ChatGPT and AI’s Impact on Payment Integrity
ChatGPT, a generative Al chatbot, gained popularity and momentum and reached 1 million users in just five days after release. A quick primer on ChatGPT – it’s a conversational Al chatbot built by OpenAl and released in November 2022. Generative Al applications, like ChatGPT (GPT stands for Generative Pretrained Transformer), are trained on vast amounts of data from the internet, including books, articles, and websites, and can generate new content like text, images, and audio that are seemingly realistic and human-like. ChatGPT/OpenAl is funded by Microsoft. Google has released an Al chatbot, Bard, similar to ChatGPT and like ChatGPT, it is powered by a language model to converse with users. With the distribution power of Google and Microsoft, the average users will be immersed in Al over the next couple of years. Despite Generative Al being in its relative infancy, with massive funding from Venture Capitalists, the technology is being incorporated in various applications and is disrupting industries and processes.
Generative Al is already having an impact on the healthcare industry and will play a crucial role in the future evolution of modern healthcare. It’s already making progress in areas like research, drug discovery, medical transcription, medical diagnosis, documentation etc. Within health plans, generative Al is presenting new opportunities to drive value and efficiency in areas like Underwriting, Sales Experience, Customer Support and Claims.
Generative Al has the potential to significantly impact payment integrity by automating the backend processes and detection of fraudulent billing and other types of improper payments.
Following are some areas within Payment Integrity that can be impacted using Generative Al and tools like ChatGPT:
Payment Integrity concepts are designed to ensure that healthcare payments are accurate, appropriate, and based on sound principles of medical necessity, cost containment and are referenceable. Generative Al can be helpful in developing concepts by providing insights and patterns that might be difficult to discern using traditional statistical or analytical methods. The use of generative Al can help identify anomalies and patterns that might not be immediately apparent, allowing payment integrity professionals to develop more effective strategies for reducing fraud, waste, and abuse in healthcare payments.
Audits are a critical component of payment integrity and can be time-consuming and resource-intensive, as auditors must review large amounts of billing data and documentation. Generative Al can automate many aspects of the auditing process by analyzing billing data and identifying patterns that indicate potential fraud or errors.
Payment Integrity teams spend time and resources keeping up with the changes to the underlying referenceable policy that drives a concept. Generative Al can help analyse large amounts of data, including claims data and regulatory filings, to identify patterns that indicate changes in CMS policies.
Leveraging Al to analyze claims data and other relevant data sources, it is possible to identify Members that may have other insurance coverage.
Speed to Deploy Concepts
With Generative Al, business users can write code even if they do not have a background in programming or computer science. Generative Al tools have natural language processing capabilities that allow users to input text-based prompts and receive code output. This means that a business user could provide a description of the functionality they require, and the generative Al tool would produce code that implements that functionality leading to faster deployment of Payment Integrity concepts while minimizing the dependency on IT resources.
Rebills and Adjustments: Al can flag claims that have been submitted multiple times or that have unusual billing patterns. Data Aggregation: Generative Al can aggregate external data from various sources and integrate it for analysis and decision making. Payment Integrity relies on certain external data and Generative Al tools can be trained to aggregate and analyze the data from various sources.
Generative Al also be used to automate the process of analyzing medical records to ensure that they accurately reflect the services that were provided. By comparing medical records to billing records, generative Al can identify instances where providers may have billed for services that were not actually provided or where they may have provided services that were not documented in the medical records.
Generative Al can be used to create synthetic data that can be used to train machine learning models to identify potential fraudulent behavior. This synthetic data can be used to simulate a wide range of billing scenarios, allowing the models to learn to identify common patterns and anomalies. Once the models have been trained, they can be used to analyze real-world data in real- time, flagging suspicious claims for further investigation.
The use of generative Al has the potential to significantly improve, accelerate and disrupt healthcare payment integrity. We at ClaimShark believe in the transformative power of Al on payment Integrity and have been hiring, deploying teams, and working with partners to accelerate the use of Al in our products. We have incorporated Al into our Clinical Audit and Coordination of Benefits platforms. We are partnering with our customers, pooling resources, and identifying areas within Payment Integrity that can benefit from Al.
Where are you in your Al journey? We’d like the opportunity to collaborate and exchange notes to broaden our collective understanding and accelerate progress together. Our team and Al advisors can better understand your goals and challenges and work with you on your strategic Al vision and roadmap. To know more about our products and how Al is being leveraged or to exchange notes, please drop us an email at firstname.lastname@example.org