Merck VP for Global Market Access & Pricing Strategic Planning Marco Rauland outlines how AI has already begun to transform pharmaceutical strategy and highlights five ways in which it can be better utilised in pricing and market access decision making.


Artificial intelligence (AI) has been creating a buzz in the pharma industry for quite some time now. Many pharmaceutical manufacturing companies are looking to model their business processes by gauging future requirements. The use of AI can help to improve the strategies for new and existing products and the next evolution is the use of AI in market access decision-making.


Why AI needs to be part of global pharma pricing & market access debates

Dynamic market conditions are forcing pharma manufacturers to rethink global launch strategies, as well as local market access strategies such as interaction with payers, prescribers and authorities in order to achieve the expected return on investment and market share. Since ‘traditional’ global product launch strategies no longer yield the expected results, industry trends require a more holistic approach and thinking to increase market share by applying analytics using Big Data and AI.


What are the industry trends and market conditions?

  • Increasing health payer pressure to reduce costs by new measures in addition to reference-pricing and clawbacks
  • Endorsing pricing and reimbursement legislation by governments to minimize pharmaceutical spending growth (e.g. Tender Management).
  • Consolidation of pharma’s target market segments requiring new marketing models for speciality drugs
  • Reimbursement approaches shifting from traditional fee-for-service to ‘Value-Based Healthcare System’ (quality/cost) via patient outcome measurements.
  • Physician prescribing decisions influenced by formulary protocols defined by healthcare policymakers and payers (e.g., Managed Care Organizations (MCOs) and Pharmacy Benefit Managers (PBMs) increasingly use formulary exclusion policies including innovative and specialty medicines).


Role of AI in Decision-Making

Given these trends, manufacturers will need to show innovation and differentiation to make a case for new drug prices scrutinized by payers, providers and policymakers. The goal is not just to set prices to recoup research investment and successfully hit profit goals as was traditionally done with International Reference Pricing (IRP) and Launch Sequence approaches.

AI can accelerate the gathering and analysis of medical evidence, market access information and the clinical trial results supporting the case which could take a long time if done by traditional means (i.e., human intensive analysis).

In addition, AI data analysis can improve product launches by identifying patterns in several key areas:

  • Local Market Trends – Identification of standardized protocols and guidelines for treating patients in local markets
  • Key Opinion Leaders (KOL) – Identify the trending KOLs for better focus on effort and resources within various areas of the business (e.g., clinical, medical affairs, commercial, sales and marketing) and identify opportunities for focused sales and marketing action
  • Patients – Integration of multiple patients’ data patterns for examination of unmet needs
  • Product potential – Identify opportunity for the brand after assessing patients’ unmet needs (i.e., Optimal pricing to demonstrate superior patient outcomes and value for payers)


So, how to move forward? How do you select an AI solution? Here are a few key points:

  • Start by identifying the core processes and business strategy, including organizational roles and interdependencies
  • Identify and assess the internal and external data sources that compile patient and customer data for analytics inclusion into the AI solution
  • Move forward with evaluating AI solutions by executing pilots


Market Access-related applications of AI

  1. AI in Pricing & Reimbursement and Health Technology Assessment (HTA)

AI brings a number of advantages when it comes to HTA evaluation, pricing and negotiations by helping companies to build value prepositions, differentiate products and provide rapid answers to questions from the HTA bodies.

AI can involve analyzing large amounts of data, for example, with AI it has been possible to conduct a GAP analysis of three major ulcerative colitis drugs, focusing on real world studies. Based on paper reporting outcomes, the system appropriately grouped together numerous outcomes for each drug in just 20 mins which would be pretty challenging without machine learning.

At the moment, pricing and reimbursement modelling requires many hours building evidence from clinical trial data, real-world evidence, historic drug submissions, pricing data and HTA appraisals.

UK AI startup Okra Technologies has launched a new AI platform called ValueScope that can predict the price as well as the likely outcome of negotiations with HTA agencies like NICE in the UK and IQWiG in Germany with more than 90 percent accuracy.

