The Future of...

The Future of AI-Driven Health and Drug Discovery

New insights depend on interconnecting the data dots

Tim Waters
The Future of AI-Driven Health and Drug Discovery

We tend to think innovation is about creating something new. But more often than not, great breakthroughs occur when existing dots are connected. For example, penicillin had its start in 1928 when Alexander Fleming returned from vacation to discover a mold had contaminated some of his bacteria cultures, and it was preventing normal growth of the bacteria. But it didn’t gain any real traction until a decade later when Howard Florey came across Fleming’s paper on the penicillium mold in a medical journal. He began work on developing penicillin with his colleague Ernst Chain, and the first human patient was successfully treated with the drug in 1942.[i] It was a series of dot connections that led to one of the most important medical breakthroughs of our time.

That’s not unlike the process where artificial intelligence (AI) and machine learning (ML) are leveraged to connect “data dots” to yield new insights. But to effectively uncover relationships between different knowledge domains, these technologies require access to large, diverse data sets. That can be a challenge in industries like health and pharma where data is collected and stored in different places and considered to be highly sensitive. But by leveraging an interconnected distributed data architecture, participants in digital health ecosystems such as providers, insurers, governments, researchers and more can share patient information safely and compliantly.

Download the Healthcare Digital Edge Playbook

Get the playbook that outlines how industry leaders can use interconnection and colocation (control points) at the digital edge, in proximity to patients, regulators and partners, to create a foundation for digital healthcare.

Download Now

What the future of AI-driven health looks like

Technologies like AI and cloud computing not only enable new connections between knowledge domains but also between physical entities and events, such as a physician performing surgery remotely or a medical device detecting a critical care event and alerting nurses and on-call physicans. A recent Optum survey of 500 U.S. health industry leaders found that 62% of respondents have an AI strategy in place, nearly double that of the previous year.[ii] There are a number of areas where AI is already helping to advance positive outcomes in the sector. For example:[iii]

  • Alibaba trained an AI model to detect the coronavirus in seconds with 96% accuracy.
  • Developed in just 12 months, an AI-designed drug to treat obsessive-compulsive disorder (OCD) will enter human clinical trial for the first time.
  • MIT researchers discovered a new antibiotic able to kill 35 drug-resistant bacteria through a deep-learning algorithm.
  • The FDA has cleared GE Healthcare’s AI platform for X-ray scans that can help radiologists prioritize cases involving collapsed lungs.

Suki AI is a voice-enabled digital assistant for doctors that records notes dictated by medical staff during a patient visit and then automatically fills out electronic health records (EHRs) based on those notes, vastly accelerating medical transcription.

Growing adoption of AI by the health and pharma industries will continue to fuel innovations like these but access to data is key to their success.

Access to standard data is a key challenge

AI is only as good as the data you feed it, and that’s a key challenge for the health industry. Although EHRs, medical journals and other digital data sources have been around for quite some time, historically they have been siloed in centralized IT infrastructures. And much of the critical patient data has been locked in different payer and provider systems in various states and formats. Future breakthroughs will depend on this data being shared, securely and compliantly, between patients, devices, health providers, insurers, governments, researchers and more. This requires a distributed interconnected ecosystem that can speed collaboration and secure data exchange between digital health ecosystem participants. The Global Interconnection Index (GXI) Volume 3, a study published by Equinix, outlines five simple steps that participants can take to achieve this digital-ready state:

  1. Network optimization: Shorten the distance between users and services by creating a digital supply chain of interconnected hubs at the digital edge across providers, payers, research/pharma companies and governments.
  2. Hybrid multicloud: Locally and directly interconnect clouds and partners and segment network traffic in the hubs for improved performance and low latency.
  3. Distributed security: Deploy and connect security controls and countermeasures in the hubs for real-time transparency and control across the digital landscape. Leverage private interconnection to avoid security risks inherent in the public internet.
  4. Distributed data: Place analytics, data lakes and data controls in the hubs to integrate data pipelines, manage massive data volumes and generate insights. Leverage healthcare and life science (HCLS) standardized data models based on existing health alliance architecture to remain compliant.
  5. Application exchange: Participate in interconnected digital health ecosystems to harvest, process and exchange patient and other medical data, generate timely insight and expand the value chain with partners and data for superior outcomes and experiences.

As the secure exchange of health data grows, it will pave the way for medical breakthroughs we can only imagine now. As a colleague of mine recently said, “The cure for cancer is already out there – we just have to find it.”

Learn more about how healthcare and pharmaceutical companies are leveraging interconnection to drive better patient outcomes.

Avatar photo
Tim Waters Sr. Manager, Equinix Research Group for FED, SLED & Healthcare and Life Sciences
Subscribe to the Equinix Blog