Machine Learning in pharma: drug discovery and clinical trials

Iflexion look at how different companies are utilizing machine learning to optimize their efforts

Back in 2013, McKinsey estimated that big data analytics and machine learning could bring up to $100 billion in pharma and healthcare annually, advancing R&D, improving the efficiency of clinical trials, introducing data-based decision support, enabling IoT and more. In 2017, another study found that 40% of the surveyed representatives from pharma and life sciences industries had already deployed AI.

Big data and analytics could bring in $100 billion to pharma and healthcare annually

In particular, drug discovery and clinical trials are in the focus. Not only are they complex, expensive, and lengthy processes, but also their outcomes are critical for pharma companies to deliver value and stay in business. And that’s how machine learning can help.

Related: Discover how to empower employees with data science


Next-gen drug discovery is underway

Pharma companies are consistently generating ideas for targeted and personalized therapies for a range of diseases. However, even good hypotheses may prove themselves wrong. AI and machine learning can accelerate this testing phase to limit investments in the wrong directions. Technology can also boost human creativity, by finding the subtle patterns and underlying causalities.


BERG’s Interrogative Biology platform

Boston-based BERG is a biopharma company with an AI-based platform called Interrogative Biology. This helps to identify disease biomarkers and use them to facilitate the discovery and development of drugs in oncology, endocrinology, and neurology.

BERG identifies disease biomarkets to discover and develop new drugs in oncology, endrocrinology and neurology

Traditionally, pharmaceutical companies target chemical libraries and hypotheses for drug discovery, but BERG approach it via the patient’s biochemistry and genomics, taking tissue samples from both healthy and sick patients and combining them with clinical data. This approach allows the company to focus on individual and targeted therapies for various types of cancers, neurodegenerative diseases, and endocrine disorders.

BenevolentAI’s AMD treatment research

BenevolentAI uses machine learning for target identification, molecular design, and clinical mechanistic stratification to advance therapy discoveries based on biomarkers. They also support clinical trials by analyzing individual patients’ responses to treatment.

Just recently, BenevolentAI announced their next target for treatment discovery – age-related macular degeneration (AMD), the leading cause for blindness in the UK. The company aims to aggregate, process, and mine millions of scientific abstracts, articles, medical scans, and medical research data related to AMD. This will allow for a comprehensive condition picture, highlight the prevention measures, and identify the areas for further R&D of a possible cure.


Ensuring the success of clinical trials

Although millions is invested into the research and discovery of a new drug, it may not be able to prove its efficiency or safety during clinical trials. The FDA reports that only 1 in 10 of medication tested on humans receives its approval.

It's therefore important for pharma companies to optimize clinical trials, refine patient matching mechanisms, and make the process safer for patients to increase the likelihood of approval.


Antidote connects patients with fitting studies

Antitdote seeks to accelerate medical breakthroughs for pharma companies by helping  patients discover relevant medical research studies to assist their conditions.

Antidote operates a machine learning-powered clinical trial matching platform. Using algorithms to review patient eligibility criteria and partnerships with non-profit organizations and healthcare advocates, companies can reach out to eligble, engaged and informed patients.

Antidote's machine learning powered platform matches clinical trials with eligble, engaged and informed patients

This increases patient enrollment and allows pharmaceutical companies to meet trial deadlines and successfully test their treatments of a diverse subject set. for all-around clinical trial management is an AI platform crested to assist researchers in managing clinical trials. The system allows operators to design studies, optimize protocols, monitor risk factors, and improve patient adherence within one system. The platform also analyzes medical abstracts and past clinical trials results to streamline and accelerate the process. has already worked with the Moores Cancer Center to cut their study’s timeframe by approximately 33%, while reducing data errors by 20%.

Read More: The FDA's top regulator believes the clinical trial system is broken


In anticipation for more tech-driven medical breakthroughs

Following in the footsteps of the value-based care and patient consumerism, the pharma industry is now investing in AI and machine learning development to improve their existing processes in a faster and safer manner.

Companies can now discover new treatments rooted in the patient's own biology and utilize underlying patterns to find new cures. The clinical trial process can be connected, analyzed and reviewed to minimize risks and give treamtents their best hope for approvals.

There is still significant room for development in this field. Currently the biggest roadblock to AI and Machine Learning adoption in Pharma lies in the scarce knowledge of both what is available and how to benefit from these technologies. If the pace of development continues, it is hoped that this barrier will be broken within the next five years.