Exploring the Role of AI in Managed Care Organizations
Artificial Intelligence (AI) technology and ChatGPT are disrupting countless industries around the world. For healthcare, the possibilities appear promising, with organizations striving to improve patient outcomes and the overall quality of care. According to Precedence Research, AI in healthcare is expected to surpass $187.95 billion dollars by 2030, growing at a CAGR of 37% during the forecast period 2022 to 2030.
From pipeline to prescription, certain healthcare spaces have been quick to adopt AI technology. In the clinical setting, AI and big data has proven helpful in evidence- and probability-based medicine. Statistically based machine learning models are able to interpret clinical evidence and help physicians diagnose and treat disease, such as cancer metastasis. In some instances, AI is even able to support nurse staffing shortages and detect the probability of adverse events among patients.
The utilization of AI in pharmaceutical brand strategy and commercialization is also under active exploration. Use cases range from improved patient engagement strategies to enhanced disease management tools and the implementation of AI-powered digital health technologies.
But not all healthcare stakeholders are embracing AI. Payers are typically slow adopters of technology and infrastructure, as demonstrated by their historical lag in adapting to vertical integration and other large-scale initiatives. Given the expected explosive growth of AI adoption in healthcare, the MMIT Index team fielded an in-depth survey and Rapid Event Primer to explore payer interest in using AI/ChatGPT. Can this new technology enable better decision-making to improve the quality of care?
Payers Undecided About AI and ChatGPT
As ChatGPT and other AI tools have become widely available, managed care organizations have begun discussions on how best to incorporate this technology—but some are unclear on the path forward.
According to our research, more than 90% of payers are generally aware of ChatGPT technology, but are looking for answers on the appropriate utility of these tools. It is still unclear how ChatGPT can understand individualized care nuances, or how the technology can remain objective with the influx of use and data input. As is the case with any new initiative, understanding how these new technologies can streamline administrative processes and reduce costs will be a huge driver of increased adoption across organizations.
Concerns about data privacy and security continue to be at the forefront of payers’ minds, especially as they will likely need to share sensitive information to ensure optimal results from an AI implementation. As concern persists surrounding the access, use, and control of patient data, payers need assurances that sensitive data and information will be closely monitored to protect patients’ privacy.
Key Uses: Patient Communication, Administrative Efficiencies
While there is some hesitation on the best ways to ensure the security of patient data, many payers think AI could be the answer to better patient-centered communication. They also see AI as a tool for potentially decreasing their administrative burden, and for ensuring optimal coverage and access for patients.
Surveyed payers indicated that AI technology could help with streamlining their prior authorization (PA) processes, while physicians and practice managers saw the benefits of using AI tools for diagnosis and treatment decisions. One payer expressed, “We are optimistic about the utility of AI/ChatGPT, as we see promise in [its] application to speed up P&T reviews with AI’s ability to provide expedited summarized comparisons of competitors.”
Payers are increasingly interested in understanding the best use cases in which AI can help to bridge the gap in member education and communication. Implementing strategies using targeted AI to improve patients’ self-managed care resonates strongly with payers, as patient adherence and improvement metrics are of primary concern. The promise of improving patient-focused metrics, such as medication management, will ensure greater utilization of AI across managed care organizations.
The majority (60%) of surveyed payers anticipate that AI and ChatGPT will be highly useful in expediting data retrieval and PA and expect that initial implementation will take place in these areas. More than 65% of surveyed payers, and 41% of practice managers and physicians, anticipate that AI implementation is likely to reduce or eliminate the administrative burden in their organizations. Most physicians are excited by the possibilities for diagnosis and treatment capabilities, and hope that further technological advancements will help to reduce physician burnout over time.
One payer of note highlighted that “the possibilities of these applications will only increase alongside their degree of sophistication.” While the truth of this statement remains to be seen, it is likely that managed care organizations will gain greater benefits as the information provided by these applications becomes more intricate and nuanced. Payers from large national plans and PBMs are excited about the prospect of “identifying members, based on pharmacy and medical claims, who may have specified diagnoses or gaps in care that, if addressed, can improve overall care and outcomes.”
An Optimal Future in Managed Care
It is important for payers to embrace the opportunities created by AI and ChatGPT when considering optimal cost of care, social determinants of health, clinical models for quality treatment, and formulary decision-making. However, it is also important for payers to ensure that their approach to coverage analysis and quality of care is unbiased, evidence-based, and driven by objectivity. As these applications gain prominence, MMIT will continue to evaluate the impact of AI/ChatGPT on managed markets, including the extent to which they ultimately impact patients’ lives.
For payer, PBM, provider and IDN perspectives on therapeutic areas and emerging trends, see MMIT’s Oncology Index and Biologics and Injectables Index.