Leveraging digital assistance and intelligent document processing to create successful patient journeys

The U.S. healthcare industry is under pressure to reduce costs and improve performance while providing a superior member experience. Pharmacy benefit managers, also known as PBMs, are feeling this growing pressure to lower increasing drug costs through optimizing operations without losing focus on members.

PBM companies are engaged in every part of managing prescription drug benefits on behalf of health plans, large employers, Medicare Advantage plans, and state Medicaid programs. PBMs are entrusted to successfully execute the benefits, optimize spend on prescriptions, and support members. PBMs are rated by their performance, including metrics such as member satisfaction as determined by NPS scores, HEDIS requirements, or STAR ratings.

While the basic exchange of a physician prescribing a drug and a patient receiving it has remained essentially the same, everything about how this process works has and continues to change through advancements such as ePrescribing. Workflow processes have evolved from being just paper-based to electronic, affecting all PBM stakeholders: payers, members, drug wholesalers, pharmacies, and pharmaceutical manufacturers. These enhancements are especially critical as the needs of members and their expectations of engagement with PBMs continue to shift. Healthcare organizations must now prioritize providing members with the same kind of seamless, personalized, and fast interactions that have become standard in industries like ecommerce.

PBMs are examining their drug benefit management system from end-to-end to find ways to improve operations, reduce costs and strengthen the patient experience. By looking at workflows that can be automated using interactive AI, machine learning, and intelligent document processing, PBMs can address both immediate operational issues and the larger goal of delighting their members.

Some examples of PBM workflows that can be augmented and automated by digital assistants and AI are:

• Analyzing network disruption files and ensuring correct member communications when notifying a member about a network change
• Analyzing formulary change files to identify members negatively impacted, which may now require a prior authorization or may need to change to a clinically equivalent drug
• Intake and managing of benefits from payers
• Rebates administration
• Clinical-based communications such as collecting missing information for completing a prior authorization or notifying members of PA status

What is a digital assistant?

A digital assistant is based on a cognitive interactive AI system that offers humanlike interactions developed specifically to the needs of a particular workflow. Digital assistants not only receive communication but can produce meaningful interactions by initiating communication with a member or provider.

Digital agents are integrated within a team at a PBM. Acting as a virtual workforce, digital agents are trained by the business users who rely on the results of these AI-driven interactions. Instead of replacing human workers, digital agents augment an existing workforce by taking on mundane and repetitive work that can be automated and digitized. This allows the human workforce to perform at the top of their license, addressing complex issues to better meet patient needs.

“If you understand your workflow and can articulate where your decision points are, a digital agent can really help you. It can take a more probabilistic approach to not just analyzing the workflow and what steps need to be taken, but how to optimize and improve that along the way,” said Amy Sayers, senior assistant vice president, pharmacy services, EXL Health.

For a digital assistant to be effective in completing a successful patient journey, the solution must execute contextually accurate, human-like interactions; be capable of advanced recall; and be available across multiple channels and in various languages. Other capabilities to drive success include:

• Contextually accurate interactions.
  Simply put, the digital assistant must understand context—especially with verbal communications. The system should have enough flexibility and intelligence to interpret if a member mispronounces a drug name, or to understand the difference between I95, the ICD-10 code for hypotension, and I-95, a major interstate highway. Moreover, the system has to reach the right conclusion with high accuracy.
• Human-like interactions.
  Legacy systems such as a traditional IVR system for call centers are transactional and do not provide interactions. A digital assistant needs to not only sound human but also analyze sentiment. Is the user happy or getting annoyed? Is the reaction strong enough that the call should be escalated to a human agent?
• Advanced recall.
  A digital assistant also must be capable of having a conversation across multiple topics. This functionality is based on the extent that the system is trained, but any AI agent should be capable of having a multi-rooted conversation. For example, a member may call to inquire about a network change. After receiving an answer to their question, the member can then easily change topics to ask about the cost of a new prescription. The digital assistant can keep up and properly engage.
• Omnichannel and multi-lingual.
  A digital assistant needs to engage with the member, physician or any other related user in the channel of their preference such as phone, text or email. The solution also must be available in multiple languages to successfully reach the largest number of people.

The right AI-based intake solution can read any unstructured, physical medium and digitizes it into a machine-readable form.

Intelligent document processing

While virtual agents focus on workflows based on phone and voice interactions in call centers, AI-based digital solutions can reduce or eliminate manual processes in the front, middle or back office.

Intelligent document processing is a solution that weaves together multiple technologies to digitize unstructured data, such. The same type of machine learning modeling used to understand telephonic and voice communication can be applied to written and textual conversations. Additionally, the same key features of a digital agent also are needed for intelligent document processing. For instance, information on network or formulary disruptions are often saved as Excel spreadsheets. Notifying members of these disruptions requires manually analyzing these files – a time-consuming process. The right intelligent document processing solution can automate this analysis, reducing the time required for patient outreach.

The right AI-based intake solution can read any unstructured, physical medium and digitizes it into a machine-readable form. Natural language processing models are developed specific to the PBM to identify content with extremely high level of accuracy. For example, the solution can tell the difference between a Member ID, a date of birth, and a claim code. This is not scanning to a PDF. The result is a workable collection of data that is contextually accurate.

Using this newly created dataset, a PBM can run algorithms and develop models to reimagine workflows that better meet member expectations. As an example, digitizing formulary disruption files can be connected to other relevant data, such as a member communication, ensuring the right member is receiving the right communication. Additionally, analytics can be run to determine the level of member impact perhaps driving a different type of member outreach, increasing overall satisfaction with the plan.

EXL’s Chat Box Solution

EXL EXELIA.AI™ is a pre-trained solution for human like interactions through voice and chat, developed across industry specific use cases. EXL Exelia.AI™ make consumers self sufficient and able to instantly solve their queries through next-gen natural language processing and contextual understanding. Our conversational AI teams up with human colleagues to solve challenges for service providers and customers, deliver world class service and measurable results.

Written by:

Amy Sayers 
Senior Assistant Vice President,
Pharmacy Services

Chris Stebbins
Vice President, Digital Transformation