Leverage AI to revolutionize and advance healthcare
What is Intel doing in the area of artificial intelligence/machine learning?
Artificial intelligence is causing a technological revolution. Intel recognizes the power AI has to transform society and industries. We are committed to democratizing AI and machine-learning innovations so that everyone has the opportunity to benefit. To that end, we’ve been doing a number of things:
• Understanding the importance of integrating hardware and software advances to create AI experiences, we have invested in acquiring companies that are innovating around the hardware and the software driving intelligent applications – companies such as Nervana, Saffron, Movidius and Mobileye.
• We launched an AI products group that offers a broad palette of AI technologies to meet diverse needs. This group focuses on solutions that make it easy to incorporate custom AI solutions into existing infrastructure. These solutions let us create intelligent ecosystems that seamlessly connect applications and devices across the spectrum, allowing people and businesses to make better use of audio, video and text wherever they are.
• Our research and software teams continue to contribute to the software foundations of AI and machine learning. Our software optimization experts work with customers and partners to build powerful, high-performance software for creating AI applications. These teams have made significant contributions to many of the important software frameworks for creating deep learning-based AI solutions. For example, we recently released our BigDL project,
an open-source, deep-learning library offering users a framework for creating deep-learning systems that run on a cluster of standard Intel servers.
How will this technology advance healthcare delivery?
AI offers the opportunity to give healthcare teams and patients access to relevant information in a timely fashion. Care teams have to process a lot of information quickly. They may be proficient at synthesizing the information they have, but they may not be able to quickly and easily find the data they need. That’s where AI comes in. As anyone in a care team knows, healthcare delivery is often hampered by disconnected siloes of care and information. Patients move from care site to care site, but information about their clinical history may not move with them. AI is playing an increasingly important role in bridging those gaps.
And AI doesn’t just benefit clinical teams, it helps patients, too. A contextually intelligent agent (CIA) is one example. A CIA system can guide a patient through the complexities of the care delivery system. Patients interact directly with CIAs, asking questions and receiving recommendations as the care situation evolves.
How are hospitals leveraging and using AI/machine learning today?
Hospitals are using AI and machine learning in various ways. For example, hospitals use CIAs to help manage post-acute care and avoid readmissions. Since CIAs are interactive and learn the communication styles of patients, crucial care information _ such as reminders to attend an appointment or take medication _ is delivered to them in a manner they are comfortable with. CIAs can even provide encouragement to patients.
AI and machine learning can be used to extract structured information from unstructured data, such as images and clinical text. Hospitals are using AI systems to facilitate the measurement of specific features in diagnostic studies. One example is automating the computation of the ejection fraction of a potential heart failure patient in an echocardiogram. In such a case, AI and ML applications amplify the work of radiologists by performing screening steps and analyzing data, ultimately putting only the most important analyses in front of the care team.
Hospitals are also using AI applications that can read clinical text to recognize key findings that may play an important role in improving outcomes and increasing reimbursement. For example, an AI-based system can interpret “hidden” information useful in caring for patients and can identify when a properly documented code has not been submitted to Medicare for capitated reimbursement.
What steps should a healthcare organization take to prepare for the future – i.e. “future-proof” their organization – for developments in AI?
I would suggest hospitals focus on three areas to start preparing for AI. First, AI relies on access to trusted data. Ensure that data governance and management processes are in place so that data
is cleaned, aggregated and ready for analysis. Second, have a strategy for integrating AI into business processes and clinical workflow. This includes having a strategy for training personnel and managing process change. Upfront education about AI can reduce organizational resistance down the road. Third, build an AI advisory council. I’d recommend looking for external partners who are subject matter experts and have broad experience with AI across different industries. They can help sort through the hype and identify the technologies and use cases that could be most valuable for your organization.
How AI is driving forward-looking healthcare orgs.