Finsights Blog

IDNs, personalized medicine and artificial intelligence

Healthcare conversations at the intersection of personalized medicine and artificial intelligence (AI)

Health care providers in the US have not historically been known for their information technology leadership, but recently some have moved aggressively to the leading edge by applying AI to their businesses. Personalized medicine is all the rage in US healthcare today as providers adopt modern information technology and as treatment and testing technologies evolve. Actionable personalized medicine relationships are perfect targets for AI algorithms, and integrated delivery networks (IDN) have ideal databases of longitudinal patient data to feed AI algorithms. The intersection of personalized medicine and AI for IDNs creates potential for very interesting conversations between providers and pharmaceutical manufacturers.

Personalized medicine – Medical decisions on testing, treatment and support being made based on granular patient segment attributes that predict the disease risk and/or response.

Artificial intelligence (AI) – Building self-learning algorithms that discover insights and make predictions based on complex and/or unstructured data.

IDNs – Integrated healthcare delivery networks that have access to longitudinal and comprehensive patient profiles and histories across their continuums of care, including visibility to clinical, cost and quality of care data.

The key for personalized medicine is the data that supports discovery, development and validation of a relationship between a patient segment attribute and a risk or response. Without AI the identifying personalized medicine relationships is limited by the experiences and innovative ideas of providers. With AI the patterns and insights hidden within a database can be discovered and applied to personalized medicine and patient care. We are already seeing conversations around these newly-discovered relationships occurring between providers, and also attempts to engage manufacturers in these conversations.

How can manufacturers contribute to these conversations?

The obvious answer is that manufacturers can apply AI to their own databases. Unfortunately most manufacturers don’t have data and analytics that are sophisticated enough for this, but it doesn’t need to be that way. We are seeing increased manufacturer interest in commercial data strategy to support emerging use cases like these personalized medicine conversations.

Commercial Data Strategy for a pharmaceutical manufacturer typically consists of at least the following components:

Understand business needs – Commercial data strategy must be grounded in the current and anticipated future business needs for commercial functions for the therapeutic areas of focus, product portfolios and individual brands.

View of customers – Many pharma manufacturers are good at managing provider-level master data, but many struggle with maintaining accurate relationships between providers, facilities and parent organizations. Enabling customer roll-ups for targeting, CRM, reporting and analysis requires new skills and data, and additional resources.

Longitudinal patient integrity – Specialty pharma manufacturers often engineer a longitudinal patient view across their specialty pharmacy and service hub data. The use cases we are talking about in this article require a more elaborate approach, i.e. an enterprise approach to PHI de-identification. An enterprise longitudinal patient view can exist across all data sets that have de-identified patient data, including specialty pharmacy, service hub, copay service provider, pharmacy claims, medical claims, testing data and EMR data.

Data sourcing – Smart data sourcing is a critical component of commercial data strategy. It means acquiring data, not based on a vendor sales pitch, but based on an objective and detailed assessment of each available data asset’s size, completeness, attributes and bias.

Data management – Data management is the unsung hero of commercial data strategy. It includes the important components of aggregation, quality control, normalization, integration and storage. A smart data platform enables a manufacturer to have analytical flexibility and adaptability as downstream tools can be changed over time, because of the robust durability of the data management platform.

People and tools for analysis – The sexy parts of the commercial data strategy are the business analysts, data scientists and analytical tools that create the reports, the insights, the predictions and the intelligence that drives the business.

As pharma manufacturers rapidly evolve their portfolios to specialty products, they need commercial data strategy to be product and indication-specific. This also means that commercial data strategy becomes a critical launch work stream that should be initiated at t-18 months to allow sufficient time to understand commercial business needs, to develop an optimal data strategy, and to build the solution to support commercial success from day 1.



Sign up to receive our Finsights Blog recaps by email.