Between last year’s and this year’s conference I think it is clear that we are starting to come off the peak of the hype cycle regarding impact of AI and Digital Health.
There was a definite focus on implementation of perhaps less blue sky but more immediately practical solutions to improve functioning and efficiency of healthcare delivery such as using NLP to capture consultations and automate EMR upload or using AI to reduce unnecessary readmissions. There was lots of talk around ensuring solutions fit workflows and practices. There was much talk about the emergence of Natural Language Processing being key to many of these solutions.
There is a growing sophistication in the quality and output from small and medium tech businesses and startups. I think the market is becoming far more discerning and less accepting of more home grown products that are not reliable or scalable.
Challenges with outdated governance, ethics and privacy processes and increasing public scrutiny and lack of trust are an international issue and a major break on the development of the technology sector.
There was a growing acknowledgement of the need for platforms that handle live and identified health information that can be used for decision support, QI etc. I spoke to the CEOs of GE and Siemens who are all building such platforms and they acknowledge that they need to collaborate to enable cross talk between systems. I anticipate there will be a small number of large multi-partner plays in this area.
In terms of decision support and development of AI applications there was a focus on integrated solutions to head of the emergence of multiple single point solutions all hitting services at once. As in our DHCRC workshops, there was a focus on the continuum of data and information capture across our lives and care contexts and the need to use this data proactively. As expected there was much talk about the need to capture social determinants of health data and integrate this into managing health and wellness. Interestingly there was talk about the failure of many small AI companies and the need for consolidation.
There was much debate regarding the future of clinical trials in terms of access to real time data from EMRs allowing for pragmatic or real life trials and reducing the need for classic randomised clinical trials. There was consensus that there was still the need for randomized trials and there was insufficient data or methodologies at this point to realistically move away from trials. A number of the CEOs of drug companies at the conference emphasized they were becoming data as much as pharmaceutical companies. I think we are going to continue to see interesting and disruptive mergers and acquisitions between big players moving into convergent markets.
It is clear that China is emerging as the front runner in terms of investment and innovation in healthcare. This is driven both by necessity, different privacy and governance rules and massive cash investment into research and development. Another key advantage China has is that they are largely building entirely new health systems so they are not facing the issues we encounter when trying to re-tool existing systems.
At the Clinical Learning Analytics Network meeting at Partners Healthcare attended by colleagues from Harvard, Stanford, NYU and University of Toronto there was much talk about the missed opportunity in EMRs to support behavior change beyond crude alerts and the need for new approaches to EMRs that embeds learning and links to maintenance of competency programs.
A number of the panel discussion are available from the conference if you are interested:
CEO Panel (Siemens, Amgen Phillips and GE) https://www.youtube.com/watch?v=OSJHoeV3pmQ
AI Investment Leaders https://www.youtube.com/watch?v=hvT3GqriTLU
China panel https://www.youtube.com/watch?v=U7TH6-SAho8
Trial optimization in the AI era https://www.youtube.com/watch?v=nwbgg2-Tt3s