How Insurance and Healthcare are Getting an AI Revamp

The influence of AI in healthcare is wide-ranging, with noteworthy contributions including better patient care, progress in research, cost reduction, and improved accessibility to healthcare services. Ashutosh Khare Centre Leader (India) & Head - Analytics at Cigna, during the ongoing Cypher2023 event, the biggest AI conference in India, delved into the different use cases where generative AI can prove to be transformative in the overall healthcare sector.

Cigna Group is headquartered in Bloomfield, Connecticut, USA. It was formed in 1982 with the merger of ILA Corporation and Connecticut General Corporation. Even though it is present in 30 different countries and serves 200 million customers worldwide. Khare said its major source of revenue is the US market. In India, Cigna has a Global Capacity Centre (GCC) in Bengaluru and is planning to hire around 1000 people in the next two years.

Healthcare chatbots

“We manage a multitude of contracts, particularly within the realm of healthcare, including millions of agreements with various vendors, agencies, pharmacies, and clinical companies across the United States. Presently, we are in the stages of developing a chatbot system that will comprehensively handle these documents,” Khare said.

This chatbot will also be integrated into an IVR (Interactive Voice Response) system, enabling it to respond to inquiries effectively and efficiently. Moreover, Cigna is working on implementing a medical chatbot that streamlines insurance-related queries. Instead of patients sifting through documentation and manual processes to determine their insurance eligibility, this chatbot swiftly analyzes patient data, insurance details, and medical history. It provides quick, accurate answers about eligibility for specific medical treatments or interventions.

Disease Diagnosis

One of the areas where healthcare companies are heavily investing is disease diagnosis. For instance, Cigna is actively working on developing AI algorithms capable of diagnosing diseases from a variety of medical imaging data. This technology can identify patterns indicative of various diseases, such as skin diseases from images of images.

Drug Discovery 

Generative AI is also a game-changer in drug discovery. It aids pharmaceutical companies in identifying new drug molecules, offering potential solutions for various ailments. The AI algorithms analyze a vast array of data sources, such as medical records, wearable device data, and genetic information, providing more accurate risk assessments.

“GPT-4 and Bard are two of the most widely used Large Language Models (LLM) right now. We use a fine-tuned version of Bard which is called BioBard and is used specifically for clinical research,” Khare said.

Managing Healthcare Costs

Cost management is a challenge in the healthcare industry. Khare said generative AI assists in optimising medical coding and billing practices, ensuring accurate submission of insurance claims, thereby reducing claim denials and payment delays.

It also aids in managing healthcare costs by reducing unnecessary medical tests and optimising provider contracts.

Ethical concerns

As AI becomes increasingly prevalent in healthcare, ethical considerations are paramount. Transparency, privacy, and data security are crucial. The decision-making process of healthcare AI should be transparent and explainable. Patients should have control over their health data and be informed about how it will be used.

Cigna, like other healthcare companies, understands the significance of maintaining a human touch in healthcare. AI should complement human expertise, not replace it. The goal is to bring about equity and accessibility in healthcare while ensuring ethical practices and accountability throughout the process.

Although there are numerous use cases in consideration, Khare mentions that for the current year, they have chosen six specific use cases to focus on. They plan to expand this to 15 for the following year.

“If we are able to implement two -three generative use cases in six months, then that’s good progress,” Khare said.