Merck KGaA’s Marco Rauland and Kaushal Kishore examine how artificial intelligence and machine learning can best be applied in pharmaceutical pricing and market access. Noting the significant uptake of such tools in other aspects of the pharma value chain, Rauland and Kishore speculate how they might come to revolutionise pricing and market access while cautioning that legal and compliance concerns may present a challenge to widespread adoption.
Abstract
Circa 2030: A commercial representative from a pharma company is walking into a contract negotiation for a recently approved drug. The steps he takes are not, however, into the office of the pharma payer but into a ‘smart negotiation room’ of his pharma company. After a quick retina scan, he gains entry into the ‘negotiation room’ and enters some keywords in a large ‘quantum dot display screen’. A recently built artificial intelligence program ‘Contractum AI’ triggers the value dossier and some scenarios of the contract offer to the payer’s office. A similar AI program receives the offer on the other side and the deal is closed in two days. Welcome to the world of ‘Next generation AI’ in market access. This situation seems to be straight out of a science fiction movie but recent advances and phenomenal growth in the area of AI and machine learning do not leave an iota of doubt that this will soon become reality.
This white paper attempts to illustrate the footprints and journey of artificial intelligence in the pharma and life sciences value chain with a specific focus on potential use cases, opportunities in pricing and market access. With opportunities come challenges and the application of AI in market access is not an exception. One of the key challenges will be the legal and compliance aspect of AI tools, especially in light of the EU regulations. Although not a bottleneck, working around these regulations can slow down the process and adoption. Pharma companies’ market access departments must embrace this opportunity not only to focus on patient value and business growth but also to avoid missing the bus. Early birds will catch the worm, as the famous saying goes.
Pharma Industry: The Upside-Down Flying Goose
Technologies evolve from simple observation to innovation
Unknown
Innovation sometimes looks like an upside-down flying goose trying to match the wind to regain momentum. It’s the magic mantra for survival and excellence. Innovation in the pharma & life sciences industry is no different and has long been the cornerstone for success.
The industry has been trying to sustain itself amidst major upheaval. Shrinking pipelines, increasing competition from biosimilars, and patent expiries all present major challenges, not to mention global supply chain issues, geo-political conflicts, and unforeseen scenarios such as COVID. All of this has led to a radical shift in the way the industry functions. Innovation has always been the need of the hour to sail through rough waters.
One such area of innovation in the pharmaceutical industry is AI which has effectively transformed pharma’s business model in the past few years. This article is going to discuss the maturity of AI across the value chain of pharma. Based on a recent research and industry assessment, it will also highlight a key gap in the later part of the value chain i.e. sales and marketing.
AI and machine learning have been talked about a lot in recent years and, following the ‘ChatGPT’ buzz, need no further explanation for readers. This paper will try to focus on the current status of AI/ML applicability in pricing and market access and explore two concrete use cases where the industry can immediately find some quick wins. Companies must take a leap of faith and create efficiencies using AI techniques, especially in the area of pricing and market access where a huge void currently exists.
AI in the Pharma and Life Sciences Value Chain
The pharmaceutical industry has been investing billions in drug R&D to create novel medicines and bring benefits to patients. Starting from R&D, the key processes in the pharma value chain are depicted below with snapshots of applicability and maturity of AI in each step. One of the inferences we can draw from the study is that the maturity and usage of AI in earlier parts of the pharma value chain is quite robust as compared to the later part of the value chain, especially in sales and marketing. It’s no surprise therefore that, as a part of this study, we could not find any relevant and applied use case for AI in pricing and market access.
Made with Visme Infographic Maker
Sooner or later, the industry will catch up with a strong AI presence in sales & marketing and we have already started to see some applicability in fields like omnichannel engagement and promotions. What’s surprising though is that in the area of commercial, especially pricing and market access, AI has not been explored significantly. As pricing has picked up as a strong topic for revenue growth and sustainability in recent years, it’s the right time to explore it as an effective tool in the pricing space.
A Study of the AI Maturity Index in Pharma and the Life Sciences
Based on a recent study conducted through primary and secondary research, the applicability of AI in pricing and market access departments in pharma has been found to be low to non-existent. This is an eye-opening finding given the fact that AI and ML technologies have been significantly used in other industries, especially in the commercial space.
