ISPOR 2023: Highlights, trends, and key takeaways

The leading global conference for health economics and outcomes research was back again this year, as ISPOR 2023 welcomed healthcare researchers, academics, payers, and providers to Boston, MA on May 7-10.

The event covered key topics across the healthcare continuum, including:

• • Clinical outcomes
  • Economic evaluation
  • Epidemiology and public health
  • Health policy and regulatory
  • Health service delivery and process of care
  • Health technology assessment
  • Medical technologies
  • Methodological and statistical research
  • Organizational practices
  • Patient-centered research
  • Real-world data and information systems
  • Real-world evidence
  • Study approaches

Hundreds of research findings and best practices were shared in plenary sessions, poster presentations, panel discussions, and workshops.

Groundbreaking research took center stage

Three plenary sessions covered healthcare access and pricing, the growing spotlight on AI, and the value of EHR data when estimating treatment effects.

Global focus on affordability and investment

This session brought together experts from the United Kingdom, Switzerland, and the U.S. to share their perspective on government interests around the world in better managing expenditure of medications, while improving access to care. Covering the U.S.’s Inflation Reduction Act (IRA), Germany’s recent pricing changes, and Spain and Japan’s inwards investment strategies, the discussion offered unique perspectives on issues that span borders.

Particular to the U.S., the IRA will bring new implications to negotiations for maximum prices for brand-name drugs covered under Medicare Part B and Part D, affecting how manufacturers think through their product life cycle. In fact, one manufacturer has already paused their development plans to “continue to evaluate the impact of the IRA”, although negotiations for selected drugs aren’t slated to begin until later this year. Looking ahead, manufacturers (and their research partners) will likely continue to closely monitor the impact of this legislation.

AI wants to chat

The big question during this session was, “will AI augment, disrupt, or distract health economics and outcomes research – or a bit of each?” There is rich data captured in notes or descriptive text in medical records and written by doctors/care providers; using AI could help that data be extracted in a way that is usable and extremely valuable for many research needs, if applied thoughtfully.

Research leaders shared their thoughts on the rise of artificial intelligence in healthcare and how this technology can support innovative discoveries, while noting important precautions to consider when bringing AI to the research table.

Using electronic health record data when estimating treatment effects

This session invited research experts to map out practical solutions to four common challenges:

• • Reducing covariates space for optimal confounding adjustment
  • Dealing with differential information content and its consequences
  • The influence of selective observability in EHR data
  • Efficiently addressing missing data in EHR datasets

Moderator Blythe Adamson, PhD, MPH, Flatiron Health, kicked the conversation off by asking, “What are some of the biggest barriers in the past to using this type of observational data for retrospective analysis and research to generate evidence about medical products?”

Mary Beth Ritchey, PhD, MSPH, FISPE, Chief Epidemiologist, FDA, CDRH, shared the importance of data sources and timing of access, stating, “We know that relevance and reliability of the data source as well as ensuring that we have purpose for that specific question is really important… in particular, the timing of when researchers get that data source and when we can generate evidence from it, it’s really critical. In addition, the detail that we’re then able to get to within the data is really important.”

Regarding applying language models to solve common HEOR challenges, Guillermo Cecchi, MSc, Principal Research Staff Member, IBM Research, offered a thoughtful perspective to the room, challenging researchers to “change the way that we think about how one extracts information from text.” He explained how data collection is often the sum of many parts, taking into consideration “controlling factors you want to adjust for, residents themselves that you want to analyze, or outcomes and endpoints that you want to estimate. [These are] generally found in places that are not in the convenience sources such as claims data that you have available, but rather hidden in that clinical text.” Looking to the future, Cecchi was optimistic AI could contribute largely to the answer the research community has been looking for, empowering teams to turn months of data collection into just a few days or hours.

An additional challenge highlighted by Jacqueline Shreibati, MD, MS, FACC, Senior Clinical Lead, Devices & Services, Google Health, was how hard and expensive it is to take these types of models and bring them up to scale from smaller startups to large companies. She also touched on the challenge of trust in AI on its ability to provide accurate diagnoses, stating:

“If an AI algorithm gives a patient a diagnosis of diabetic retinopathy, do they actually follow up that information to see the ophthalmologist? If they don’t trust it, what are the repercussions for additional cost to repeat that exam, and how do you effectively model that? I think there will be further investigation to involve interviews with patients, and just [gain] better understanding of how patients and consumers trust and respond to output, and how to effectively model that into our cost utility analysis.”

Inovalon poster presentations

Our Data Solutions team was proud to present six posters at ISPOR 2023:

• • Characteristics, Healthcare Resource Utilization, Cost, and Quality Outcomes of Medicare Beneficiaries While Commercially Insured Before and After Transition to Medicare Advantage or Traditional Medicare Fee-for-Service
  • An Assessment of the Direct and Indirect Costs of Bladder Cancer Preceding and Following a Cystectomy: A Real-World Evidence Study.
  • Long COVID-19 Associated with 5-Fold Increase in Total Medical Costs and Greater Use of Inpatient and Outpatient Services Compared to Non-Long COVID-19: An Analysis of Real-World Data
  • Assessment of Medically-Attended COVID-19 Patient Risk Profiles and Health Care Resource Utilization (HCRU) and Costs in the U.S.
  • Treatment Patterns and Healthcare Resource Utilization in Patients with Relapsed/Refractory Multiple Myeloma (RRMM), Classified by Urbanicity and Ethnicity 
  • Impact of Social Determinants of Health on Access to Device Aided Therapy Services for Medicare Fee-for-Service Beneficiaries with Advanced Parkinson’s Disease

Most noteworthy was our study on the impact of SDOH on DAT. Conducted in partnership with AbbVie, this research was recognized as a top 5% out of over 2,000 accepted abstracts. The study found the proportion of APD patients who receive DAT varies by race/ethnicity, income, disability status, education level, and vehicle ownership. These factors may contribute to access issues resulting in health disparities for APD patients who may benefit from DAT.

Key takeaways from ISPOR 2023

From policy and technology to researchers’ perspectives and patient trust, conversations at ISPOR 2023 tackled the biggest topics in healthcare research.

Across HEOR, data and information systems, epidemiology and public health, and beyond, it is real-world insights – made possible by dedicated researchers, new methods of data extraction and analyses, and global partnerships – that will transform healthcare.

Our Data Solutions team is partnering with organizations across the industry to push the boundaries of what research can do. Visit our Research & Publications page to discover their latest findings.

---

Inovalon and design^® and Inovalon^® are trademarks of Inovalon, Inc.