The evolving landscape of healthcare has placed great importance on data-driven strategies that prioritize and automate member-centric approaches. The use of high-quality data is crucial for enabling health plans to use artificial intelligence (AI) to improve member experience and to optimize operations.
When exploring how to use AI to optimize processes and outcomes, health plans need to consider that its reach expands beyond the widely known generative models, encompassing classical numerical AI models that anticipate member behavior and optimize interactions. Furthermore, the adoption of AI requires careful consideration of data privacy and the production of synthetic data, especially in such highly regulated environment.
AI Models and Use Case Considerations
The use of AI is not a new phenomenon in healthcare. While generative AI (GenAI) and large language models (LLMs) garner the lion’s share of attention, classical numerical AI models have been around for some time, and they remain pivotal in health plan use cases for improving the member experience. Regression and clustering techniques, for example, are foundational for tasks such as predicting member retention, correlating outcomes and optimizing interactions to improve CAHPS scores. Straddling the line between analytics and AI, numerical AI models have presaged a broader use of AI among health plans for the purpose of driving better health outcomes through data.
In using AI to improve health outcomes, healthcare organizations typically break down potential use cases according to risk levels, with factors that include how well PHI is protected and whether AI decisioning will have a direct impact on patient care. For payers, the adoption of AI could start with low-risk applications like automating appointment reminders or enhancing patient registration processes. As confidence in data accuracy and AI grows, its use can expand to more critical areas like risk assessments and personalized care recommendations, which have a direct impact on patient outcomes and costs.
AI enablement in healthcare is multifaceted, impacting various stakeholders from patients and providers to marketers and back-office operations. Identifying key use cases, understanding how AI will be applied and establishing key performance metrics upfront is important for successful implementation – particularly when AI will have a direct impact on member satisfaction and retention.
The Role of Data Quality on AI Performance
When expanding the use of AI to include GenAI capabilities – such as conversational AI using LLMs – it is vital for health plans to consider the health of data being used to power AI applications. High-quality data is the backbone of any successful AI implementation. Health plans often have data siloed across various departments and disjointed systems. On top of this, structured data, unstructured data, and member communications are typically managed separately. While AI can help bring these data sets together for deeper insights, flawed, incorrectly matched, or incomplete data can hinder the accuracy of AI models. Profiling data to eliminate inaccuracies such as incorrect phone numbers and maintaining data integrity across use cases are crucial steps. For AI to be effective in tasks like patient engagement or risk assessment, the underlying data must be up-to-date, accurate and trustworthy.
Trust in data is particularly important in how healthcare organizations approach the use of AI given stringent data privacy regulations such as HIPAA. Protecting PHI is paramount, and any use of AI must abide by existing safeguards. One solution is the production of synthetic data – artificially generated data that mirrors real data without compromising individual privacy. For AI training, as an example, it’s possible to alter existing data to create new, artificial data that is similar to the real data, but different in a key respect in that it does not represent an individual member – and is thus not subject to HIPAA requirements. This model of “synthetic” members can then be used to drive training and ask questions to improve member satisfaction and the member experience.
AI and Coordinated Care: An Ongoing Story
Leveraging AI and other tech innovations gives health payers a unique opportunity to gain an edge by focusing on building truly personal 1:1 member relationships. And what’s more personal than providing better guidance throughout each member’s care journey? AI is increasingly being used to improve risk assessment based on sound patient data, enabling timely interventions that reduce complications and costs, for patients and health payers alike.
Health plans have a vested interest in ensuring members receive timely care, as delays can lead to increased complications and higher costs. AI can assist in identifying at-risk populations and guiding them to appropriate care providers, aligning with CMS guidelines of closing care gaps and maximizing incentives for screenings and preventive care. This proactive approach not only improves health outcomes but also supports financial sustainability.
A historic opportunity for technological innovation is unfolding for health payers, one that can radically revolutionize how individual member relationships are managed. Whether through anticipating member needs, optimizing interactions or promoting healthier behavior, AI has the power to transform healthcare by creating a more personalized, efficient and member-centric system. As trust in AI grows, its applications will become even more integral in shaping the future of healthcare. However, failing to create a strategic vision and a strong data foundation now could leave health payers struggling to keep pace with the industry in the near future.