The research focused on some mid-size and big pharma majors and concluded that almost all of them were at the initial stages of AI usage in this function. The below diagram depicts the summary of the findings
Made with Visme Infographic Maker
Possible AI Use Cases in Pricing and Market Access
Pricing processes, data and tools have matured over the past 10 years. The retail and consumer goods industries have already been using AI-based pricing decisions, especially in dynamic pricing for quite some years and this has been very effective. In pharmaceuticals though this opportunity is rather underutilized. In pharma, prices are heavily negotiated, with negotiation outcomes not always driven by rational aspects and different negotiating bodies often having different interests. AI and ML algorithms can spot patterns that highlight these kinds of aspects which we currently don’t have information on. In the traditional approach, we judge the value compared to existing products, and this is the basis for the price decision, but we do not explore the other factors which would help with negotiations. It can complement the old approach by identifying hidden pricing drivers. Currently, there is a plethora of data on products, prices, discount patterns, negotiation parameters etc which can be leveraged through AI tools and algorithms for better and faster pricing outcomes.
Another key area where AI-based algorithms can make the process faster and more efficient is pricing governance. Currently, this process is heavily manual and a typical pricing business case evaluation process takes 2 to 3 weeks within the pharma company. Interestingly most of the parameters of approvals are repetitive and in the approval chain, information is available in a structured digital format. This makes this use case a promising one for AI/ML-based programming. The approval time can be drastically reduced by AI and ML programs with pre-filled logic and pattern analysis. In the evolving generative AI space, it is going to become much easier to take effective decisions through these AI tools. The way these programs are structured, AI can help us not only to generate smart pricing negotiation models but will also enable quick machine-based pricing decisions. Let us have a quick look at the conceptual structure of such an AI-based engine which will be run by the ‘pricing prompt library’ concept.
Made with Visme Infographic Maker
AI Use Case for HTA Submissions
The pharma industry of today focuses heavily on value-based pricing. Depending on the structure of the healthcare system, data is needed with evidence to assess comparative effectiveness. This evidence is submitted through HTA (Health Technology Assessment) documents or value dossiers. Payers investigate this data, and then based on the additional value in a product the reimbursed price is provided. This means that if a product is less efficacious than others, it will be priced at a lower level than existing ones. This is a very data-driven negotiation. As data collection and processing have enriched over the years, some pharma companies have started using an MIS (Master Information System) in the market access function to draft Global Value Dossiers, which are the basis for local value assessments as well as pricing and reimbursement negotiations. With the advent of generative AI such as ChatGPT, we can foresee a situation in the near future where this system, with the help of AI and ML, can identify trends and concrete evidence on topics like which patient populations benefit the most, hence likely shifting to a population-based pricing model combined with the value benefit story leading to a better price for the targeted population.
Sooner or later, this process will be transformed through AI-based language processing which can help compile publicly available data, real-world evidence, the unmet need, current therapies, and so on. For new launches, as healthcare systems and payers get better equipped to collect data and analyze data for reassessments, collecting RWE (Real World Evidence) data will become much easier and the ability to map it to pricing decisions will become seamless. AI algorithms can check if the drug works in a real-life setting and compare it to other existing therapies. This approach can be used on both the payer and pharma company side for real-time evidence-based negotiations. If computers can play chess together, the days are not far away when negotiation will happen between two parties through these models. We have examples from other industries eg Walmart where robotic programs already play a role in contract deal negotiations with a successful output.
The current process of price approval post-drug registration has a long lead time because of how long the negotiation takes. The entire process is quite resource intensive and consumes extensive human effort even today. A typical initiation, review and approval process takes up to 50 weeks i.e. close to 1 year. In an era of the 3D printing of human organs and the Toyota Summer Olympics Humanoid Robot perfecting long-range basketball three-pointers, the below process looks too vintage and outdated. It is probably time to connect the dots and take a leap of faith.
This critical time can be shortened using AI and natural language processing tools which are already being used for technical writing in other industries. In the end, this will have a huge impact on patients as we bring the drug to the market with the right price more quickly.
The Road Ahead: The Future is Almost Now
It’s a no-brainer that pricing and market access departments need the applicability of AI and ML to quicken the decision-making process and create efficiencies in operational processes. What is needed is a leap of faith and an experimentational approach to creating next-generation AI-based working models. A staged four-step approach is needed which is depicted below:
Made with Visme Infographic Maker
In a nutshell, it’s not the rules of the game but the game itself which is changing. The industry has to get on the road before we miss the bus.
Summary and Recommendations
All of this is easier said than done. Finding resources and budgets amidst the current challenges and priorities is always a tough ask. Change management is another key challenge that companies can face and there are certain legal, regulatory, ethical and privacy issues which need to be figured out. There are challenges galore and the industry must find a way to initiate the set-up for larger-scale benefits.
In the longer run, AI has a clear business case to be a more efficient and cost-effective solution vis a vis the current processes and tools. The operational and decision-making times are going to shrink and hence the go-to-market time will be faster. This will also improve the transparency around our pricing and market access processes. Because of this new approach, there will be a tremendous quality improvement in analysis and government submissions. Negotiations will become easier and last but not least, patients will benefit. Human intervention and validation will always be needed, but AI can be leveraged for efficiency and accuracy.
