“There will be abuses of power that involve AI, just as there will be advances in science and humanitarian efforts that also involve AI. Unfortunately, there are certain trend lines that are likely to create massive instability.” 

Danah Boyd, Principal researcher for Microsoft

Wisdom involves applying knowledge and using judgment and insight for relatively nuanced and complex situations. This highest level of the data-to-wisdom hierarchy is still a human endeavor. Perhaps with “super” information resources like the large language models that have recently arrived, knowledge and intelligence as well as wisdom can be expedited in healthcare to improve the Quintuple Aim.

AI in biomedicine failed in earlier eras, mainly due to a lack of computational speed and big data. The result was an undesired disruption to clinicians’ workflow. This disadvantage of slowing down the work process coupled with a lack of large volumes of data created unfavorable deployment. The clinicians back then preferred not to use these AI tools even though the performance of was good and even superior to groups of consultants.

The recent re-emergence of artificial intelligence this past decade after two previous “winters” resulted from a convergence of advanced algorithms, cloud computing, and computational power as well as big data. These elements all improved dramatically in the past decade, and this convergence has accelerated the progress of AI as well as raised the expectations of AI in healthcare for the future.

Most healthcare data is unstructured (without tables with rows and columns) and has been synthesized only in the last 5-10 years. Data is the foundational layer of the data-information-knowledge-intelligence-wisdom pyramid. Good data science and artificial intelligence projects will need good quality data that is relatively complete and accurate. Big data (volume, velocity, variety, and veracity) is exponentially increasing especially with genomic data and real-world data.  

Real-world data (RWD) and real-world experience (RWE) are becoming more relevant in clinical research. They can be very complementary to traditional data collected in randomized controlled trials (RCT) and clinical research registries. While RCTs focus on efficacy and safety with an experimental type of study with intensive monitoring and follow-up, RWD and RWE focus on effectiveness and value with an observational study with variable amounts of monitoring and duration of follow-up.

Most time on a healthcare ML/AI project is spent on data access, collection, and processing. While the algorithm design and model deployment can also be time-consuming, this later step is less cumbersome than the initial phase of curating the healthcare data. The very first step of the ML workflow, however, should always be the formulation of a clinically relevant question with potential impact (avoiding type III error- testing the wrong or irrelevant hypothesis). 

The importance of these observations of artificial intelligence in healthcare will be part of the topics of discussion at the in-person AIMed Global Summit 2023 scheduled for June 4-7th of 2023 in San Diego. The remainder of the week will be other exciting AI in medicine events like the Stanford AIMI Symposium on June 8th. Book your place now!

We at AIMed believe in changing healthcare one connection at a time. If you are interested in discussing the contents of this article or connecting, please drop me a line – [email protected]