Making an Impact in Healthcare

RPA, AI, and the Future

Key Takeaways

AI is becoming a component of traditional RPA solutions and offering more predictive tools across numerous use cases. Traditional rules-based approaches to remote process automation (RPA) are gradually being enhanced through AI to offer enhanced predictive ability and efficiencies.

CAQH estimates that automation has already reduced administrative costs by $122B and could cut another $16B from administrative costs. Coupling AI with RPA may offer even more opportunities to automate tasks and reduce work burdens for staff as well as save money. In contrast to other fields where automation may result in lost jobs in healthcare, we see less concern about job losses, given the current shortage of trained professionals that may grow even more severe in coming years as health professionals leave the market. Some administrative job losses may occur, but there will also be new administrative roles focused on managing digital technologies such as AI.

More studies of the impact of RPA, cost of implementation, and emerging ethical issues around AI and RPA are needed. Comprehensive studies of the market for AI-based RPA in healthcare are needed to assist stakeholders in understanding the complexity of implementation, impact on work burdens and costs, and understanding emerging ethical issues associated with RPA use. Chilmark Research is launching a new study on the vendors and the solutions in the AI-based RPA space in healthcare.

Introduction

AI and machine learning have been making their impact felt in the robotic process automation (RPA) domain for quite some time and healthcare is beginning to experience the benefits of automation of some administrative tasks that consume a great deal of time and money. There is also a fair amount of hype in the market and many consumers of RPA are unclear about the differences between traditional RPA and RPA enhanced by AI/ML. For these reasons, we decided it was time to launch the research for a Market Trends Report on RPA that will help clarify the use cases, maturity, risks and future possible trajectories for RPA/AI in healthcare across the administrative and operations domains.

First, some notes on the definitions we plan to use to guide our research. Traditional RPA has been deployed for tasks that humans often must do manually via computers, fax machines, and telephone calls. These tasks include copying files and data, or extracting data from structured databases. Often this includes patient demographic data, specific fields of medical records, credit care, and payment data. RPA bots can be created that follow clear rules-based tasks that have pre-determined actions that they are programmed to follow.

AI/ML-based approaches are quite different, in that they can provide more sophisticated approaches that recognize patterns in images and text or use natural language processing to interpret data (text, voice, images) from structured, semi-structured, and unstructured data. These bots can learn over time and make decisions and provide predictive insights as well. An example would be deploying a conversational agent to manage a call from a health plan member, where the agent would have the ability to predict some of the questions or data the member may need, potentially providing a more efficient remedy to the member’s needs and questions without long wait times and similar frustrations.  The focus of our research will be on this latter category of AI-based RPA services.

Use Cases of RPA in Healthcare

Automation in healthcare, frequently through use of AI/ML, has the potential to reduce administrative costs and waste substantially. The Center for Affordable Quality Healthcare (CAQH) has estimated that automation has already reduced administrative costs by $122B and can readily save another $16B.[1] In some areas such as utilization management where clinical expertise is needed but in short supply, AI-based RPA can resolve important bottlenecks and delays for patients, providers, and even payers due easing the work burden for nurses involved in reviewing cases. The current workforce shortage for nurses has become a challenge for many institutions.

For our study we will be looking at vendors providing RPA products in the following areas:

• Scheduling/new patient registration/care coordination
• Recommendation engine/chatbot-like apps
• Claims management/adjudication/billing/coding
• RCM/Prior authorization/benefits
• Regulatory/Compliance (HIPAA, etc.)
• Supply chains/dashboards
• Patient engagement
• Discharge planning
• Data entry/migration/extraction

Analysis of Vendors

Across these use cases we want to develop criteria that can differentiate the vendors in the ecosystems and assess the overall strength of offerings. For this we are developing some criteria for rating the products and services we find in the AI-based RPA space. These will include:

• Maturity of AI models (process complexity that they can solve for)
• Number and scale of clients
• Mapping tools for client engagement/strategic implementation
• Time savings/work burden reduction/process improvements/impact on pricing of services/impact on bottom line of service
• Number of processes they solve for (and sub-processes)
• Code-based or low-code solution and extensibility of solution
• Attended/unattended or hybrid solutions
• Durability of solution: cost of maintenance, replacement
• Customer experience improvements
• Dashboards and alert controls/governance processes
• Single or standalone platform
• Data integrity and security
• Contribution to innovation/digital transformation
• Configurability/flexibility/ease of bot configuration by non-pros
• Consulting services (architecture, deployment of bots/bot sprawl, change management, etc.)
• Sales model/pricing model
• Implementation cycle/cost of implementation
• Data intake/preparation

We think these criteria are important for assessing the robustness of solutions and will also give users of RPA a solid set of criteria for assessing vendors that they may want to engage with as partners. If you are familiar with this space and feel we are leaving out any major criteria, please reach out and we can have a discussion. We are also not including RPA solutions that focus more on clinical use cases, such as medical imaging.

If you have knowledge of this space and are a vendor and would like to provide feedback on our research process or would like to be included in the study which involves responding to a survey and providing a demo, please feel free to reach out to jody at chilmarkresearch.com.

---

[1] 2020 CAQH Index. “Closing the Gap: The Industry Continues to Improve, but Opportunities for Automation Remain”  https://www.caqh.org/sites/default/files/explorations/index/2020-caqh-index.pdf