Three areas artificial intelligence will impact healthcare in 2018

Benjamin Yu, M.D., Ph.D., Vice President, Medical Informatics and Genomics at Interpreta -

Artificial intelligence (AI), machine learning (ML), and deep learning (DL) are again top technology topics for 2018. In the past several months, these strategic technologies have rapidly gained traction in financial markets, big data analytics, and consumer devices to improve trend analysis, automation, and user learning.

Furthermore, the accessibility of AI tools to the general public has vastly improved with the introduction of AI-optimized neural chips and open-source AI solutions, e.g. TensorFlow (Google), Turi Create (Apple), and Gluon (Amazon, Microsoft).

While interrelated and often used interchangeably, AI, ML, and DL refer to distinct aspects of making machines smarter. Artificial intelligence, for instance, is the over-arching term that refers to technologies that mimic aspects of human intelligence, such as recognizing an image without significant instruction. Machine learning is the method of achieving AI, teaching AI via training on reiterations of input and feedback. Deep learning is a type of machine learning where artificial networks learn to discover patterns in data. Advancement of these technologies and their promise to help save costs and expedite discovery is helping to propel their adoption forward in healthcare and other industries.

How will AI impact healthcare, hospitals, and providers in 2018? Here are top opportunities and considerations for AI in healthcare.

Opportunities
Discovery: In 2018, AI will make a big impact on life science discovery and diagnostics. AI can help uncover patterns between millions of data points from multiple sources and correlate them with thousands of outcomes. Coupled with the emergence of deep molecular data such as genomics, proteomics, and other -omic big data, AI is already making its presence felt in diagnostics and biomarker discovery. Last Fall, biopharmaceutical company, Berg, in collaboration with the Department of Defense, claimed to have the first commercial use of AI in the discovery of disease biomarkers: network machine learning was applied to proteomic data in prostate cancer. In diagnostics, AI-based image pattern recognition has been reported to be effective in the early detection of melanoma and lung cancer. In the pharmaceutical industry and at a growing number of new startups, AI is increasingly being applied to drug discovery and accelerating compound screening. Industry leaders expect big things for AI in the pharmaceutical industry for years to come.

Clinical: AI-driven solutions will make an immediate impact in the clinic in 2018. Clinical optimization often involves streamlining workflows, improving net value for patient outreach, predicting engagement, and identifying synergies between risk factors for better disease management. Technology solutions such as those used by Interpreta capitalize on AI to help healthcare providers and payers prioritize high-risk members in a timely manner. Using AI to evaluate constantly-changing health parameters, quality gaps, and other health needs, this type of real-time automation can enable a relatively small number of healthcare workers to nimbly manage multimillion-member populations in a cost-efficient and coordinated manner.

Information technologies: AI will also impact health IT and complex data processes. One example of this is a recent collaboration among investigators at Google; University of California, San Francisco; University of Chicago, and Stanford. Together, they reported on the use of AI methodologies to identify patterns in electronic medical record systems. While this approach was primarily used for research and development, similar approaches could be used to accelerate real-time needs in the clinical setting. AI will be poised to provide early identification of high-risk conditions from multiple sources of health record data including claims, electronic health records, and biometric readings in the upcoming year.

Considerations
Health IT: With any new and disruptive technology such as AI, health IT concerns will be raised. One area of concern is where machine learning physically takes place, e.g. on a consumer device, a public cloud, or a private cloud. Each of these possesses new risks to healthcare IT because of the potential exposure of personal health data. Furthermore, AI is becoming more ingrained in consumer and mobile technologies, such as with facial recognition and natural language processing. Many organizations could be using such AI capabilities inadvertently and need to understand the extent of privacy and other IT risks.

Value: As AI-powered applications grow, savvy vendors are likely to tout AI capabilities. This poses challenges for providers and payers in evaluating which applications use AI to solve an unmet need or simply tout AI for its marketing value. Understanding how AI can improve healthcare efficiency and solve new problems, rather than using AI for AI’s sake, are key questions healthcare IT needs to address.

Implementation: Lastly, the biggest hurdle for AI in healthcare is implementation. AI use in diagnostics, new discoveries, and novel treatment guidance will face FDA scrutiny to properly address the true value and transparency of AI-based solutions. In the clinic, clinicians are already faced with vast heterogeneity of data and decisions. In order for AI to make an impact in 2018, implementation of AI must further streamline clinical workflow rather than burden already overtasked clinicians.

Conclusion
As AI, ML, and DL continue to improve and grow, they will undoubtedly make an impact in healthcare in 2018. Healthcare executives will need to learn how to best implement these technologies in their organizations for years to come.

References

Narain, Diers, Lee et al., Identification of Filamin-A and -B as potential biomarkers for prostate cancer, Future Science OA, March 3, 2017.
Esteva, Kuprel, Novoa et al., Dermatologist-level Classification of Skin Cancer with Deep Neural Networks, Nature, February 2, 2017.
Velazquez, Parmar, Liu et al., Somatic Mutations Drive Distinct Imaging Phenotypes in Lung Cancer, Cancer Research, July 15, 2017.

Benjamin Yu, M.D., Ph.D. is Vice President, Medical Informatics and Genomics at Interpreta. His company provides a real-time analytics engine that continuously updates, interprets, and synchronizes clinical and genomics data, creating a personalized roadmap and enabling the orchestration of timely care. It is online at www.interpreta.com. For correspondence, please email ben@interpreta.com

The views, opinions and positions expressed within these guest posts are those of the author alone and do not represent those of Becker's Hospital Review/Becker's Healthcare. The accuracy, completeness and validity of any statements made within this article are not guaranteed. We accept no liability for any errors, omissions or representations. The copyright of this content belongs to the author and any liability with regards to infringement of intellectual property rights remains with them.

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.