Deana Rhoades (deana.rhoades@nttdata.com) is Principal, Automation, Health Plan Innovation and Consulting at NTT DATA Services in Plano, TX. John Wells (jwellsnow@hotmail.com) is a Medicare Compliance Executive in Baltimore. Sherrie Ryder, JD CHC (sherrie.ryder@nttdata.com) is Compliance Senior Advisor at NTT DATA Services in Phoenix.
File this one under hard to believe but true: By 2020, the average person will have moreconversations with botsthan their spouse.[1] By 2021, more than 50% of enterprises will spend more annually on botsand chatbot creation than traditional mobile app development.[2] There are even more predictions about workforce automation, such as a complete machine takeover by artificial intelligence (AI) and a robotic workforce in the banking industry. By 2025, it is predicted that a virtual workforce will displace 110 million full-time banking workers around the world.[3]
We often hear the word “disruption” used to describe the automation of the human workforce. These predictions―and the evidence of automation in business today―indicate this is no longer a buzzword; disruption is happening right now. Industries such as banking, finance, and retail (think Amazon) are far ahead of the game. Historically, healthcare as an industry has been slow to adopt new technology, even when the technology can help improve workflow and patient care, and the same can be said for robotic process automation (RPA) and AI tools.
RPA and AI are often confused, but they’re not the same. RPA is automation of human activities performed by robots. AI simulates the intelligence of humans; it’s machine learning, working to perceive its environment and analyze the correct steps it needs to achieve goals. In its simplest terms, think of RPA as doing and AI as thinking. Although both types of technology are important, RPA is more widely used and somewhat easier to implement, because RPA is process-driven data (i.e., the same actions are taken repeatedly and with the same outcome). With RPA, companies are realizing greater productivity, more consistency as routine tasks are done the same way every time, and better accuracy, among other benefits.
RPA and healthcare
RPA is an emerging and important concept in healthcare. All the largest health plans in the US have adopted RPA and machine learning, but many others are, at the very least, considering an automation strategy. Those who have not yet considered an automation strategy will need to do so sooner rather than later to remain competitive in the future.
Industry regulations and business strategy are some of the barriers health plans face when thinking about automation. For example, in the Medicare Advantage space, beneficiaries have generally been resistant to interacting with virtual agents. Medicare members typically have many questions about how their coverage will change: Is my doctor in-network, or is my medication covered? Their desire to speak with a live agent, coupled with a generation not always comfortable with technology advancements, has caused more than a few stumbling blocks when plans embrace automation.
Furthermore, robotics and automation cannot address every possible scenario―there will always be “one off” cases that will require a level of human decision-making or human interaction. Also, there are some situations in which regulators will not allow for virtual agents. The Medicare Communications and Marketing Guidelines (MCMG), Section 80.6, states that “All Part D sponsors must operate a toll-free call center with live customer service representatives available to respond to providers or enrollees for information related to coverage determinations, including exceptions, prior authorizations, and appeals.”
There are clearly benefits to RPA and AI that can lead to better member engagement and consumer experience. Human call center representatives would never be able to understand a member’s entire history and make key decisions based on their profile in real-time, but a bot can.
Several Medicare Advantage plans have embraced the sales concept of “needs assessment” during the initial sales process. The purpose of this assessment is to aid the consumer or Medicare beneficiary in selecting the type of plan, whether it is Medicare Advantage, Medicare Supplemental insurance, or original Medicare with a Part D Prescription Drug plan, which best meets their medical and social needs. RPA and AI are tools to assist sales agents. This automation is done through a series of survey questions, which maps responses to the type of plan that would be the best fit for the consumer. In essence, this is predictive modeling based on consumer responses to specific preferences.
The question becomes how a bot can mimic the human response to conversational tone and emotion to predict grievances and other issues. Tools with these capabilities are already being used today and quickly advancing. Emotionally intelligent AI is using face-tracking software to read people’s facial expressions and interpret emotions.
When it comes to compliance, health plans, like financial institutions, have specific concerns. RPA is undoubtedly making companies more productive, but there’s concern that automation may not be compliant enough, or allow for new compliance regulations to be activated quickly enough. The concerns are understandable. The cost of being noncompliant with a state, Centers for Medicare & Medicaid Services (CMS), or the Office for Civil Rights for Health Insurance Portability and Accountability Act (HIPAA) is high and usually at the very top of any health plan’s list of concerns.
Yet, this is where RPA can help, not hinder, a company’s compliance efforts. For instance, consider the specific rules that govern every aspect of a health plan’s actions in a highly regulated industry, from call documentation to claims processing. These actions are performed the same way each time. With automation, there’s a history documenting each step and showing each compliance box is checked.
Because of RPA’s strong capability in data gathering during processing of transactions, there’s a huge advantage to using this data to create a compliance risk assessment. For example, if compliance participates in the business requirements gathering for the automations, they have an opportunity to create a log with a wealth of information related to that specific process that can be mined during risk assessments. RPA may also “see” other opportunities during its daily work.
For compliance auditing, it’s similar: The information in the RPA logs that the operations auditing team requires may not be completely what is needed for compliance auditing purposes, but compliance has an opportunity to define their needs at the same time the robots are being “trained” to do their jobs. Need details on appeals processing? Get the bots to collect it while creating the required member and provider letters. Need additional data for your Encounter Data System (EDS) submissions? Get the bots to log it and send it to your Encounter Data Warehouse (EDW), all while adjudicating the claims in their work queue. Want to improve your Consumer Assessment of Healthcare Providers and Systems (CAHPS) scores? Try using a virtual agent to assist a live call center agent with understanding the sentiment of the caller and suggesting the next best action. With RPA, the opportunities are endless.
Lack of clarity about compliance and regulation can be a hurdle that causes some companies to resist automation, but it shouldn’t stop them. Compliance shouldn’t be thought of as a silo when considering corporate strategy and governance. Operations and IT teams need to include the compliance business unit in an automation strategy to create a smooth path. Health plans operate in a highly regulated industry, and compliance acts as the foundation, along with operations and IT, from which the business operates. Corporate governance over change management for the robotic and virtual workforce is important. Just as is the case for human tasks, when a business process is changed, the change agents need to consider how regulatory requirements may affect the feasibility of the change. It goes vice versa as well: When regulatory changes are made, the governance team needs to be involved with compliance so the robots can be retrained if needed.
The work that robots do needs to be treated as if it is being performed by people. Plans need to audit transactions, manage scheduling, and treat them as if they are members of the workforce, just like human employees. They need to be retrained when something changes, or they may need to be retired (aka robot mortality). But, they are also software. Because of that, they also must be overseen just like all other IT systems. For many reasons, this causes a double layer of oversight.
Even though there is an additional layer of oversight required, there are many benefits to RPA and creating a team of virtual workers. Specific benefits of using RPA for health plans includedecreased turnaround times, reduced labor costs, increased accuracy (with the elimination of “human error”), increased adherence to standardized processes, better capability to handle volume, and increased data privacy. For health plans, the last benefit is significant; private health information no longer must be viewed by a human in automated processes.