In 2018, we dove into the topic of AI and machine learning in healthcare

While AI/ML are critical tools for solving big problems in healthcare, we’re concerned by hype causing confusion in the space, and by solutions being developed without sufficient regard for their efficacy and business value. We built on our white paper, Demystifying AI and Machine Learning in Healthcare, to offer perspective on what the launch point of AI in healthcare really is—where it’s ripe for use, and where we won’t see movement for years to come.

Attendees came away with:

  • An evaluation framework for AI/ML algorithms

  • Understanding of the major venture trends in AI/ML-powered digital health companies

  • Tools to manage major threats to adoption and scale

  • Practical guidance for overcoming hurdles such as dealing with dumb and dirty data, incorporating AI into an enterprise business strategy, and navigating regulatory uncertainty

  • A new network of companies and academics with the most promising solutions

 
 
AIML-Spectrum-1200x696.png
 
 

Read our blog post about the report—and for full access to all of our research (including our proprietary funding database of every digital health deal since 2011), get in touch.


past agenda highlights

Brodgesell_RockHealth_Enterprise2018_065.jpg
Brodgesell_RockHealth_Enterprise2018_451.jpg

 Is AI Ready for its bedside debut?

A bounty of new studies suggests machine learning algorithms are more effective at diagnosis than physicians are, but machines are not diagnosing real patients—yet. Providers are fairly asking some tough questions of these tools as they relate to care delivery—are they effective, will I get paid, and does this put me at greater risk for malpractice? We’ll discuss these implementation hurdles and lay out the timeline for today’s solutions to fully realize the opportunity they represent for patients and providers—lower costs, faster service, and better treatment.

Emerging ethical issues for healthcare in the age of ai

Despite all their promise, algorithms trained on data which may contain unforeseen omissions or biases can result in subtly—and perhaps dangerously—biased predictions. In this session, we’ll explore the ethical issues algorithms pose, including: what margin of error will be allowed for computers in healthcare? What rights do patients have with regards to the data and algorithms involved in their care? Experts will advise on how to proactively surface and mitigate bias, as well as offer a framework for addressing unresolved ethical issues as a community.

The enterprise playbook for identifying breakthrough innovation

It has never been more crucial for enterprise to identify which startups are both optimally using AI/ML and ready to scale within healthcare. Yet the path forward for effective integration between startup technology and enterprise business lines is overwhelming—leaders must decide which type of contract is most helpful in managing “readiness risk,” disentangle the different validation signals—from the FDA to academic studies— to assess a startup’s potential to scale, and evaluate compatibility with their own internal data stack and workflows. Investors and enterprise leaders will discuss how companies should navigate the AI/ML startup space in order to differentiate between hype and real business value.

 
 

WHAT AI REALLY MEANS FOR HEALTHCARE BUSINESS LEADERS

 
 
Brodgesell_RockHealth_Enterprise2018_256.jpg
Brodgesell_RockHealth_Enterprise2018_405.jpg

what attendees learned in 2018

Attendees may not have left as an academically trained technologists (that is, unless they came as such), but they were prepared with the tools needed to tackle the most pressing challenges facing their healthcare business in the age of AI. Attendees left the room knowing how to:

  • Internally structure themselves to integrate AI solutions into their businesses

  • Think proactively about the ethical implications of AI specific to their businesses

  • Assess the startup ecosystem and separate hype from practical use cases

This event aimed to support enterprise business leaders—including those at hospitals and health systems, pharmaceutical companies, payers, and technology companies—in navigating and leveraging AI/ML solutions. Here’s a breakdown of what attendees, by type, came away with:

  • Pharma, Life Sciences, and Devices

    • How entrepreneur Iya Khalil of GNS Healthcare is making precision medicine a reality by leveraging causal machine learning for drug discovery

    • Perspective from ZS Associates’ John Piccone on working alongside pharma and med device companies using AI to bend the cost curve

    • A look into GSK’s AI-focused group from John Baldoni on how the company has set itself up for the adoption of AI and successful partnerships with startups

  • Providers

    • How entrepreneurs from leading startups, such as Fabien Beckers of Arterys, work in tandem with physicians to earn their trust and leverage algorithms actively used for patient care

    • Learnings from Dr. Robert Watcher of UCSF about how he selects the right opportunities to plug in ML solutions across his organization—and the impact it’s had so far

    • Strategy for implementing AI-powered bedside workflow tools, shared by Tom Cassels of Leidos Health

  • Payers & Self-Insured Employers

    • Guidance around the practice application of machine learning to optimize risk stratification and map your members to appropriate interventions

    • An approach to creating a framework for addressing unresolved ethical issues as a medical community, from Ian Blumenfeld of Clover Health

  • Tech

    • Thoughts from prominent entrepreneur-turned-venture capitalist Vinod Khosla on how he evaluates potential investments in the crowded AI/ML space

    • A case study from renowned mathematician Gunnar Carlson of tech companies’ role in supporting enterprises in automating previously manual processes using machine learning

    • A playbook of how tech companies are becoming “AI-first”—and why so many are entering the healthcare space

  • Academia

    • How private-public-academic partnerships are addressing the elephant in the room: the need for robust, well-structured, clean data sets to train models

    • An explanation of the gaps and limitations of startups and enterprise companies—and the opportunity for academia to plug in and accelerate the advancement of algorithms, data sets, and full-stack solutions

    • First-hand accounts of creating successful partnerships with startups—such as Brandon Ballinger of Cardiogram and UCSF’s joint effort to study DeepHeart, a deep neural network intended to predict cardiovascular risk