Healthcare providers and payers need better ways to resolve patients’ complaints. AI and advanced automation can be the tools that make this happen — giving providers an opportunity to restore the trust of patients who in the past got caught in complex appeals and grievances (A&G) processes. This is especially crucial because of audit reporting rules that use A&G data to assess how well companies handle A&G claims. Audits that go badly can cause countless problems for healthcare companies.

What’s wrong with A&G processes

Conventional A&G processes face many challenges. On one side, we have the patients whose care (or insurance claim) has gone badly enough to motivate them to file an appeal or grievance. On the other side, we have payers working with disparate, inefficient processes that draw out complaint resolution — and invite unwelcome audit scrutiny.

What are the biggest A&G process challenges? Inefficiencies and inaccuracies cause the most problems, while others include:

Multiple disconnected intake channels
Tedious manual processes for document creation, management, storage, and versioning
Poor (or nonexistent) data validation during case creation
Lack of SLA tracking and inventory management dashboards
Disconnected correspondence and poor tracking
Manual, delayed, and inconsistent messaging to patients and providers
Long time spans for research and resolving claims

Add in the requirements for external and physician reviews and it becomes evident that A&G processes are cumbersome and time-consuming. In fact, Virtusa has documented at least a dozen places where bottlenecks occur in A&G processes from our experience helping healthcare organizations implement the latest technologies.

Fortunately, our artificial intelligence, machine learning, and robotic process automation capabilities can help transform A&G processes to benefit patients, providers, and payers alike.

Five levels of A&G evolution

Companies can improve A&G claims with the right combination of technology and human intervention. To help with this, Virtusa’s developed a five-part framework. It starts with the familiar concept of maturity: starting with automating basic repetitive tasks and evolving into applications of natural language processing and other technologies that mimic human thought processes.

Level 1: Robust processes. Digital process engineering lays a foundation for automation maturity. At this level, we deploy digital process automation to streamline claim triaging and unify the overall A&G workflow.

We also include an integrated correspondence module to keep lines of communication open and consistent. These processes can be customized for different lines of business within a healthcare organization.

Level 2: Robotic automation. When the level 1 processes run smoothly, we automate specific tasks using virtual agents (chatbots).

These chatbots work across multiple channels (email, phone calls, SMS, website forms, etc.). Robotic automation also integrates discrete and disconnected systems, dismantling the bottlenecks resulting from having to reconcile data flowing from multiple directions. We also pull in data from external systems here.

Level 3: Intelligent processes. Now, it’s time to automate workflows, using learning algorithms to optimize everyday processes. We’ve found that implementing intelligent process automation minimizes pre/post-service appeals. Our methodology uses intelligent workforce management for people who triage A&G claims and coordinate cases. The software automatically forwards complaints to the proper medical professional based on a skill inventory. It also conducts a dynamic SLA derivation for case follow-ups.

Level 4: Cognitive analytics. Here, we deploy natural language processing and text analytics to mimic basic human judgment. Data input in this phase is more accurate than conventional optical character recognition (OCR), handwritten character recognition (HCR), paper, and fax processes, reducing the errors that often delay appeals.

We also automate appeal and grievance classifications and conduct automated voice analysis of the tone of people’s calls to get a clearer picture of the kinds of things that upset or irritate patients.

Level 5: Focused AI. Now, we’re in the most ambitious phase of advanced learning automation. We’re not just mimicking human judgment — we’re augmenting it with machine learning and decision management software.

These tools can help predict the likely outcomes of nurse and provider research. Moreover, pattern-matching algorithms can help identify cases of fraud, waste, and abuse. Data-driven predictions and auto-corrections also are possible.

Services

Industries

Insights

Careers

About