An article written by Fabio Caversan, originally posted at Forbes.
Today, AI is used with increasing regularity across nearly every industry, with AI-based systems and technologies introducing new efficiencies, unlocking extraordinary opportunities and delivering powerful new insights and capabilities that were previously unattainable — perhaps even unthinkable.
Not only are health care and pharmaceuticals no exception to that rule, but life sciences actually represents one of the most innovative and exciting new frontiers for AI technology and machine learning. In recent years, AI usage has exploded in pharma, health care and biotech. Life sciences companies and institutions have used AI to develop and test new drugs, advance new therapeutics and treatment protocols, and, in some cases, completely transform the drug development and distribution process.
A Virtual Lifeline
The power and potential of AI-based technology in life sciences has arguably never been more important. The value of safe and speedy medical innovation is clear, especially since the critical work of pharmaceutical companies and healthcare organizations has been significantly disrupted by the Covid-19 pandemic. Early in the pandemic, at least 440 clinical trials in the U.S. had been halted in the face of logistical difficulties, safety concerns or elevated exposure risks to participants.
These urgent new challenges can be particularly difficult to overcome for a process that was already plagued with structural inefficiencies: coordinating multiple systems, products and vendors, siloed applications yielding siloed data, error-prone manual processes and complex environments affecting collaboration, data flow, data access, analytics and reporting.
Some cutting-edge solutions offer comprehensive AI-powered platforms that allow organizations to transform how they conduct and monitor clinical trials. They include everything from digital study design tools, to options for conducting smart, hybrid and virtual trials, to the ability to provision devices and concierge on a unified platform — complete with automated data collection and orchestration, analytics and cognitive AI.
Research, Big Data And Clinical Applications
In drug discovery and research, an overwhelming amount of data is generated. With machine learning, extraordinarily large and complex datasets can be processed through AI and recursive training methods to yield insights and critical information. At virtually every step of the development process — from preclinical drug discovery (screening thousands of molecules), to initial lead screening and lead optimization, to target selection and target identification — AI-based tech can save enormous amounts of time and money.
On the clinical side, AI is also making new breakthrough tactics and treatments possible. Genentech has used machine learning to process vast volumes of data from cancer patient data through sophisticated computer modeling to identify new cancer therapy targets. Familiar names like Cyclica and Bayer use AI technology to sift through large numbers of candidates and identify promising pharmacological profiles. And Bayer and Merck & Co. were granted the FDA’s Breakthrough Device Designation for AI-powered software designed to facilitate better clinical decision-making for CTEPH, a serious and often misdiagnosed pulmonary condition.
Transforming The Cost-Benefit Equation
The traditional drug discovery and research process is a costly and time-consuming ordeal. Identifying the right molecular target for research — or the right compound or combination of compounds with the therapeutic promise — can take decades, consuming hundreds of millions of dollars.
Even advancing to the trial phase is no guarantee of success. A study published by Biostatistics found that less than 14% of drugs pass clinical trials. The financial fallout from that inefficient, labor-intensive process is considerable. From research and clinical trials to FDA approval, the median cost of bringing a drug to market approaches a staggering $1 billion — and can easily double that. Because new AI solutions can recognize patterns, draw connections and identify potential targets with minimal manual intervention — much faster and more accurately — AI is a potential bottom-line game changer. Drugs get to market faster and more affordably, and resources previously dedicated to conventional drug discovery processes can be invested in new therapies and improved patient care.
One of the most fascinating AI frontiers in the life sciences is the expanding ecosystem of social listening technologies and solutions. Social listening — monitoring social media channels for brand and product mentions, competitor activity and other relevant information — has exploded in popularity in recent years.
For life sciences organizations, social listening isn’t just a vital way to monitor and understand the sentiment around their brand but a tool that can address a range of impactful issues:
- Identify customer needs.
- Screen for patient identification and selection for trials and therapies.
- Identify geographic areas or populations that are potential trial candidates.
- Monitor disease prevalence or symptoms.
- Conduct competitor analysis.
- Patient engagement and retention by tracking experiences and perceptions over time.
A Bright Future
It’s important to mention that, although these exciting AI possibilities are out there, taking advantage of them requires care and attention. Organizations considering AI solutions — or those facing challenges collecting results from AI initiatives — should do the following:
1. Find partners or build teams with knowledgeable and experienced experts not just in AI but in life sciences and health care. This field requires more human-centered thinking than others, and focusing too heavily on technology can lead you astray.
2. Recognize both the possibilities and limitations of current technology. Start slowly by automating simple processes before searching for cures for novel diseases.
3. Prepare the data. Most of the real-world examples referenced above are data-hungry solutions. Start to think about how much structured data you have. Even unstructured data may provide good material to work with.
4. Prioritize explainability in AI solutions, especially if your goals involve making or supporting decisions. In this particular area, it’s important to understand the reasoning behind AI decisions in order to achieve confidence in the process.
For pharmaceutical and health care organizations grappling with the challenges of a global pandemic, a competitive marketplace and daunting logistical and financial realities, AI-powered technologies offer something rare: a true paradigm shift in the way whole swathes of industries operate. From research and development, to trials and treatments, to long-term monitoring and engagement, AI isn’t just opening new doors; it’s unlocking new opportunities — at the precise time when they are needed most.