It’s easy to get excited (and maybe a little nervous) about artificial intelligence (AI). In the simplest sense, AI is an advanced capability of machines to learn from what it experiences and make inferences from data – similar to how a brain processes information[1]. It’s a big shift from healthcare’s trial and error approach to apply AI’s vast processing power, ability to simulate different outcomes, and make decisions.
The prospect of quicker diagnoses, new treatments, reduced risk of medical errors, and voice-enabled advice is thrilling. Stories of the promise of AI are everywhere, but what does it mean for healthcare? And what will it take to get there?
Over the past three days, the 2018 World Medical Innovation Forum brought together senior healthcare leaders to share perspectives on AI. They discussed how cognitive computation, machine learning, and big data are starting to change how clinicians deliver care and how patients experience it.
Across healthcare – providers, payers, medical device, and pharmaceutical manufacturers – there have been some big takeaways:
AI is already providing diagnosis and treatment recommendations, such as diabetic retinopathy. As with many great innovations through the ages where skills have shifted, AI could impact diagnostic jobs – radiology, pathology, and more. Some say it will make them obsolete, others say it will be a shift toward augmented intelligence, where AI increases a person’s ability to accurately accomplish their tasks.
The volume of data, literature, and best practices is growing exponentially. A person is never going to be able to constantly evaluate dynamic changes in a patient’s wearable that may have clinical significance but a machine could. AI has a role in processing and making sense of disparate information and making clinical recommendations.
Traditionally, payers look at claims data to understand member needs. With AI, we can begin to address social determinants of health based on data outside of a health record – access to transportation, food security, social isolation, infrastructure, and more. For example, it may be possible to understand the causes of an individual’s medication adherence issues and then deploy personalized resources to help.
Right now, the reimbursement system incentivizes volume and complexity. But AI could help us shift from sick care to health and care. It could predict which people will become sick over time, and interventions could be implemented before a patient has an expensive, acute health crisis. This could save lives and reduce costs for patients and health systems.
High quality, well-curated data can help speed discovery by simulating the impact of a molecule on its protein target. With that information and data pulled from literature, AI could make suggestions about areas researchers should explore.
Right now, determining clinical trial sites and patient identification is a major challenge. It’s worse to find a lot of patients who don’t take the medication and drop out of the trial than it is to find a small group of the right participants. AI can be used to help analyze and provide recommendations that save time and reduce wasted resources.
Beyond algorithms, cutting-edge care and innovation start with the network. With AI, there will be massive amounts of data. The data needs to be stored – on premise, in the cloud, or in the multicloud domain. The data needs to be secure, because of its sensitive nature and privacy regulations. And the data needs to be used in real time – in a hospital, research lab, or in the field.
These requirements mean a provider, payer, or pharmaceutical organization requires a solid and modern IT infrastructure, and better, more secure ways to collaborate and share across diverse teams that could include clinicians, researchers, clinical trial participants, and data scientists.
One thing is clear – AI is here. It’s not part of some distant future, although we may be a ways away from scale. Want to explore more AI in healthcare topics? Watch the sessions from the World Medical Innovation Forum and learn about the role of the network in delivering innovation across connected healthcare.
[1] “Wielding the power of AI to improve healthcare,” Dr. Anne Klibanski and Dr. Gregg Meyer, World Medical Innovation Forum, 2018.
This is awesome information about AI in healthcare. I like this blog.
There is a lot of potential in AI in Healthcare, it is preemptive healthcare rather than acute.
Sourcing from many clinical trials and determining the best option for individuals will certainly be a great improvement and reduce cost in the long run.