Happy New Year: Digital Health perspectives on AI / Cognitive Computing for 2018


As we close out and retrospect 2017, it is clear that artificial intelligence (AI) and cognitive computing were major themes. For Digital Health executives, entrepreneurs, and experience designers, below are some great resources we shortlisted. We hope it makes it into your 2018 strategic thoughts.

  1. Value-based care (VBC) + AI on a continuum. Much of 2017 was a transition year for U.S. providers and payers understanding how to transition to newer VBC models. As physician leaders, Toussaint and Halamka have both deployed innovations effectively that have improved: population health, outcomes, and cost reduction. Toussaint has deployed VBC models at scale using core Lean principles, while Halamka runs BIDMC innovation like a lean startup targeting pragmatic care problems. In his last quarterly HIT policy update, he highlighted on pragmatic use of AI within exam rooms. Imagine asking “Where is Doctor Smith?” and getting answers back.
  2. The Patient-Physician interface. AMA’s Madera re-emphasized physician-led change in order to tackle U.S. Healthcare’s $3T costs with “puddles” of interoperability in an ocean of siloed big data. His vision of AI focuses on the patient-physician(s) interface with care coordinated by cognitive computing. He clearly emphasizes how physicians can improve care via intelligence augmentation (IA) versus EHR data entry tasks. In 2018, those who design the AI-powered experiences well will win.
  3. Real world AI via cognitive computing. This year, many Digital Health startups launched using AI as a disruptive technology, while industry conservatives such as medical associations strategically positioned for relevance. This ranges from Andrew Ng’s Woebot as a mental health digital platform or AMA’s Health2047 portfolio companies such as Switch. Davenport, et al. anticipates mainstream going real-world AI and is a great synthesis of previous Deloitte and IBM thought leadership on cognitive computing captured in a MOOC (October 2015) and enhanced with recent field interviews. The punch line? Focus on automating non-value added human tasks to free up more cognitive value add.
  4. Going deeper and staying relevant. Most of the recent AI hype is within an area of machine learning (ML) called deep learning. In 2016, Mount Sinai clinical researchers proved real-world use of deep learning for EHR use cases and representing a “deep patient”. If you are feeling out of the AI loop and want to dig deeper, Andrew Ng’s deep learning specialization offers high-yield content for a wide audience. This has been updated in August 2017 and reflects Ng’s AI industry experiences after returning from Baidu. If you are business-focused and want a quick breadth level pass, auditing his “Heroes of Deep Learning” interviews quickly surveys recent AI research and industry trends. If you are technology-savvy and want the details, five-course modules full of Python notebooks with neural networks await you.

At OA, we advise on cognitive computing strategy and implementation in Healthcare. One of our innovation efforts on the use of conversational agents using IBM Watson versus Microsoft Cognitive technologies got noticed here. Curious? Drop us a line. Look for more relevant posts in the new year.