OR WAIT null SECS
Julian Upton is Pharmaceutical Executive's Online and European Editor. He can be reached at email@example.com
Pharm Exec speaks to Katya Svoboda, Taneesha Chawla, and Tanvi Ahuja about the growing practice of applying AI and machine learning in value-based and outcomes-based contracts.
Pharm Exec speaks to Katya Svoboda, Taneesha Chawla, and Tanvi Ahuja — authors of the ICON report, The Impact of Artificial Intelligence on Outcomes Based Contracting — about the growing practice of applying AI and machine learning in value-based and outcomes-based contracts (VBCs/OBCs), the current obstacles, and how the area is likely to evolve in the next few years.
Katya Svoboda, Taneesha Chawla, and Tanvi Ahuja:
AI can be very helpful in informing the design of OBCs and VBCs by identifying the appropriate population, identifying outcomes and metrics, and predicting costs. AI can first be used to identify the disease that could benefit from this type of arrangement as well as patient populations and sub-populations to include in the contract. AI can help remove some uncertainty about how effective a therapy is in specific populations, which patients are most likely to respond, and how and when to intervene. It can also be used to understand where a product can be placed in a treatment algorithm for a disease to achieve better outcomes. Identifying appropriate metrics to assess outcomes tied to use of the therapy is another key use of AI. The deep insights gained from using AI can help payers and manufacturers align on the contract design as well as expected outcomes and costs.
Pharma companies and payers are using various technology platforms based on their requirements. These technologies or software can be developed in-house, but that takes significant resources, so external vendors or third party companies are also available. One company, Inovalon, uses natural language processing to assess risk scores of patients much faster than any traditional methods. They also predict in real time the potential adverse event for a particular patient. Payers, manufacturers, and the government have used Inovalon for AI. Payers are currently using AI to identify high risk patients and even automate intervention. Pharmaceutical companies are using AI to support clinical trials by predicting biomarkers and clinical measures of interest, for personalized medicine, adherence support, and even drug discovery to predict molecular targets efficiently. Some examples we have seen include:
• BMS partnered with Concerto Health AI, to analyze real-world oncology data to generate insights and real-world evidence for a range of data sources, cancer types, and activities including clinical trials, protocol design, and precision oncology treatments
• Gilead has a strategic collaboration with Insitro to create disease models for NASH and find targets that affect the disease's progression
• AiCure has a technology that improved adherence in patients in an AbbVie phase 2 schizophrenia trial by visually confirming medication ingestion on smartphones
Using AI in designing VBCs and OBCs is still in its early stages. Although the era of AI and machine learning has arrived and it is being used extensively in various fields, use in OBCs has been limited. Using AI in OBC or VBC is very new. There is lack of familiarity and no definite examples, so many stakeholders are reluctant to enter into these contracts. There is also still a lack of infrastructure and resources to develop VBCs and OBCs as these contracts are increasing, but not widespread yet in the US. Another major challenge is the lack of trust between payers and manufacturers, a major reason contracts don’t move forward. This can lead to disagreement between both parties on contract design as well as execution. While third parties are available to assist with OBCs, selection of these parties and trusting these parties to be unbiased are challenges for payers and manufacturers. With AI still very new, it is likely to take some time before it is able to demonstrate it improves objectivity and reduces uncertainties. In addition, these contracts are very data dependent, and data security and prevention of data breaches become a concern.
Currently there is some mistrust between payers and providers around measuring outcomes in an OBC. Using a mutually agreed-upon AI algorithm would help improve data accuracy and remove biases, potentially making this process more efficient and accepted. Either a trusted third party can conduct these analyses to alleviate concerns, or both parties can conduct analyses and identify discrepancies, but it is expected that AI will reduce the discrepancies seen and improve trust. If a contract is formulated pre-launch, before real world utilization data is available, a manufacturer’s clinical trial data needs to be used. It makes sense to reassess the contract using the payer’s own membership data after real world utilization and outcomes data is available, and adjust the contract specifics if needed.
Yes, we have seen a willingness to use AI for VBC and OBC, to mitigate risk, predict outcomes, and streamline the process. Companies are already investing in developing algorithms and finding ways to use AI to simplify tasks and assist in designing and managing OBCs. It has not been used much yet because of the lack of internal resources and uncertainties about how to use AI since it is still new. Once this technology is easier and has been tested, we expect use to pick up.
Implementing and administering AI in OBC is challenging currently given the lack of resources at most payer companies. Costs involved are high especially for smaller health plans with fewer resources. In the future, we may expect larger plans to get AI resources in-house but smaller health plans will struggle to integrate AI capabilities into their companies. Outsourcing or buying AI is a valid option given many payers express willingness to involve an unbiased and independent third party vendor to assist with AI resources and data analysis for outcomes based contracting with manufacturers.
The current challenges surrounding AI in OBCs such as lack of familiarity with AI, lack of big data, and hurdles associated with aligning on outcomes between payers and manufacturers are expected to decrease over the years with the increasing use of AI for OBCs. More payers would be willing to implement OBCs pre-launch and in early stages given AI would be efficient in predicting outcomes for the future. Using AI for OBCs would expand the disease areas in which payers would be willing to implement contracts since AI would be successful in predicting outcomes across diverse disease areas, and also lead to increasing implementation of different types of OBCs that are currently thought to be too risky or have attributes that are difficult to align with between manufacturers and payers.
Katya Svoboda is Senior Principal, Global Pricing & Market Access at ICON plc. Taneesha Chawla and Tanvi Ahuja are Senior Analysts at ICON plc.