Guest Column | November 18, 2019

Why Healthcare Should Double-Down On AI-Powered BI For Reporting

By Scott Hampel, MedeAnalytics

According to the Centers for Disease Control, there were 2.5 million reported TBIs in the United States in 2010, resulting in direct and indirect costs of approximately $76.5 billion.

Understanding data is critical to every business, but especially to those of us in healthcare.

Understanding data often is the difference between success and failure.

  • Between positive and negative health outcomes.
  • Between improving and worsening patient satisfaction.
  • Between increasing or decreasing revenue.

Each of these areas and many others in healthcare rely not only on the collection of data but, importantly, the ability to decipher and act upon it. In that intersection, reporting was born. After the advent of reporting, healthcare organizations looked to business intelligence (BI) to understand and act more quickly on the information created by data. Today, we look to artificial intelligence (AI) to move data to the next level: predictive insights powering real-time decision making.

Back To The Future

Before diving into the opportunity of AI reporting in healthcare, exploring where reporting has been helps lay the groundwork.

The differences between traditional reporting, business intelligence and AI, as seen through the lens of healthcare information collection and action, are:

  • Traditional reporting is static, labor-intensive, quickly out-of-date
  • Business intelligence is cloud-based, tells a story with data, near real-time
  • AI is automated, uses algorithms to predict future challenges, positive trends or potential issues and provides real-time, actionable data

Healthcare organizations and providers are catching on. They understand traditional reporting often is inadequate when comes to really understanding data. In revenue cycle management, for example, “Coding and billing is a challenge in a market driven, fee for service healthcare system,” writes Stefano Bini, M.D., Chief Technology Officer, Department of Orthopaedic Surgery at the University of California San Francisco. Billing is currently done manually, sometimes remotely and always asynchronously. Any missing information that would lead to a more accurate or billable code cannot be gathered once the patient has left the office. Further, it is the kind of highly complex, but rule driven task that computers excel at.”

Traditional Reporting

By today’s technology standards, traditional reporting seems rather archaic and out-of-date. However, there will likely always be a place for it in healthcare.

The process can be time-consuming:

  • End users request the reports.
  • Others, frequently in an organization’s IT department, interact with the systems containing the desired information.
  • Reports are run and sent back to the requestor.

This linear, assembly-line progression is highly inefficient. The process takes days, but more often weeks, to complete. By the time the person requesting the information receives the report, it’s frequently out-of-date. These retrospective reports are served to users as a static document, which makes acting on the information extremely difficult. It delivers as much value as it would if the report was never created in the first place. For the organizations that continue to use this process—and thousands do—doing so may impact revenue, patient satisfaction and many other internal and external stakeholders.

Generally, traditional reporting is inefficient, takes a considerable amount of time to process, only looks to the past, isn’t timely and makes it difficult to collaborate among different teams.

Research published by the Journal of the American Medical Informatics Association compared manual chart reviews with “computerized surveillance systems” to understand which was more accurate. In this study, the group looked at hospital-associated infections and found the computerized system detected more infections (92 percent) than manual chart review (34 percent). “Healthcare organizations generally rely on manual chart review to retrospectively measure quality and safety. Yet this ‘gold standard’ is too time-intensive and costly to be the sole means of routinely identifying patient events of interest,” according to the publication.

While computer systems used in the study weren’t infallible on all measures, it’s important to note the study was published 17 years ago and employed computers that have been surpassed today by far more sophisticated technology. Nevertheless, the point about computer system accuracy is apparent today as digital technologies have improved with the ability to parse terabytes of data in record time.

Traditional reporting, while appropriate in many situations, is and will remain an inherently slow and manual process. This isn’t necessarily a negative, but it’s important to understand business intelligence and AI offer many additional ways to help healthcare organizations succeed financially and improve quality.

BI: The Steppingstone To AI-Powered Information

Business intelligence—the use of applications, infrastructure and tools to access and analyze information to improve and optimize decisions and performance—is a dramatic and significant step forward in healthcare industry reporting and a natural transition to artificial intelligence (AI) enabled real-time insights.

Business intelligence is a natural progression in this three-part series that looks at traditional reporting, BI and, finally, AI as each relates to methods to access healthcare data and information. Even though BI is a step forward from traditional reporting, the overall trend for data is toward a centralized, always-on resource, think cloud computing, which is well supported by AI.

As a methodology, BI is an improvement over traditional reporting. BI users, often known as “citizen data scientists,” go directly to the information source. The users have the necessary experience and the ability to get into the data set and extract the pertinent information on their own. The model is self-service, which makes it more convenient for users. (Nevertheless, users need a fair amount of knowledge and training to understand how to submit and retrieve reports.)

