There is a mountain of administrative waste in healthcare. Hacking away at that waste could save mountains of cash and improve hospital and health system margins.
There are IT tools that can help provider organizations reduce administrative waste, tools such as artificial intelligence, machine learning, revenue cycle data analytics and others. Knowing how to deploy and use these tools is the key.
Which is why we sat down with Brian Robertson, CEO of VisiQuate, a vendor of advanced revenue cycle analytics, intelligent workflow and AI-powered automation. Here, Robertson discusses what administrative waste looks like, how AI and machine learning can be used to combat revenue cycle anomalies, how strategic automation and revenue cycle data analytics can diminish waste, and how healthcare data, analytics and AI chatbots can help administrative staff get provider organizations on more stable financial ground.
Q. You cite nearly $1 trillion in administrative waste in healthcare. What does this look like? Why is this happening?
A. At the most fundamental level, it is the sheer complexity of our U.S. healthcare system of nearly 1,000 distinct payers and what is still largely a fee-for-service reimbursement system, where on average 40% of claims are not paid electronically on the first pass, and in many cases are still often worked for resolution on a one-by-one basis.
Other countries that operate a single payer system have inherently much more standardization, particularly as it relates to administrative costs. In the U.S. healthcare market, the average hospital or health system has multiple systems of record, including multiple peripheral systems, bolt-on applications and integration of third-party data sets.
These market realities result in significant data fragmentation and operational friction as complex data sets are not often cleansed, normalized and curated to enable highly efficient operational workflow. The overall waste or excess overhead or spending is a combination of administrative, operational, and clinical support systems and functions, and each is wrought with tremendous duplication, workflow redundancy, and other low-value or process waste inefficiencies.
All in, most studies suggest that an average of 25% and growing of every U.S. healthcare dollar, in a $4 trillion dollar industry, exists in some form of industry waste. And according to most studies, the largest and perhaps easiest area to gain some real traction in optimizing value is in the areas of non-clinical waste.
In addition to inherent industry fundamentals, the conventional wisdom is that U.S. healthcare is a decade or longer behind in overall technology adoption, including leveraging modern, high-value technologies such as AI and machine learning.
The current adoption of AI and machine learning in the RCM arena is still nascent and in the early adopter phase of the technology adoption lifecycle. And while there’s been a lot accomplished as it relates to purely digitizing information assets, enabling data for action and leveraging more advanced technologies such as cognitive or intelligent process automation to reduce process is still lagging.
Q. How can artificial intelligence and machine learning be used to battle revenue cycle anomalies that erode hospital margins?
A. Many hospitals and health systems have begun their journey, with various industry surveys suggesting two-thirds have begun to invest in and implement some form of AI. Robotic process automation appears to have gained the most traction up to this point, including task automation in key areas such as eligibility, pre-authorization, and patient account follow-up and collections management.
In addition to robotic process automation, there are additional AI subsets such as machine learning, predictive analytics, natural language processing and cognitive process automation.
Cognitive process automation in a broader context focuses on leveraging large data sets, and machine learning to enable various bots and AI engines to learn how to complete more advanced operational tasks. Cognitive automation mimics behavior of more complex processes that require analysis, judgement and recommendations for actions.
Advanced data curation and crowdsourcing enables deep learning pattern analysis and more automated anomaly detection to decipher different data signals that can be trained to look for process inefficiencies and bottlenecks with delineation between chronic or systemic versus more acute challenges.
AI and machine learning can analyze how process waste is preventing people and systems from working more efficiently and focusing on the most important tasks. Ferreting out this information historically would require an army of data analysts and data scientists to uncover patterns and defects.
But AI, machine learning and computer vision can do this work exponentially faster. This allows you to get to action at scale in a fraction of the time.
The idea that you can train a piece of technology to work and perform 24 hours a day, every day, is extremely powerful. AI and machine learning can be used to constantly monitor and contextualize what’s going on across all clients and systems, from the procedure diagnosis level to the claims entity level to medical records coding, looking to identify bottlenecks that cause waste.
Revenue cycle FTEs are capable of processing much higher impact claims exceptions versus administrative and more clerical errors like constant problems with missing information. Examples include complex medical necessity or coding justification versus fixing missing financial and demographic information.
Q. What can the combination of strategic automation and revenue cycle data analytics do to reduce waste?
A. Although the definition of administrative waste certainly includes unnecessary and broken processes and rework, a broader view of the term includes inefficient tasks and processes that can be made more efficient or even eliminated using modern technology-driven workflow and methods.
One such technology framework that is disrupting traditional methodologies in the revenue cycle today is intelligent process automation, which further broadens the AI umbrella to include a combination of computer vision, cognitive automation, machine learning, optical character recognition, natural language processing and robotic process automation. Think smart and deep data that can predict and even fix problems before they hit a rework desk.
The primary goal behind applying intelligent automation in many cases is to improve the efficacy of problem diagnosis and root cause analysis and to then minimize the often routine and rote processing tasks, allowing staff to focus on more complex exceptions that require human discernment.
This prioritization allows resources to be allocated to tasks that require additional judgment, and which may involve analyzing complex and unstructured data to reach a decision on what actions to take next. Of course, the panacea would be a touchless revenue cycle, but that can only come with more overall market maturity and advanced adoption across all key stakeholder groups of healthcare’s complex revenue cycle.
If staff can be focused on much higher-value exception processing, and core automation is directed to process lower value bulk transactions that can be distilled to well-defined business rules and structured-information flows, then significant improvements in operating incomes can be realized. Robotic process automation is particularly well suited to off-loading those tasks that are high volume and low complexity, with relatively little variation from one transaction to another.
While automation technology can assist across many aspects of this problem, there are specific business processes in the revenue cycle that are better suited to one type of intelligent automation than another.
Data science and AI-driven payment and collections management with continuous algorithmic updates from crowdsourced data wisdom from claims, remittances and third-party data is a growing area where an always-evolving recommendation engine enables humans to process much more inventory faster and with higher degrees of accuracy.
Q. How can healthcare data, analytics and AI chatbots help administrative staff get provider organizations on more stable financial ground?
A. It’s about leveraging automation across all the separate and distinct core processes of the revenue cycle. If you implement a technology and workflow framework that leverages curated operational data, agile lean methodologies, and the umbrella of AI technologies, with a focus on gaining traction with high-value impact areas, then there is tremendous potential to achieve compelling ROI.
As many industry articles on the topic of the potential of AI in the revenue cycle have argued and articulated, AI tends to have boundless potential in high-volume transaction environments where there are endless codified business rules and vast cube farms of semi-skilled labor. And AI technologies are ready to work 24/7 to solve top-line and bottom-line problems that improve multi-stakeholder satisfaction and experience.
But as is often discussed, the most critical ingredient remains and will always be buy-in from the top and actual executive involvement and continuous support in the process of moving up a data- and AI-driven maturity curve. What doesn’t work is half-baked ivory tower support and incremental investment.
Email the writer: firstname.lastname@example.org
Healthcare IT News is a HIMSS Media publication.