The AI was built using data from more than 1,700 drugs that have been launched in Europe, creating a virtual model for HTA negotiations. It was put through its paces in Germany, and hit the 90 percent accuracy threshold when predicting the outcome of appraisals and the negotiated price of Phase III treatments.


  1. AI in Payers’ Informed Decisions

On Feb 2021, Humana, the payer and technology company partnered with IBM Watson Health with the aim of improving and personalizing member engagement, particularly engagement related to healthcare spending and benefits information through AI.

The AI-enabled virtual assistant will be connected to the IBM Watson Health cloud to provide accurate benefits, costs, and provider information to agents, employers, and plan members. The payer also plans to leverage AI to reduce members’ healthcare spending by helping them predict their healthcare costs.


  1. AI in Outcome-based Contracting (OBC) & Value-Based Programs

AI can be very helpful in the design of OBCs by identifying the appropriate population, identifying outcomes and metrics, and predicting costs.

  • Identify qualifying populations – AI can be used to identify the patient populations and sub-populations to include, that could benefit from this type of arrangement.
  • Level of risk – AI can help remove some uncertainty about how effective a therapy is in specific populations.
  • Disease or product selection – AI can inform where a product may best be placed in the treatment algorithm by analyzing currently available treatment options.
  • Identifying, measuring and tracking outcomes – AI can be used to identify appropriate metrics to assess outcomes for the therapy.

In Dec 2020, Centene Corporation acquired Apixio to leverage its AI solutions to broaden the evidence base that the payer uses for value-based reimbursement. Apixio will provide Centene with an AI platform that can both find and examine unstructured patient data. Upon organizing this data, the platform can present payers and providers with quality measures information and feedback on services for value-based care reimbursement.


  1. AI to predict roadblocks for Field Reimbursement Managers (FRMs)

IntegriChain is utilizing AI and machine learning to develop risk score models that will allow Field Reimbursement Managers (FRMs) and similar functions to proactively intervene on behalf of patients who are likely to experience roadblocks in their journey — before the roadblocks occur. Patients can contain additional information such as insurance benefit details, channel, diagnosis code, and out-of-pocket cost. Also, it can bridge patient status data to other relevant datasets such as dispensing, copay, and population data, to engineer new features for predictive models.

Key stakeholders such as Patient Services and FRM team leaders can then foresee cases and prioritize interventions for patients with higher discontinuation risk scores.


  1. AI to predict best pricing

Predictive Acquisition Cost (PAC), a drug pricing measure developed by Glass Box Analytics and distributed by Elsevier/Gold standard, shows how predictive analytics can be used to optimize Maximum Allowable Cost (MAC). PAC, is a new drug price type that more closely tracks true acquisition cost by leveraging proven concepts from other industries and applying the power of predictive analytics to drug pricing. PAC helps Payers, PBM’s and pharmacies through MAC Optimization by consistently tracking the actual drug acquisition cost more effectively than AWP (Average Wholesale Price) or any other existing price type.

PAC, which has been selected by Oklahoma State Medicaid to better manage its State Maximum Allowable Cost (SMAC) drug price, employs a multi-tiered predictive analytics model to estimate acquisition cost by considering a variety of factors such as: MAC benchmarks, published price lists, existing price benchmarks, drug dispensing, behavioural metrics, supply and demand and survey-based acquisition costs.



There is no doubt that AI is changing our daily lives, and it is likely to take a more active role in life sciences and healthcare in the future, from the development and prescription of medicines to their access to the market. Although the adoption of AI has started in healthcare, ethical, regulatory and technical challenges have emerged. With the increasing amount of medical data available, companies that manage to exploit the AI potential will gain a competitive advantage by rapidly generating insights that traditional human-intensive analytics are unable to access. Through algorithms able to model patient outcomes, prescription rates, patient subgroups and identify KOLs, AI will also continue to play a more significant role in the pricing, reimbursement, and market access of medicines.