tide of the Life Sciences data ocean, including streams of data from research papers, clinical trials, patents, medical records, etc.. This data has information and insights from some of the top scientists who have spent their lives developing unique drug solutions. Leveraging the data to identify new connections, anticipate unknown side effects, and enhance molecule structures ushered in a revolution in pharma.

 

The data-driven drug development approach focuses on leveraging Big Data of Life Sciences and Pharma to accelerate the creation and identification of novel therapies and to realize this potential, the use of Artificial Intelligence is essential.

 

2) Faster insights can be generated with a data-driven approach harnessing AI

 

With so much data to analyze in such a short time, and under tremendous pressure, only innovative technologies, such as AI, can help leverage this influx of insights coming from publications, congresses, medical records, thesis, dissertations, patient data, trial data, and many other sources. Densen estimated that by 2020 Life Science data will double in just 73 days. For example, a new immunology paper is published every 30 minutes. Considering that a researcher takes one hour to read and annotate a research paper and additionally, 2.5 hrs, i.e. 600 hrs per year, on manual information research, the time that could be spent to generate new, useful insights is spent curating facts rather than generating novel theories for better therapies.

 

Lawrence Ganti, the Global President of Life Science at Innoplexus, noted, “With an increase in data, there will be an increased need to analyze data and do it fast. This is when pharma companies will need computing power, AI, and machine learning.”​ Already, companies such as Pfizer, Roche, and Sanofi have not only embraced the data-driven approach but also learned to manage big data in pharma using AI technologies such as computer vision, entity normalization, network analysis, machine learning etc.

 

Lagging behind at this stage in the boom of data-driven drug discovery could lead pharma companies to become extinct. As faster modes of discovery are being enabled by leveraging life sciences data using AI, pharma is moving towards changing its game in the drug development landscape.

 

3) Data holds the key to discovery

 

In the year 2007, AI caused a scientific discovery for the first time when a robot named Adam identified the function of a yeast gene. It did so by searching public databases and afterward using robotics to physically test it in a lab. The use of robotics is not yet welcomed in labs all over the world. However, when it comes to searching public databases, that should be doable for any company now! With AI by pharma’s side, today, searching vast amounts of life sciences data is not only possible, but it’s also feasible to analyze relationships between biological entities such as genes, proteins, diseases, drugs etc. and form novel therapies based on thus obtained research graphs. Understanding and verifying entity relationship is the most difficult work to do manually, especially when tens of thousands of molecules are involved. New opportunities for better therapies and interventions surface through with the revelation of unique connections between biological entities, thereby, accelerating research. It can assist in the identification of new targets, which will enable drug repurposing as well.

 

A data-driven approach can increase research efficiency, propel lead and target discovery, redefine repurposing, and result in efficient clinical developments.

 

4) Data-Driven Drug Development: faster, more accurate

 

A drug usually hits the market, if approved, in 10-12 years. This long timeline is divided into various stages; drug discovery, pre-clinical, and clinical phases, before the commercialization starts. For faster development, it is necessary to reduce the time taken at every stage. While data-driven drug discovery could lead to better target identification, it can also shorten the pre-clinical timeline. Researchers spend days testing various molecules to be sent to the clinical stage. A data-driven approach can change the dynamics of drug development. So, instead of experiments, researchers can optimize the chances of success of a molecule by using data repositories effectively.

 

Gunjan Quote

 

Data is there, full of previously untapped potential, unexplored connections, and relevant insights. All one needs is to know how to use it efficiently. Data can provide information on the success rate of similar drugs in a particular indication (eg. lung cancer). It can also provide insights useful to accelerate patient enrollments and map out previously successful sites for selected indications. This would result in more successful clinical trials. Apart from this, the identification of key opinion leaders could lead to better research collaborations, good investors, and successful commercialization. Analyzing social media and patient forums for sentiment analysis could provide an overview of side-effects post drug launch. Insights that would lead to the identification of unmet needs would help in better pipeline prioritization.

 

5) Better Drug Repurposing

 

Drug repurposing is the fastest way to bring cures to the market. By analyzing trillions of data points hidden in vast repositories of life sciences data ocean, and mapping all the connections through research graphs, anything can be made discoverable. This is where pharma can leverage the data-driven approach the most. Through technologies such as network analysis and machine learning, using life sciences domain ontology (instead of just a subdomain ontology, ie. gene ontology), more relevant data can be generated in real time. This data will help in repurposing drugs with minimal experimenting and testing efforts.

 

Ganti believes “Eventually, the pharma industry will see more drug repurposing and companies will do more with what they have because AI can identify the subset of the population for whom a drug will work for.”​ He added, “Whether it’s a failed or new drug, pharma companies will move toward approving drugs in specific disease areas with certain biomarkers.”​

 

Additional benefit of adopting the data-driven approach: No more data silos

 

The data-driven approach will not only accelerate drug approvals but also encourage data sharing. More data means more meaningful insights, better training data for machine learning to analyze it, and considerably valuable business decisions. As and when pharmaceutical companies realize the value of unshared Life Sciences data, silos will be a thing of the past. With the use of blockchain technology, companies would share data from failed or abandoned clinical trials, offering insights on what had worked and what did not. This would reduce the doubling of efforts and will offer more opportunities for redefining strategies in drug development.

 

The data-driven approach is not just about faster developments or better research. It’s about utilizing the information we already have in one form or another, through any source whether from medical devices, publications, or IoTs, and leveraging it to increase ROI through every data-generated opportunity.

Author

Lawrence Ganti

Lawrence Ganti is the Global President of Life Sciences, CEO of  Americas, and Member of the Global Executive Community. Prior to joining Innoplexus, Lawrence spent 15 years leading teams across functions and geographies in the pharmaceutical industry. He also worked as a research associate for McKinsey & Company and holds an MBA from IMD in Switzerland and a bachelors from Babson College in the United States.

Download Technology Capabilities
close slider

Download Technology Capabilities