The convergence of artificial intelligence (AI) and the health care sector was inevitable. The advent of machine-learning and deep-learning technologies capable of analyzing and synthesizing massive amounts of data with algorithms designed to mimic human-level decision-making seems a natural fit for an industry in dire need of greater efficiency.

From improving the rates of drug discovery and clinical successes to reducing the costs associated with incorrect diagnoses and unnecessary treatments to helping patients avoid pricey procedures down the road through predictive monitoring, the possibilities for AI run the gamut across biopharma and med tech. And there's no shortage of enthusiasm – some might even call it 'hype' – as more and more companies, hospital and agencies are investing in AI platforms. But figuring out how to best integrate and use AI is still a work in progress.

In this 28-part joint BioWorld and BioWorld MedTech series, we take a deep dive into the trends, opportunities and challenges surrounding AI in health care and look for early signals to determine whether AI will be just another tool in the ongoing efforts for greater efficiency or whether it will result in a tectonic shift along the entire health care landscape, bringing drastic changes from the lab to the bedside.

On the rise

The series kicks off with parallel stories from BioWorld's analyst, Karen Carey, looking at the increase in AI-related dealmaking in both the biopharma and med-tech sectors. "Biopharmas move into deep waters of massive data, high R&D potential" details the rapid rise in the only the last few years involving AI-focused firms and technology, including some potential billion-dollar affairs, as firms seek machine-learning help in finding new drug targets in diseases such as Alzheimer's to identifying patients for clinical studies. In "AI propels med tech into new partnerships, new possibilities," Carey highlights recent activity in med tech, from identifying predictive biomarkers that could help prevent disease to creating AI-driven implants that could improve memory in patients with neurodegenerative disease.

BioWorld Insight Editor Peter Winter takes on the biopharma fundraising angle. In "VC meeting at the crossroads of biopharma R&D and computation," he notes the flurry of capital going into AI for drug development, including an excess of $1 billion in venture capital raised just in the past two years. Fundraising for AI in med tech, too, has been increasing at a strong clip, as Staff Writer Liz Hollis describes in "Digital health financings sees continued momentum."

And while uses for AI abound, their applications in clinical trials have been particularly compelling. In their respective stories, "AI in the clinic opens new vistas for biopharma investigators," and "AI starting to improve efficiency in med-tech clinical trials," BioWorld News Editor Michael Fitzhugh and BioWorld MedTech Staff Writer Meg Bryant offer their takes on an area in which AI is already starting to make a difference.

Nor is the interest for AI limited to the U.S. and Europe. In "AI is expected to drive health care effectiveness, increase jobs in Australia," Staff Writer Tamra Sami explains how Australia is making moves to adopt AI across the health care spectrum, from diagnosis to record-keeping. Staff Writer David Ho, meanwhile, looks at what's going on with AI in the Association of Southeast Asian Nations, which, though lagging behind its Western counterparts in terms of implementation and policy, is yet making early headway incorporating AI into drug discovery efforts, with Malaysia and Singapore leading the effort.

Rules and regulations, or lack thereof

One of the biggest challenges ahead will be adapting AI to global regulatory environments. Accustomed to operating within the technology space, AI platforms and those that create and use them will have to adhere to the more rigorous regulations of the health care sector.

Starting with the U.S., Regulatory Editors Mari Serebrov and Mark McCarty examine the regulations and policies facing biopharmas and med-tech companies, respectively. In "FDA looking to streamline drug development, approval process," Serebrov details some of the issues facing legislators, while a sidebar Q&A with the FDA's Sean Khozin highlights both the promise and challenges for the FDA when integrating AI into drug development. McCarty's "U.S. AI med-tech regs in development 'not for the faint of heart,'" meanwhile, dives into the agency's complicated regulations pertaining to AI.

Staff Writer Jihyun Kim takes at a look at Japan in "Japan to develop assessment system to speed up AI development." As a country with a well-developed and carefully regulated health care sector, Japan is by far the most advanced in Asia in terms of adopting AI. For instance, the Japanese government has pledged to invest more than $100 million to open 10 AI-based hospitals by the end of 2022.

In China and Hong Kong, the governments have yet to finalize any rules governing AI, according to Staff Writer Elise Mak, in "China on the cusp of regulatory AI-based devices," but the country has outlined a three-step strategy, with the aim of becoming a global center for AI technologies by 2030.

In Australia, the Therapeutics Goods Administration (TGA) has already decided to classify AI and machine-learning technologies under software as a medical device, or SaMD. In "Australia: No plans to regulate AI, machine-learning via separate pathway," BioWorld's Sami details current designations and uses for AI, while also looking at challenges faced by the TGA, a small regulatory agency struggling to keep up with changes in the global AI space.

In other regions, regulations are slow in coming. In "South Korea's insurance, personal information law slows use of AI devices," Kim describes the insurance coverage issue as the biggest barrier to adoption of AI technologies in South Korea, while Staff Writer T.V. Padma discusses how the lack of clear rules surrounding AI is frustrating health care innovators in India in "India lags in regulating AI as development proceeds."

Rethinking needed

Beyond changes at the regulatory level, the successful integration of AI will require rethinking the way industry operates.

For one, while the promise of AI is to do what human brains can – only better and faster and at a much vaster scale – humans still remain an important part of the equation, particularly humans with the skills to program, operate and understand the results generated by an AI platform. As Staff Writer Lee Landenberger notes in "Across the great (skills) divide," those are not the typical skillsets in demand in traditional R&D roles.

And the importance of having the right people is emphasized in Senior Science Editor Anette Breindl's story, "AI's superpowers are in questions, not answers," which looks at the scientific perspective. AI can analyze data on a large scale, but its biggest strength may come from asking better questions to generate predictions; those predictions still need to be validated in the lab.

Hospitals already are starting to use AI to address problems such as diagnosis and image analysis, according to Staff Writer Stacy Lawrence, in "Research hospitals push AI beyond pattern recognition," though she highlights one of the challenges is not overselling the abilities of AI.

And, lest you think technical usage is the only place companies can get tripped up by AI, there is the ethics debate and patent issues to consider. Staff Writer Nuala Moran delves into the former, in "Ethical issues trailing AI are writ large in health care," looking at some of the ethical issues that have from patient privacy, to incorrect diagnoses to biases that might be unwittingly embedded in the system, along with questions of who gets to use the data generated by AI and who gets to profit from it.

And, speaking of ownership, in "Supreme Court's Alice decision leaves drug, device firms in AI wonderland," Serebrov digs into a barrier facing U.S. firms using AI in their drug and device efforts: the highly problematic Section 101 of the Patent Act.

Wrapping up the series is "At the crossroads people and technology," in which Landenberger offers up some key takeaways and points to some of the problems the industry will need to resolve so that humans and AI can co-exist to provide the best outcome for patients and the global health care economy.

And companies will need to resolve those issue, because, as BioWorld analyst Carey notes in the opening piece of the series, there's one message on AI in health care that has become increasingly clear the last few years: This is the future. Will your company be ready?


Glossary of terms

Artificial Intelligence

Software platform comprising complex algorithms to analyze data, reach conclusions and anticipate problems with human-level expertise but without requiring direct human input.

Machine Learning

Continuous and repetitive process by which machines are capable of integrating new data and improving performance over time without the need for explicitly programmed instructions.

Deep Learning

A complex version of machine learning that involves multiple layers of abstract variables, for analysis at a scope and speed far beyond human capability.

No Comments