The Synergies in Combining Artificial Intelligence and Market Access


Although AI has simplified various market access processes, this technology does also present its share of barriers, so what can be done to find a happy medium?

Dan Sontupe

Dan Sontupe

Not long ago, big data was the big buzz across pharma. It continues to be, though now it’s animating the latest innovation: artificial intelligence (AI). AI is in the news every day via expanded drug discovery, post-market surveillance, earlier and innovative diagnoses, and, well, every part of healthcare, including market access.

The large language model of ChatGPT introduced a new, easily accessible form of AI. Millions of people of all ages are now using it. Before we turn to its application in market access, below are just two examples of how AI is already embedded in healthcare:

  • In education: Harvard’s Artificial Intelligence in Medicine program (AIM) has a mission1 of accelerating the application of AI algorithms in medical sciences and clinical practice.
  • In the literature: The publisher of the New England Journal of Medicine (NEJM) launched a new journal, NEJM AI,2 to identify and evaluate state-of-the-art applications of AI to clinical medicine.

Onward and presumably upward

NEJM has already weighed in with how AI is influencing and improving market access, in part by improving reimbursement through predictive analytics:

  • Billing and coding: For example, examine a large set of claims rejections, identify the reasons, and create a predictive model based on level of complexity that alerts to coding mistakes, thereby generating cost savings and reducing manual adjudication.
  • Prior authorization (PA): Input data for plan member’s clinical history, benefits, medical and pharmacy claims, and compare with past or similar requests. This can result in reducing steps in the PA process, reducing turnaround times, and improving patient and provider experience.

Smart AI claims processing is a breakthrough that is already happening; by one estimate, 85% of claim denials could be avoided. For example, AI can find markers that move automatic adjudication to manual adjudication, and support both by assessing utilization management and health records, as well as identify claims issues (eg, coding) early to avoid appeals.

Other market access stakeholders are asking, what can AI do for me?

Triple value: A blast from the past in the present and future

The term “Triple Aim” may not be as popular as it was a few years ago, but its essence—value-based healthcare—certainly is. AI is very much in play for the triple domains of better care/patient experience/ population health.

Patient care and patient experience are ripe for AI enhancement—for example, through insights about, and interventions for, social and digital determinants of health, whether in traditional care or telehealth, or surfacing insights about patient risk factors in surveys with fine-toothed machine learning. The ability to mine data in more efficient ways can help efforts to improve population care management and financial outcomes.

Let’s not overlook the barriers

John Knutila

John Knutila

The challenges in the use of AI such as privacy, bias, and inaccuracy are neither simple nor simply resolved. For example, federal medical privacy protections, such as the Health Insurance Portability and Accountability Act (HIPAA), apply in the private sector, but different regulations apply to government channels like Medicare. In addition, different states have different privacy regulations. Ensuring individual authorization for disclosing data is a complex task, without even considering how data from wearables may or may not be protected. Much work is required to clarify these and many more privacy variables for AI for all stakeholders.

Data biases must also be addressed as AI comes online, since race and ethnicity already impact healthcare inequalities, and US data may not align with regional needs. Removing data biases is a complex task. But then, AI is already contributing solutions that support personalization and inclusiveness.

A 2023 study reported in JAMA Oncology3 entered prompts for how to treat Stage I breast cancer, with answers compared with National Comprehensive Cancer Network (NCCN) guidelines. While most responses were correct, 12.5% were deemed nonsensical or incorrect and called hallucinations. So, perhaps industry is not fully ready for unassisted disease management.

Who’s ready for next steps?

Pharma stakeholders sensitive to value are paying attention. Think ever-increasing momentum toward value-based systems, pushback on specialty drugs (2% of the US population use biologics, which account for 40% of biopharmaceutical spending), and the push for accelerated drug approval, for starters. The promise of AI, using big and granular data to produce fast insights customized for specific stakeholders, is hard to ignore.

AI and market access decision-making

New tools are most useful when solving serious unmet needs. AI can address new challenges in global launch strategies, interacting more effectively with local stakeholders, or consolidating market segments with new marketing models for specialty drugs. All the buzzwords—innovation, disruption, etc—take on new meaning with the analytical power of AI. Organizations can start by identifying core processes and business strategies with internal and external data sources and move forward smartly with a pilot program.

AI undeniably will be a significant part of all our futures. But in market access as elsewhere, it’s also happening now.

About the Authors

Dan Sontupe is Managing Director and Associate Partner, while John Knutila is VP and Creative Director of Copy, both at The Value Builders.


1. About AIM. Harvard University Artificial Intelligence in Medicine Program.

2. NEJM AI homepage. NEJM AI.

3. Chen S, Kann BH, Foote MB, et al. Use of Artificial Intelligence Chatbots for Cancer Treatment Information. JAMA Oncol. 2023;9(10):1459–1462. doi:10.1001/jamaoncol.2023.2954

Related Videos
© 2024 MJH Life Sciences

All rights reserved.