Contextualizing Artificial Intelligence for HEOR in 2023

  • This page as PDF


ChatGPT marked an AI inflection point decades in the making—with implications for health economics and outcomes research.

In this series, Avalere is identifying the top trends in health economics and outcomes research (HEOR) that are shaping the landscape in 2023 and beyond. In this Insight, Avalere experts dive into trend #6: the use of artificial intelligence (AI).

AI Advancements Are Transformative for HEOR

In collaboration with IBM, Harvard, and Bell Laboratories, J. McCarthy et. al. coined the term “artificial intelligence” in their “Proposal for the Dartmouth Summer Research Project on Artificial Intelligence” on August 31, 1955. In the decades since, parallel developments in digital health technologies and advanced analytics have continued to increase the potential of AI to transform healthcare and life sciences. Integrating digital and analytics (DnA) capabilities offers new tools and methods for exploring HEOR use cases. Specifically, supervised or unsupervised machine learning (ML) techniques or newer generative AI models can be applied to analyze either structured or unstructured real-world data to generate real-world evidence (RWE).

Figure 1. DnA Capabilities Combine Digital Health and Evidence Generation to Advance Innovations in HEOR
Figure 1. DnA Capabilities Combine Digital Health and Evidence Generation to Advance Innovations in HEOR

Top HEOR Uses Cases for AI and Predictive Analytics

HEOR researchers are applying AI algorithms to administrative claims and electronic health records data to undertake outcomes analyses such as:

  • Comparing effectiveness of traditional patient interventions versus AI/ML augmented interventions (e.g., pain care management)
  • Modeling decision and budget impact on patient journey and healthcare resource utilization
  • Incorporating clinical trial arms that model response to therapeutic alternatives

ML has been used frequently in these functional and therapeutic areas:

  • Imaging: AI-assisted diagnostic imaging is currently the leading use for AI applications in healthcare, with 222 AI-augmented devices approved in the US and 240 in Europe in 2015–2020. AI achieves better diagnostic performance than human experts in several medical specialties like pneumonia (radiology), dermatology (clinical images), and pathology (detection of metastases in breast cancer) and helps reduce wait times and increase precision.
  • Rare Disease: AI-assisted diagnostics and disease awareness have reduced undiagnosed and misdiagnosed rare diseases. ML-based techniques are used in clinical trial recruitment and identifying genomic variants. Pharmaceutical manufacturers are investing heavily in synthetic drug development.
  • Patient Care and Administration: Clinicians (e.g., oncologists) use AI/ML techniques where time to treat is especially impactful on clinical outcomes. AI-assisted pharmacovigilance using wearable technology helps monitor patient status and drug effectiveness in clinical trials. Additionally, billing staff use AI-based automation for prior authorization and coding.

Responsible Use of Generative AI Applications for HEOR

Generative pre-trained transformer (GPT) models are neural networks that use transformer architecture. Large language models (LLMs) based on GPT are trained to identify patterns in large text-based data sets. HEOR researchers can use application programming interfaces to leverage public, standard-purpose LLMs like ChatGPT, Bard, Claude, or Cohere; medical LLMs like BioGPT-JSl or Hippocratic AI; or bespoke “closed” LLMs on proprietary document libraries to conduct systematic literature reviews, synthesize text, and generate first drafts and content variations for HEOR analysis.

As Avalere previously reported, it remains important to mitigate bias in training data and apply responsible AI frameworks to cultivate AI alignment with human values. The legal landscape is constantly evolving to address AI issues related to privacy, intellectual property, liability, bias, and explainability. On October 30, 2023, the Biden–Harris administration issued an executive order on AI to regulate and safeguard the research and development and use of AI models. The executive order includes provisions addressing “synthetic biology” and the “incorporation of equity principles in AI-enabled technologies used in the health and human services sector.”

How Avalere Can Help

Look for future Avalere Insights and webinars on HEOR top trends, including value assessment, health equity, and policy pressures, and view the related webinar on AI for HEOR. To learn more about how Avalere’s evidence and strategy experts can help you stay on top of this evolving landscape and support your HEOR or RWE initiatives, connect with us.

Webinar | A Closer Look at Patient Support On June 6 at 2 PM ET, Avalere experts will explore how potential implications of the Inflation Reduction Act (IRA)’s out-of-pocket cap, in addition to other key regulatory and policy activities shaping benefit design and patient cost-share (e.g., EHB), could impact patient commercial and foundation assistance. Learn More
From beginning to end, our team synergy
produces measurable results. Let's work together.

Sign up to receive more insights about Digital Health
Please enter your email address to be notified when new Digital Health insights are published.

Back To Top