The benefits of BI as a reporting tool are:

  • Self-service reporting
  • Immediate report availability
  • Interactive, dynamic dashboards
  • Data storytelling
  • Advanced visualizations
  • Collaboration
  • Alerts

Each element is more advanced than traditional, manual reporting, as BI supports near real-time queries. Modern BI reporting has been available for approximately 15 years and organizations of all types have squeezed out everything it offers. Modern BI is the notion of data storytelling through interactive dashboards, which is in demand from the market.

Modern BI gets us to the almost-automated, flexible and intuitive nature AI. Which, in the end, is the goal.

Why AI Augments Healthcare BI Reporting

Is artificial intelligence in healthcare simply hype? You can’t go an hour without seeing an article extolling the virtues of AI. It selects your movies, it pursues you as you shop online and helps drive your car. It’s Skynet. It will rule the world. Or save it.

AI is all these things and none of them. In some ways, AI already controls our world. In others, the technology is not quite ready to complete its takeover.

While AI was first described in the early 1950s by Alan Turing and solidified as a concept a few years later by John McCarthy, it’s really come into its own in the past few years. When you see a suggested purchase on Amazon or look at recommended movies on Netflix, you’re interacting with AI.

Healthcare has been slower to adopt AI, but many organizations are deep into their journey, while others are committed to implementing it in the coming years. AI is an important way to help the healthcare industry parse the enormous amount of data created every day by patients, providers and payers.

Recently AI techniques have sent vast waves across healthcare…,” according to researchers publishing in the journal Stroke and Vascular Neurology. “The increasing availability of healthcare data and rapid development of Big Data analytic methods has made possible the recent successful applications of AI in healthcare. Guided by relevant clinical questions, powerful AI techniques can unlock clinically relevant information hidden in the massive amount of data, which in turn can assist clinical decision making.

“AI can use sophisticated algorithms to ‘learn’ features from a large volume of healthcare data, and then use the obtained insights to assist clinical practice. It also can be equipped with learning and self-correcting abilities to improve its accuracy based on feedback.”

The clear message—from healthcare organizations and researchers alike—is: AI is integral to the business of healthcare, whether you are a payer, an organization or a provider. And AI only will continue to grow in importance as the healthcare industry adapts to market changes and recognizes AI’s value in improving operations and patient satisfaction.

Understanding the potential capabilities of AI helps healthcare organizations understand the dramatic differences between reporting, business intelligence and AI.

Healthcare Industry Invests

In 2019, Gartner found 27 percent of healthcare payers would be increasing funding for AI in 2019. No one, not surprisingly, said they would decrease their AI spend in 2019.

In addition, 52 percent of healthcare payers say they’re going to fund a business intelligence or analytics solutions, according to the 2019 Gartner Healthcare CIO Survey. Only 3 percent will decrease funding in 2019, Gartner reported.

Healthcare providers, like hospitals and academic medical centers, recognize the value of using AI in reporting operations.

“Artificial Intelligence tools can be used to mine the medical record for both structured and unstructured data, suggest the appropriate billable codes and request missing information in case of insufficient data-entry in real time,” explains Stefano Bini, M.D., Chief Technology Officer, Department of Orthopaedic Surgery at the University of California San Francisco. “If done correctly and accurately, AI supported coding is also an exceedingly valuable mechanism by which to collect the type of detailed, high quality data that analytic software thrives on.”

AI has many strengths when applied to the healthcare industry:

  • Automate routine, repeatable data analysis;
  • Create data insights delivered directly to users;
  • Build analytics, such as chatbots;
  • Posit “what if” scenarios; and
  • Identify data clusters, forecasting and anomalies using algorithms.

When data is loaded, algorithms do the searching and the culled information is served up in reports and dashboards. Specific insights are baked into the resulting report. Automation discovers positive and negative trends and offers prescriptive actions to correct any downward movements of the business. This is akin to a business rules engine that is smart with AI and proposes the best actions for a healthcare organization.

How Wayne Gretzky Predicted AI

The inability to understand and act on data is often the difference between success and failure in healthcare. Using AI in concert with traditional reporting and business intelligence ensures healthcare organizations get the opportunity to receive highly-relevant information in real-time, which makes it extremely easy to act on.

As famously attributed to hockey superstar Wayne Gretzky, “Skate to where the puck is going, not where it has been.” Using AI to power healthcare decision making not only gets you to where the puck is going but helps you predict its movement in the future.

About The Author

Scott Hampel is President of MedeAnalytics.