Separating the hype from reality in healthcare AI

Separating the hype from reality in healthcare AI

Artificial intelligence (AI) and machine learning technology are sweeping most tech sectors and industries, and healthcare is no exception. In fact, at HIMSS18, no technology was hotter than AI.

"Artificial intelligence has been around for a while, but why all the buzz around it now?" said HIMSS 2018 AI panelist Pamela Peele, chief analytics officer at University of Pittsburgh Medical Center Health Plan and UPMC Enterprises. "It's because we have dense, robust algorithms, tons of data and the ability to handle it computationally. It's the perfect storm."1.

But there are major barriers to overcome that stand between the baseless claims and productive reality. The most important is the availability of large quantities of high-quality data that can be used to train algorithms. In many organizations, the data isn’t in a single place or in a useable format, or it contains biases that can lead to bad decisions. Organizations that want to prepare for effective use of AI and machine learning must first assess existing information systems and data flows to distinguish the areas that are ready for automation from those where more investment is needed2.

Although this assessment must take place across industries, discrepancy in data flow can be especially true in healthcare where data has existed in specialty silos inside and outside the organization.

According to Medicomp Systems CEO David Lareau3, “these much hyped ‘innovations’ cannot learn to be effective and accurate from poor quality data which is a major challenge in healthcare, especially in the clinical realm where data is complex and often unstructured. The progress of machine learning largely depends on the precision of the data being processed by the algorithms.”

Even for organizations actually deploying AI and machine learning systems, poor data quality hinders ROI and slows adoption.   

In a recent INFOSYS market survey, nearly half of all respondents (49 percent) reported that their organization is unable to deploy the AI technologies they want because their data is not ready to support them.4 As such, approximately 77 percent of IT decision makers said that their organization is investing in data management, particularly in India (91 percent) and the United States (89 percent).

Kaiser CIO Dick Daniels told “CIO” that AI is “the one technology that could make a huge and dramatic change" because it can digest large amounts of data quickly, and turn that into usable information for clinical decision-making.5

And therein lies the rub, which is that these AI and machine learning systems must be able to continually and quickly ingest large amounts of healthcare data, but without the right infrastructure and process, they can’t. Noted in a previous post, that in order for companies to take advantage of AI and machine learning, a data strategy should be put in place to foster success.

So what does a data strategy for AI and machine learning look like?

Companies interested in maximizing success with AI and machine learning need to invest in data management technologies.

But, the needs of these complex systems are not the same as traditional business intelligence such as dashboards or data warehouses.  The newer, more urgent needs are all about rapid ingestion of data from all sources.

Companies should look to data management technologies that support the IDC Third Platform principles. Third Platform technologies enable businesses to accelerate digital transformation (something healthcare surely needs) and are anchored by four key technology areas: Big Data & Analytics, Cloud, Mobile, and Social. IDC notes that AI and machine learning are part of the Fourth Platform for digital transformation, so it makes sense to leverage Third Platform technologies for Fourth Platform success.

In addition to the right technology mix, companies need to examine information assets and information flows with an eye toward logistics – not logistics in the physical sense of moving items, but information logistics that feed the information factory.  AI and machine learning can simplistically be thought of as hungry algorithms, much like an assembly line that require continuous inputs to be productive.

The goals of information logistics mirror physical logistics: deliver the right product (information) in the right format, at the right place at the right time for the right people (i.e. algorithms in the case of AI and machine learning).

It is essential for companies to spend time understanding the data flows into AI and machine learning systems in real time, what the quality of that data looks like, and how it can be stored, augmented and used for future training.

What does a successful AI program in healthcare look like?

An excellent outline of the key elements of a successful AI implementation comes from Chris DeRienzo, chief quality officer at Asheville, North Carolina-based Mission Health.6 DeRienzo outlines several key ingredients necessary to create the kind of continuous improvement environment needed for successful AI and machine learning. Most of them, perhaps unsurprisingly, have to do with people, not technology:

  • Reliable data is foundational;
  • Strong clinical program leadership with a deep understanding of business and clinical practice;
  • Engineers to support and surround the clinical team with expertise in clinical or manufacturing to translate process;
  • Business analysts who can visualize data and build adaptive processes into clinical workflows; and
  • Clinical support from respected subject matter experts who are committed to improvements.

It is interesting to note that only one of the five items outlined is really about technology.  Getting data clean is a monumental but very important task. It is part of the process of establishing a data strategy, both in how to manage data and how to actually put it into practice.

Once organizations put the right technologies, teams and process analysis pieces in place, AI and machine learning programs will be poised to achieve the kinds of success expected of such hyped technology.

“I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do. – HAL 9000 from 2001: A Space Odyssey

[1] Sutner, Shaun. “HIMSS 2018 Focuses on AI in Healthcare.” SearchHealthIT, 28 Feb. 2018, searchhealthit.techtarget.com/news/252435933/HIMSS-2018-focuses-on-AI-in-healthcare.

[2] Noga, Dan WellersTimo ElliottMarkus. “8 Ways Machine Learning Is Improving Companies' Work Processes.” Harvard Business Review, 7 June 2017, hbr.org/2017/05/8-ways-machine-learning-is-improving-companies-work-processes.

[3] Gutierrez, Daniel. “HIMSS 2018: Perspectives on Health Industry Use of AI and Machine Learning.” InsideBIGDATA, 6 Mar. 2018, insidebigdata.com/2018/03/05/himss-2018-perspectives-health-industry-use-ai-machine-learning/

[4] “Leadership in the Age of AI – Adapting, Investing and Reskilling to Work Alongside AI.”InsideBIGDATA, INFOSYS, Feb. 2018, insidebigdata.com/white-paper/leadership-age-ai-adapting-investing-reskilling-work-alongside-ai/.

[5] Sweeney, Evan. “Plenty of Buzz for AI in Healthcare, but Are Any Systems Actually Using It?”FierceHealthcare, 11 May 2017, www.fiercehealthcare.com/analytics/plenty-buzz-for-ai-healthcare-but-any….

[6] Miliard, Mike. “What Does a Successful AI and Analytics Program Look like?” Healthcare IT News, HIMSS Media, 12 Feb. 2018, www.healthcareitnews.com/news/what-does-successful-ai-and-analytics-prog…?

 

This article is automatically posted by WP-AutoPost Plugin

Source URL:http://www.healthcareitnews.com/news/separating-hype-reality-healthcare-ai