Solving Health Care’s End-to-End Challenges With Generative AI

Health care can no longer afford to ignore tech disruption and enablement. From disparate data sources across multiple stakeholders to tedious, error-prone operational tasks, the industry’s challenges are uniquely suited to transformation driven by generative AI. Every day, the industry increasingly embraces these disrupters, with breakthrough transformation examples like AWS HealthScribe for clinical transcription and Epic integrating GPT-4 into its electronic health record.

In dollars, a McKinsey analysis shows that implementing a well-known set of interventions—in care delivery transformation, administrative simplification, clinical productivity, and technology enablement—could generate a collective opportunity of more than $1 trillion and potentially up to $1.5 trillion through 2027. With careful considerations of bias, accuracy, and transparency, health care is now looking at taking another step toward generative AI as an enabler of process excellence.

Many aspects of health care make it a perfect match for the benefits of generative AI—perhaps the most centrally complex service fragmentation inherent to the member/patient journey. Generative AI can serve as the connective tissue that knits together existing capabilities and sources of process-driven data, which already exist like:

• AI/cognitive services like Optical Character Recognition (OCR), Natural Language Processing (NLP), and Automated Speech Recognition (ASR)
• Allied AI assets like speech analytics and intelligent document processing
• Application platforms from Robotic Process Automation (RPA) to business intelligence and assistive technology

Here we highlight how generative AI and its robust end-to-end competencies can connect touchpoints and amplify health care outcomes, including front-office engagement, back-office transformation, and clinical excellence.

Front-Office Engagement

The modernized front office is the fulcrum point for the optimized member and provider engagement in demand for payers and the consumerized patient experience that providers must deliver. Generative AI can deliver the required experience to engage and retain, with value-based benchmarks integral to service differentiators. Generative AI can help health plans perform more efficiently and save costs while providing enhanced service to members and patients throughout the lifecycle. Tasks ripe for disruption and improvement as per the catalyst generative AI span payer and provider services and include examples like call summarization for after-call work (ACW) elimination, detecting sentiments for proactive interventions, call mining for quality, and interaction disposition analysis.

Additionally, call listening features that facilitate next best actions and the ability to deploy personalized scenario adaptive omnichannel levers like chatbots to answer service questions for members and providers to improve the quality and efficiency of the overall interactions.

Back-Office Transformation

The often siloed back-office work and administrative functions are laden with manual processes to slow turnaround time and fragmented legacy systems to prevent ease of data sharing and synergy. Generative AI can optimize the back-office across the ecosystem to support multiple touchpoints across the claims lifecycle to improve the quality and turnaround time for resolution. Examples include mining of denied and manual claims for process corrections, complex claim abstraction, and leakages prevention by identifying wastage and abuse patterns. Appeals and grievances management, which includes triage and identification of true resolutions, is also an area of applicability. Letter generation use cases tailored to specific explanation of benefits, remittances, and appeals have high automation potential.

Clinical Excellence

The operational streamlining through AI with unstructured data conversion and real-time benefits verification can address high costs of lengthy turnaround and unlock bottlenecks at critical areas like prior authorization. The current prior authorization process takes 10 days on average, according to the Centers for Medicare & Medicaid Services (CMS). Clinical interpretation of content is one of the most handle time-heavy processes because it requires the highest degree of quality and accuracy of outcomes. Generative AI, from a clinical excellence point of view, augments the clinician through medical record interpretation and review. Classic use cases can be clinical summarization and abstraction, clinical reviews involving assisted prior authorizations, clinical coding of medical records, the suggestion of alternative treatments and down coding, and finally, information retrieval of clinical corpora and guidelines.

With these 3 focus areas, there are aspects to keep in mind in terms of health care applications. Generative AI is prone to bias, and this tech might lack domain specialization. As innovation hub experts, Business Process Management (BPM) partners have the scale and experience to unify disparate data sources. These partners often have the retail and e-tail customer experience (CX) skills to implement strategies like generative AI. With an understanding of how to circumvent the risks of these disruptors, they can integrate tech offerings with generative AI for significantly more impactful ROI.

-

Abhishek Danturti Sharma is Principal of Business Transformation at Sagility where he leads the company’s cognitive and AI initiatives in a consulting and solution delivery capacity. Abhishek is responsible for executing the digital product and innovation agenda across the health care organization. He is an alumnus of the University of Maryland, College Park, and BITS-Pilani, India.